Search bar for issues/pulls (#530)

tokarchuk/v1.17
Ethan Koenig 8 years ago committed by Lunny Xiao
parent 8bc431952f
commit 833f8b94c2
  1. 1
      .gitignore
  2. 4
      conf/app.ini
  3. 70
      models/issue.go
  4. 18
      models/issue_comment.go
  5. 183
      models/issue_indexer.go
  6. 4
      models/models.go
  7. 14
      modules/indexer/indexer.go
  8. 6
      modules/setting/setting.go
  9. 25
      modules/util/util.go
  10. 4
      public/css/index.css
  11. 6
      routers/api/v1/repo/issue.go
  12. 2
      routers/api/v1/repo/issue_comment.go
  13. 2
      routers/init.go
  14. 34
      routers/repo/issue.go
  15. 5
      routers/user/home.go
  16. 43
      templates/repo/issue/list.tmpl
  17. 2
      templates/repo/issue/navbar.tmpl
  18. 13
      templates/repo/issue/search.tmpl
  19. 16
      vendor/github.com/blevesearch/bleve/CONTRIBUTING.md
  20. 202
      vendor/github.com/blevesearch/bleve/LICENSE
  21. 62
      vendor/github.com/blevesearch/bleve/README.md
  22. 46
      vendor/github.com/blevesearch/bleve/analysis/analyzer/simple/simple.go
  23. 52
      vendor/github.com/blevesearch/bleve/analysis/analyzer/standard/standard.go
  24. 64
      vendor/github.com/blevesearch/bleve/analysis/datetime/flexible/flexible.go
  25. 45
      vendor/github.com/blevesearch/bleve/analysis/datetime/optional/optional.go
  26. 111
      vendor/github.com/blevesearch/bleve/analysis/freq.go
  27. 70
      vendor/github.com/blevesearch/bleve/analysis/lang/en/analyzer_en.go
  28. 67
      vendor/github.com/blevesearch/bleve/analysis/lang/en/possessive_filter_en.go
  29. 33
      vendor/github.com/blevesearch/bleve/analysis/lang/en/stop_filter_en.go
  30. 344
      vendor/github.com/blevesearch/bleve/analysis/lang/en/stop_words_en.go
  31. 7
      vendor/github.com/blevesearch/bleve/analysis/test_words.txt
  32. 105
      vendor/github.com/blevesearch/bleve/analysis/token/lowercase/lowercase.go
  33. 53
      vendor/github.com/blevesearch/bleve/analysis/token/porter/porter.go
  34. 70
      vendor/github.com/blevesearch/bleve/analysis/token/stop/stop.go
  35. 76
      vendor/github.com/blevesearch/bleve/analysis/tokenizer/character/character.go
  36. 33
      vendor/github.com/blevesearch/bleve/analysis/tokenizer/letter/letter.go
  37. 131
      vendor/github.com/blevesearch/bleve/analysis/tokenizer/unicode/unicode.go
  38. 76
      vendor/github.com/blevesearch/bleve/analysis/tokenmap.go
  39. 103
      vendor/github.com/blevesearch/bleve/analysis/type.go
  40. 92
      vendor/github.com/blevesearch/bleve/analysis/util.go
  41. 88
      vendor/github.com/blevesearch/bleve/config.go
  42. 23
      vendor/github.com/blevesearch/bleve/config_app.go
  43. 25
      vendor/github.com/blevesearch/bleve/config_disk.go
  44. 38
      vendor/github.com/blevesearch/bleve/doc.go
  45. 75
      vendor/github.com/blevesearch/bleve/document/document.go
  46. 39
      vendor/github.com/blevesearch/bleve/document/field.go
  47. 107
      vendor/github.com/blevesearch/bleve/document/field_boolean.go
  48. 99
      vendor/github.com/blevesearch/bleve/document/field_composite.go
  49. 144
      vendor/github.com/blevesearch/bleve/document/field_datetime.go
  50. 130
      vendor/github.com/blevesearch/bleve/document/field_numeric.go
  51. 119
      vendor/github.com/blevesearch/bleve/document/field_text.go
  52. 55
      vendor/github.com/blevesearch/bleve/document/indexing_options.go
  53. 52
      vendor/github.com/blevesearch/bleve/error.go
  54. 243
      vendor/github.com/blevesearch/bleve/index.go
  55. 83
      vendor/github.com/blevesearch/bleve/index/analysis.go
  56. 88
      vendor/github.com/blevesearch/bleve/index/field_cache.go
  57. 239
      vendor/github.com/blevesearch/bleve/index/index.go
  58. 62
      vendor/github.com/blevesearch/bleve/index/store/batch.go
  59. 85
      vendor/github.com/blevesearch/bleve/index/store/boltdb/iterator.go
  60. 73
      vendor/github.com/blevesearch/bleve/index/store/boltdb/reader.go
  61. 26
      vendor/github.com/blevesearch/bleve/index/store/boltdb/stats.go
  62. 175
      vendor/github.com/blevesearch/bleve/index/store/boltdb/store.go
  63. 95
      vendor/github.com/blevesearch/bleve/index/store/boltdb/writer.go
  64. 152
      vendor/github.com/blevesearch/bleve/index/store/gtreap/iterator.go
  65. 66
      vendor/github.com/blevesearch/bleve/index/store/gtreap/reader.go
  66. 82
      vendor/github.com/blevesearch/bleve/index/store/gtreap/store.go
  67. 76
      vendor/github.com/blevesearch/bleve/index/store/gtreap/writer.go
  68. 174
      vendor/github.com/blevesearch/bleve/index/store/kvstore.go
  69. 64
      vendor/github.com/blevesearch/bleve/index/store/merge.go
  70. 33
      vendor/github.com/blevesearch/bleve/index/store/multiget.go
  71. 110
      vendor/github.com/blevesearch/bleve/index/upsidedown/analysis.go
  72. 8
      vendor/github.com/blevesearch/bleve/index/upsidedown/benchmark_all.sh
  73. 172
      vendor/github.com/blevesearch/bleve/index/upsidedown/dump.go
  74. 78
      vendor/github.com/blevesearch/bleve/index/upsidedown/field_dict.go
  75. 189
      vendor/github.com/blevesearch/bleve/index/upsidedown/index_reader.go
  76. 325
      vendor/github.com/blevesearch/bleve/index/upsidedown/reader.go
  77. 853
      vendor/github.com/blevesearch/bleve/index/upsidedown/row.go
  78. 76
      vendor/github.com/blevesearch/bleve/index/upsidedown/row_merge.go
  79. 55
      vendor/github.com/blevesearch/bleve/index/upsidedown/stats.go
  80. 1037
      vendor/github.com/blevesearch/bleve/index/upsidedown/upsidedown.go
  81. 684
      vendor/github.com/blevesearch/bleve/index/upsidedown/upsidedown.pb.go
  82. 14
      vendor/github.com/blevesearch/bleve/index/upsidedown/upsidedown.proto
  83. 37
      vendor/github.com/blevesearch/bleve/index_alias.go
  84. 605
      vendor/github.com/blevesearch/bleve/index_alias_impl.go
  85. 729
      vendor/github.com/blevesearch/bleve/index_impl.go
  86. 96
      vendor/github.com/blevesearch/bleve/index_meta.go
  87. 75
      vendor/github.com/blevesearch/bleve/index_stats.go
  88. 61
      vendor/github.com/blevesearch/bleve/mapping.go
  89. 99
      vendor/github.com/blevesearch/bleve/mapping/analysis.go
  90. 490
      vendor/github.com/blevesearch/bleve/mapping/document.go
  91. 296
      vendor/github.com/blevesearch/bleve/mapping/field.go
  92. 430
      vendor/github.com/blevesearch/bleve/mapping/index.go
  93. 49
      vendor/github.com/blevesearch/bleve/mapping/mapping.go
  94. 89
      vendor/github.com/blevesearch/bleve/mapping/reflect.go
  95. 34
      vendor/github.com/blevesearch/bleve/numeric/float.go
  96. 92
      vendor/github.com/blevesearch/bleve/numeric/prefix_coded.go
  97. 186
      vendor/github.com/blevesearch/bleve/query.go
  98. 89
      vendor/github.com/blevesearch/bleve/registry/analyzer.go
  99. 87
      vendor/github.com/blevesearch/bleve/registry/cache.go
  100. 89
      vendor/github.com/blevesearch/bleve/registry/char_filter.go
  101. Some files were not shown because too many files have changed in this diff Show More

1
.gitignore vendored

@ -41,5 +41,6 @@ coverage.out
/dist
/custom
/data
/indexers
/log
/public/img/avatar

4
conf/app.ini vendored

@ -158,6 +158,10 @@ SSL_MODE = disable
; For "sqlite3" and "tidb", use absolute path when you start as service
PATH = data/gitea.db
[indexer]
ISSUE_INDEXER_PATH = indexers/issues.bleve
UPDATE_BUFFER_LEN = 20
[admin]
[security]

@ -17,6 +17,7 @@ import (
"code.gitea.io/gitea/modules/base"
"code.gitea.io/gitea/modules/log"
"code.gitea.io/gitea/modules/setting"
"code.gitea.io/gitea/modules/util"
)
var (
@ -451,8 +452,11 @@ func (issue *Issue) ReadBy(userID int64) error {
}
func updateIssueCols(e Engine, issue *Issue, cols ...string) error {
_, err := e.Id(issue.ID).Cols(cols...).Update(issue)
if _, err := e.Id(issue.ID).Cols(cols...).Update(issue); err != nil {
return err
}
UpdateIssueIndexer(issue)
return nil
}
// UpdateIssueCols only updates values of specific columns for given issue.
@ -733,6 +737,8 @@ func newIssue(e *xorm.Session, opts NewIssueOptions) (err error) {
return err
}
UpdateIssueIndexer(opts.Issue)
if len(opts.Attachments) > 0 {
attachments, err := getAttachmentsByUUIDs(e, opts.Attachments)
if err != nil {
@ -865,10 +871,11 @@ type IssuesOptions struct {
MilestoneID int64
RepoIDs []int64
Page int
IsClosed bool
IsPull bool
IsClosed util.OptionalBool
IsPull util.OptionalBool
Labels string
SortType string
IssueIDs []int64
}
// sortIssuesSession sort an issues-related session based on the provided
@ -894,11 +901,23 @@ func sortIssuesSession(sess *xorm.Session, sortType string) {
// Issues returns a list of issues by given conditions.
func Issues(opts *IssuesOptions) ([]*Issue, error) {
if opts.Page <= 0 {
opts.Page = 1
var sess *xorm.Session
if opts.Page >= 0 {
var start int
if opts.Page == 0 {
start = 0
} else {
start = (opts.Page - 1) * setting.UI.IssuePagingNum
}
sess = x.Limit(setting.UI.IssuePagingNum, start)
} else {
sess = x.NewSession()
defer sess.Close()
}
sess := x.Limit(setting.UI.IssuePagingNum, (opts.Page-1)*setting.UI.IssuePagingNum)
if len(opts.IssueIDs) > 0 {
sess.In("issue.id", opts.IssueIDs)
}
if opts.RepoID > 0 {
sess.And("issue.repo_id=?", opts.RepoID)
@ -906,7 +925,13 @@ func Issues(opts *IssuesOptions) ([]*Issue, error) {
// In case repository IDs are provided but actually no repository has issue.
sess.In("issue.repo_id", opts.RepoIDs)
}
sess.And("issue.is_closed=?", opts.IsClosed)
switch opts.IsClosed {
case util.OptionalBoolTrue:
sess.And("issue.is_closed=true")
case util.OptionalBoolFalse:
sess.And("issue.is_closed=false")
}
if opts.AssigneeID > 0 {
sess.And("issue.assignee_id=?", opts.AssigneeID)
@ -926,7 +951,12 @@ func Issues(opts *IssuesOptions) ([]*Issue, error) {
sess.And("issue.milestone_id=?", opts.MilestoneID)
}
sess.And("issue.is_pull=?", opts.IsPull)
switch opts.IsPull {
case util.OptionalBoolTrue:
sess.And("issue.is_pull=true")
case util.OptionalBoolFalse:
sess.And("issue.is_pull=false")
}
sortIssuesSession(sess, opts.SortType)
@ -1168,10 +1198,11 @@ type IssueStatsOptions struct {
MentionedID int64
PosterID int64
IsPull bool
IssueIDs []int64
}
// GetIssueStats returns issue statistic information by given conditions.
func GetIssueStats(opts *IssueStatsOptions) *IssueStats {
func GetIssueStats(opts *IssueStatsOptions) (*IssueStats, error) {
stats := &IssueStats{}
countSession := func(opts *IssueStatsOptions) *xorm.Session {
@ -1179,6 +1210,10 @@ func GetIssueStats(opts *IssueStatsOptions) *IssueStats {
Where("issue.repo_id = ?", opts.RepoID).
And("is_pull = ?", opts.IsPull)
if len(opts.IssueIDs) > 0 {
sess.In("issue.id", opts.IssueIDs)
}
if len(opts.Labels) > 0 && opts.Labels != "0" {
labelIDs, err := base.StringsToInt64s(strings.Split(opts.Labels, ","))
if err != nil {
@ -1210,13 +1245,20 @@ func GetIssueStats(opts *IssueStatsOptions) *IssueStats {
return sess
}
stats.OpenCount, _ = countSession(opts).
var err error
stats.OpenCount, err = countSession(opts).
And("is_closed = ?", false).
Count(&Issue{})
stats.ClosedCount, _ = countSession(opts).
if err != nil {
return nil, err
}
stats.ClosedCount, err = countSession(opts).
And("is_closed = ?", true).
Count(&Issue{})
return stats
if err != nil {
return nil, err
}
return stats, nil
}
// GetUserIssueStats returns issue statistic information for dashboard by given conditions.
@ -1294,7 +1336,11 @@ func GetRepoIssueStats(repoID, uid int64, filterMode int, isPull bool) (numOpen
func updateIssue(e Engine, issue *Issue) error {
_, err := e.Id(issue.ID).AllCols().Update(issue)
if err != nil {
return err
}
UpdateIssueIndexer(issue)
return nil
}
// UpdateIssue updates all fields of given issue.

@ -454,28 +454,20 @@ func UpdateComment(c *Comment) error {
return err
}
// DeleteCommentByID deletes the comment by given ID.
func DeleteCommentByID(id int64) error {
comment, err := GetCommentByID(id)
if err != nil {
if IsErrCommentNotExist(err) {
return nil
}
return err
}
// DeleteComment deletes the comment
func DeleteComment(comment *Comment) error {
sess := x.NewSession()
defer sessionRelease(sess)
if err = sess.Begin(); err != nil {
if err := sess.Begin(); err != nil {
return err
}
if _, err = sess.Id(comment.ID).Delete(new(Comment)); err != nil {
if _, err := sess.Id(comment.ID).Delete(new(Comment)); err != nil {
return err
}
if comment.Type == CommentTypeComment {
if _, err = sess.Exec("UPDATE `issue` SET num_comments = num_comments - 1 WHERE id = ?", comment.IssueID); err != nil {
if _, err := sess.Exec("UPDATE `issue` SET num_comments = num_comments - 1 WHERE id = ?", comment.IssueID); err != nil {
return err
}
}

@ -0,0 +1,183 @@
// Copyright 2017 The Gitea Authors. All rights reserved.
// Use of this source code is governed by a MIT-style
// license that can be found in the LICENSE file.
package models
import (
"fmt"
"os"
"strconv"
"strings"
"code.gitea.io/gitea/modules/log"
"code.gitea.io/gitea/modules/setting"
"code.gitea.io/gitea/modules/util"
"github.com/blevesearch/bleve"
"github.com/blevesearch/bleve/analysis/analyzer/simple"
"github.com/blevesearch/bleve/search/query"
)
// issueIndexerUpdateQueue queue of issues that need to be updated in the issues
// indexer
var issueIndexerUpdateQueue chan *Issue
// issueIndexer (thread-safe) index for searching issues
var issueIndexer bleve.Index
// issueIndexerData data stored in the issue indexer
type issueIndexerData struct {
ID int64
RepoID int64
Title string
Content string
}
// numericQuery an numeric-equality query for the given value and field
func numericQuery(value int64, field string) *query.NumericRangeQuery {
f := float64(value)
tru := true
q := bleve.NewNumericRangeInclusiveQuery(&f, &f, &tru, &tru)
q.SetField(field)
return q
}
// SearchIssuesByKeyword searches for issues by given conditions.
// Returns the matching issue IDs
func SearchIssuesByKeyword(repoID int64, keyword string) ([]int64, error) {
fields := strings.Fields(strings.ToLower(keyword))
indexerQuery := bleve.NewConjunctionQuery(
numericQuery(repoID, "RepoID"),
bleve.NewDisjunctionQuery(
bleve.NewPhraseQuery(fields, "Title"),
bleve.NewPhraseQuery(fields, "Content"),
))
search := bleve.NewSearchRequestOptions(indexerQuery, 2147483647, 0, false)
search.Fields = []string{"ID"}
result, err := issueIndexer.Search(search)
if err != nil {
return nil, err
}
issueIDs := make([]int64, len(result.Hits))
for i, hit := range result.Hits {
issueIDs[i] = int64(hit.Fields["ID"].(float64))
}
return issueIDs, nil
}
// InitIssueIndexer initialize issue indexer
func InitIssueIndexer() {
_, err := os.Stat(setting.Indexer.IssuePath)
if err != nil {
if os.IsNotExist(err) {
if err = createIssueIndexer(); err != nil {
log.Fatal(4, "CreateIssuesIndexer: %v", err)
}
if err = populateIssueIndexer(); err != nil {
log.Fatal(4, "PopulateIssuesIndex: %v", err)
}
} else {
log.Fatal(4, "InitIssuesIndexer: %v", err)
}
} else {
issueIndexer, err = bleve.Open(setting.Indexer.IssuePath)
if err != nil {
log.Fatal(4, "InitIssuesIndexer, open index: %v", err)
}
}
issueIndexerUpdateQueue = make(chan *Issue, setting.Indexer.UpdateQueueLength)
go processIssueIndexerUpdateQueue()
// TODO close issueIndexer when Gitea closes
}
// createIssueIndexer create an issue indexer if one does not already exist
func createIssueIndexer() error {
mapping := bleve.NewIndexMapping()
docMapping := bleve.NewDocumentMapping()
docMapping.AddFieldMappingsAt("ID", bleve.NewNumericFieldMapping())
docMapping.AddFieldMappingsAt("RepoID", bleve.NewNumericFieldMapping())
textFieldMapping := bleve.NewTextFieldMapping()
textFieldMapping.Analyzer = simple.Name
docMapping.AddFieldMappingsAt("Title", textFieldMapping)
docMapping.AddFieldMappingsAt("Content", textFieldMapping)
mapping.AddDocumentMapping("issues", docMapping)
var err error
issueIndexer, err = bleve.New(setting.Indexer.IssuePath, mapping)
return err
}
// populateIssueIndexer populate the issue indexer with issue data
func populateIssueIndexer() error {
for page := 1; ; page++ {
repos, err := Repositories(&SearchRepoOptions{
Page: page,
PageSize: 10,
})
if err != nil {
return fmt.Errorf("Repositories: %v", err)
}
if len(repos) == 0 {
return nil
}
batch := issueIndexer.NewBatch()
for _, repo := range repos {
issues, err := Issues(&IssuesOptions{
RepoID: repo.ID,
IsClosed: util.OptionalBoolNone,
IsPull: util.OptionalBoolNone,
Page: -1, // do not page
})
if err != nil {
return fmt.Errorf("Issues: %v", err)
}
for _, issue := range issues {
err = batch.Index(issue.indexUID(), issue.issueData())
if err != nil {
return fmt.Errorf("batch.Index: %v", err)
}
}
}
if err = issueIndexer.Batch(batch); err != nil {
return fmt.Errorf("index.Batch: %v", err)
}
}
}
func processIssueIndexerUpdateQueue() {
for {
select {
case issue := <-issueIndexerUpdateQueue:
if err := issueIndexer.Index(issue.indexUID(), issue.issueData()); err != nil {
log.Error(4, "issuesIndexer.Index: %v", err)
}
}
}
}
// indexUID a unique identifier for an issue used in full-text indices
func (issue *Issue) indexUID() string {
return strconv.FormatInt(issue.ID, 36)
}
func (issue *Issue) issueData() *issueIndexerData {
return &issueIndexerData{
ID: issue.ID,
RepoID: issue.RepoID,
Title: issue.Title,
Content: issue.Content,
}
}
// UpdateIssueIndexer add/update an issue to the issue indexer
func UpdateIssueIndexer(issue *Issue) {
go func() {
issueIndexerUpdateQueue <- issue
}()
}

@ -138,6 +138,10 @@ func LoadConfigs() {
}
DbCfg.SSLMode = sec.Key("SSL_MODE").String()
DbCfg.Path = sec.Key("PATH").MustString("data/gitea.db")
sec = setting.Cfg.Section("indexer")
setting.Indexer.IssuePath = sec.Key("ISSUE_INDEXER_PATH").MustString("indexers/issues.bleve")
setting.Indexer.UpdateQueueLength = sec.Key("UPDATE_BUFFER_LEN").MustInt(20)
}
// parsePostgreSQLHostPort parses given input in various forms defined in

@ -0,0 +1,14 @@
// Copyright 2016 The Gitea Authors. All rights reserved.
// Use of this source code is governed by a MIT-style
// license that can be found in the LICENSE file.
package indexer
import (
"code.gitea.io/gitea/models"
)
// NewContext start indexer service
func NewContext() {
models.InitIssueIndexer()
}

@ -123,6 +123,12 @@ var (
UsePostgreSQL bool
UseTiDB bool
// Indexer settings
Indexer struct {
IssuePath string
UpdateQueueLength int
}
// Webhook settings
Webhook = struct {
QueueLength int

@ -0,0 +1,25 @@
// Copyright 2017 The Gitea Authors. All rights reserved.
// Use of this source code is governed by a MIT-style
// license that can be found in the LICENSE file.
package util
// OptionalBool a boolean that can be "null"
type OptionalBool byte
const (
// OptionalBoolNone a "null" boolean value
OptionalBoolNone = iota
// OptionalBoolTrue a "true" boolean value
OptionalBoolTrue
// OptionalBoolFalse a "false" boolean value
OptionalBoolFalse
)
// OptionalBoolOf get the corresponding OptionalBool of a bool
func OptionalBoolOf(b bool) OptionalBool {
if b {
return OptionalBoolTrue
}
return OptionalBoolFalse
}

@ -2926,6 +2926,10 @@ footer .ui.language .menu {
width: 16px;
text-align: center;
}
.navbar {
display: flex;
justify-content: space-between;
}
.ui.repository.list .item {
padding-bottom: 25px;
}

@ -13,14 +13,16 @@ import (
"code.gitea.io/gitea/models"
"code.gitea.io/gitea/modules/context"
"code.gitea.io/gitea/modules/setting"
"code.gitea.io/gitea/modules/util"
)
// ListIssues list the issues of a repository
func ListIssues(ctx *context.APIContext) {
isClosed := ctx.Query("state") == "closed"
issueOpts := models.IssuesOptions{
RepoID: ctx.Repo.Repository.ID,
Page: ctx.QueryInt("page"),
IsClosed: ctx.Query("state") == "closed",
IsClosed: util.OptionalBoolOf(isClosed),
}
issues, err := models.Issues(&issueOpts)
@ -29,7 +31,7 @@ func ListIssues(ctx *context.APIContext) {
return
}
if ctx.Query("state") == "all" {
issueOpts.IsClosed = !issueOpts.IsClosed
issueOpts.IsClosed = util.OptionalBoolOf(!isClosed)
tempIssues, err := models.Issues(&issueOpts)
if err != nil {
ctx.Error(500, "Issues", err)

@ -125,7 +125,7 @@ func DeleteIssueComment(ctx *context.APIContext) {
return
}
if err = models.DeleteCommentByID(comment.ID); err != nil {
if err = models.DeleteComment(comment); err != nil {
ctx.Error(500, "DeleteCommentByID", err)
return
}

@ -18,6 +18,7 @@ import (
"code.gitea.io/gitea/modules/setting"
"code.gitea.io/gitea/modules/ssh"
macaron "gopkg.in/macaron.v1"
"code.gitea.io/gitea/modules/indexer"
)
func checkRunMode() {
@ -36,6 +37,7 @@ func checkRunMode() {
func NewServices() {
setting.NewServices()
mailer.NewContext()
indexer.NewContext()
}
// GlobalInit is for global configuration reload-able.

@ -5,6 +5,7 @@
package repo
import (
"bytes"
"errors"
"fmt"
"io"
@ -25,6 +26,7 @@ import (
"code.gitea.io/gitea/modules/markdown"
"code.gitea.io/gitea/modules/notification"
"code.gitea.io/gitea/modules/setting"
"code.gitea.io/gitea/modules/util"
)
const (
@ -158,20 +160,39 @@ func Issues(ctx *context.Context) {
milestoneID := ctx.QueryInt64("milestone")
isShowClosed := ctx.Query("state") == "closed"
keyword := ctx.Query("q")
if bytes.Contains([]byte(keyword), []byte{0x00}) {
keyword = ""
}
var issueIDs []int64
var err error
if len(keyword) > 0 {
issueIDs, err = models.SearchIssuesByKeyword(repo.ID, keyword)
if len(issueIDs) == 0 {
forceEmpty = true
}
}
var issueStats *models.IssueStats
if forceEmpty {
issueStats = &models.IssueStats{}
} else {
issueStats = models.GetIssueStats(&models.IssueStatsOptions{
var err error
issueStats, err = models.GetIssueStats(&models.IssueStatsOptions{
RepoID: repo.ID,
Labels: selectLabels,
MilestoneID: milestoneID,
AssigneeID: assigneeID,
MentionedID: mentionedID,
IsPull: isPullList,
IssueIDs: issueIDs,
})
if err != nil {
ctx.Error(500, "GetSearchIssueStats")
return
}
}
page := ctx.QueryInt("page")
if page <= 1 {
page = 1
@ -190,7 +211,6 @@ func Issues(ctx *context.Context) {
if forceEmpty {
issues = []*models.Issue{}
} else {
var err error
issues, err = models.Issues(&models.IssuesOptions{
AssigneeID: assigneeID,
RepoID: repo.ID,
@ -198,10 +218,11 @@ func Issues(ctx *context.Context) {
MentionedID: mentionedID,
MilestoneID: milestoneID,
Page: pager.Current(),
IsClosed: isShowClosed,
IsPull: isPullList,
IsClosed: util.OptionalBoolOf(isShowClosed),
IsPull: util.OptionalBoolOf(isPullList),
Labels: selectLabels,
SortType: sortType,
IssueIDs: issueIDs,
})
if err != nil {
ctx.Handle(500, "Issues", err)
@ -258,6 +279,7 @@ func Issues(ctx *context.Context) {
ctx.Data["MilestoneID"] = milestoneID
ctx.Data["AssigneeID"] = assigneeID
ctx.Data["IsShowClosed"] = isShowClosed
ctx.Data["Keyword"] = keyword
if isShowClosed {
ctx.Data["State"] = "closed"
} else {
@ -934,7 +956,7 @@ func DeleteComment(ctx *context.Context) {
return
}
if err = models.DeleteCommentByID(comment.ID); err != nil {
if err = models.DeleteComment(comment); err != nil {
ctx.Handle(500, "DeleteCommentByID", err)
return
}

@ -15,6 +15,7 @@ import (
"code.gitea.io/gitea/modules/base"
"code.gitea.io/gitea/modules/context"
"code.gitea.io/gitea/modules/setting"
"code.gitea.io/gitea/modules/util"
)
const (
@ -277,8 +278,8 @@ func Issues(ctx *context.Context) {
PosterID: posterID,
RepoIDs: repoIDs,
Page: page,
IsClosed: isShowClosed,
IsPull: isPullList,
IsClosed: util.OptionalBoolOf(isShowClosed),
IsPull: util.OptionalBoolOf(isPullList),
SortType: sortType,
})
if err != nil {

@ -4,6 +4,7 @@
<div class="ui container">
<div class="navbar">
{{template "repo/issue/navbar" .}}
{{template "repo/issue/search" .}}
<div class="ui right">
{{if .PageIsIssueList}}
<a class="ui green button" href="{{.RepoLink}}/issues/new">{{.i18n.Tr "repo.issues.new"}}</a>
@ -14,11 +15,11 @@
</div>
<div class="ui divider"></div>
<div class="ui tiny basic status buttons">
<a class="ui {{if not .IsShowClosed}}green active{{end}} basic button" href="{{$.Link}}?type={{$.ViewType}}&sort={{$.SortType}}&state=open&labels={{.SelectLabels}}&milestone={{.MilestoneID}}&assignee={{.AssigneeID}}">
<a class="ui {{if not .IsShowClosed}}green active{{end}} basic button" href="{{$.Link}}?q={{$.Keyword}}&type={{$.ViewType}}&sort={{$.SortType}}&state=open&labels={{.SelectLabels}}&milestone={{.MilestoneID}}&assignee={{.AssigneeID}}">
<i class="octicon octicon-issue-opened"></i>
{{.i18n.Tr "repo.issues.open_tab" .IssueStats.OpenCount}}
</a>
<a class="ui {{if .IsShowClosed}}red active{{end}} basic button" href="{{$.Link}}?type={{.ViewType}}&sort={{$.SortType}}&state=closed&labels={{.SelectLabels}}&milestone={{.MilestoneID}}&assignee={{.AssigneeID}}">
<a class="ui {{if .IsShowClosed}}red active{{end}} basic button" href="{{$.Link}}?q={{$.Keyword}}&type={{.ViewType}}&sort={{$.SortType}}&state=closed&labels={{.SelectLabels}}&milestone={{.MilestoneID}}&assignee={{.AssigneeID}}">
<i class="octicon octicon-issue-closed"></i>
{{.i18n.Tr "repo.issues.close_tab" .IssueStats.ClosedCount}}
</a>
@ -31,9 +32,9 @@
<i class="dropdown icon"></i>
</span>
<div class="menu">
<a class="item" href="{{$.Link}}?type={{$.ViewType}}&sort={{$.SortType}}&state={{$.State}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_label_no_select"}}</a>
<a class="item" href="{{$.Link}}?q={{$.Keyword}}&type={{$.ViewType}}&sort={{$.SortType}}&state={{$.State}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_label_no_select"}}</a>
{{range .Labels}}
<a class="item" href="{{$.Link}}?type={{$.ViewType}}&sort={{$.SortType}}&state={{$.State}}&labels={{.ID}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}"><span class="octicon {{if eq $.SelectLabels .ID}}octicon-check{{end}}"></span><span class="label color" style="background-color: {{.Color}}"></span> {{.Name}}</a>
<a class="item" href="{{$.Link}}?q={{$.Keyword}}&type={{$.ViewType}}&sort={{$.SortType}}&state={{$.State}}&labels={{.ID}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}"><span class="octicon {{if eq $.SelectLabels .ID}}octicon-check{{end}}"></span><span class="label color" style="background-color: {{.Color}}"></span> {{.Name}}</a>
{{end}}
</div>
</div>
@ -45,7 +46,7 @@
<i class="dropdown icon"></i>
</span>
<div class="menu">
<a class="item" href="{{$.Link}}?type={{$.ViewType}}&sort={{$.SortType}}&state={{$.State}}&labels={{.SelectLabels}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_milestone_no_select"}}</a>
<a class="item" href="{{$.Link}}?q={{$.Keyword}}&type={{$.ViewType}}&sort={{$.SortType}}&state={{$.State}}&labels={{.SelectLabels}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_milestone_no_select"}}</a>
{{range .Milestones}}
<a class="{{if eq $.MilestoneID .ID}}active selected{{end}} item" href="{{$.Link}}?type={{$.ViewType}}&sort={{$.SortType}}&state={{$.State}}&labels={{$.SelectLabels}}&milestone={{.ID}}&assignee={{$.AssigneeID}}">{{.Name}}</a>
{{end}}
@ -59,7 +60,7 @@
<i class="dropdown icon"></i>
</span>
<div class="menu">
<a class="item" href="{{$.Link}}?type={{$.ViewType}}&sort={{$.SortType}}&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}">{{.i18n.Tr "repo.issues.filter_assginee_no_select"}}</a>
<a class="item" href="{{$.Link}}?q={{$.Keyword}}&type={{$.ViewType}}&sort={{$.SortType}}&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}">{{.i18n.Tr "repo.issues.filter_assginee_no_select"}}</a>
{{range .Assignees}}
<a class="{{if eq $.AssigneeID .ID}}active selected{{end}} item" href="{{$.Link}}?type={{$.ViewType}}&sort={{$.SortType}}&state={{$.State}}&labels={{$.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{.ID}}"><img src="{{.RelAvatarLink}}"> {{.Name}}</a>
{{end}}
@ -73,10 +74,10 @@
<i class="dropdown icon"></i>
</span>
<div class="menu">
<a class="{{if eq .ViewType "all"}}active{{end}} item" href="{{$.Link}}?type=all&sort={{$.SortType}}&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_type.all_issues"}}</a>
<a class="{{if eq .ViewType "assigned"}}active{{end}} item" href="{{$.Link}}?type=assigned&sort={{$.SortType}}&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_type.assigned_to_you"}}</a>
<a class="{{if eq .ViewType "created_by"}}active{{end}} item" href="{{$.Link}}?type=created_by&sort={{$.SortType}}&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_type.created_by_you"}}</a>
<a class="{{if eq .ViewType "mentioned"}}active{{end}} item" href="{{$.Link}}?type=mentioned&sort={{$.SortType}}&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_type.mentioning_you"}}</a>
<a class="{{if eq .ViewType "all"}}active{{end}} item" href="{{$.Link}}?q={{$.Keyword}}&type=all&sort={{$.SortType}}&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_type.all_issues"}}</a>
<a class="{{if eq .ViewType "assigned"}}active{{end}} item" href="{{$.Link}}?q={{$.Keyword}}&type=assigned&sort={{$.SortType}}&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_type.assigned_to_you"}}</a>
<a class="{{if eq .ViewType "created_by"}}active{{end}} item" href="{{$.Link}}?q={{$.Keyword}}&type=created_by&sort={{$.SortType}}&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_type.created_by_you"}}</a>
<a class="{{if eq .ViewType "mentioned"}}active{{end}} item" href="{{$.Link}}?q={{$.Keyword}}&type=mentioned&sort={{$.SortType}}&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_type.mentioning_you"}}</a>
</div>
</div>
@ -87,12 +88,12 @@
<i class="dropdown icon"></i>
</span>
<div class="menu">
<a class="{{if or (eq .SortType "latest") (not .SortType)}}active{{end}} item" href="{{$.Link}}?type={{$.ViewType}}&sort=latest&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_sort.latest"}}</a>
<a class="{{if eq .SortType "oldest"}}active{{end}} item" href="{{$.Link}}?type={{$.ViewType}}&sort=oldest&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_sort.oldest"}}</a>
<a class="{{if eq .SortType "recentupdate"}}active{{end}} item" href="{{$.Link}}?type={{$.ViewType}}&sort=recentupdate&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_sort.recentupdate"}}</a>
<a class="{{if eq .SortType "leastupdate"}}active{{end}} item" href="{{$.Link}}?type={{$.ViewType}}&sort=leastupdate&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_sort.leastupdate"}}</a>
<a class="{{if eq .SortType "mostcomment"}}active{{end}} item" href="{{$.Link}}?type={{$.ViewType}}&sort=mostcomment&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_sort.mostcomment"}}</a>
<a class="{{if eq .SortType "leastcomment"}}active{{end}} item" href="{{$.Link}}?type={{$.ViewType}}&sort=leastcomment&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_sort.leastcomment"}}</a>
<a class="{{if or (eq .SortType "latest") (not .SortType)}}active{{end}} item" href="{{$.Link}}?q={{$.Keyword}}&type={{$.ViewType}}&sort=latest&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_sort.latest"}}</a>
<a class="{{if eq .SortType "oldest"}}active{{end}} item" href="{{$.Link}}?q={{$.Keyword}}&type={{$.ViewType}}&sort=oldest&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_sort.oldest"}}</a>
<a class="{{if eq .SortType "recentupdate"}}active{{end}} item" href="{{$.Link}}?q={{$.Keyword}}&type={{$.ViewType}}&sort=recentupdate&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_sort.recentupdate"}}</a>
<a class="{{if eq .SortType "leastupdate"}}active{{end}} item" href="{{$.Link}}?q={{$.Keyword}}&type={{$.ViewType}}&sort=leastupdate&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_sort.leastupdate"}}</a>
<a class="{{if eq .SortType "mostcomment"}}active{{end}} item" href="{{$.Link}}?q={{$.Keyword}}&type={{$.ViewType}}&sort=mostcomment&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_sort.mostcomment"}}</a>
<a class="{{if eq .SortType "leastcomment"}}active{{end}} item" href="{{$.Link}}?q={{$.Keyword}}&type={{$.ViewType}}&sort=leastcomment&state={{$.State}}&labels={{.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}">{{.i18n.Tr "repo.issues.filter_sort.leastcomment"}}</a>
</div>
</div>
</div>
@ -105,7 +106,7 @@
<a class="title has-emoji" href="{{$.Link}}/{{.Index}}">{{.Title}}</a>
{{range .Labels}}
<a class="ui label" href="{{$.Link}}?type={{$.ViewType}}&state={{$.State}}&labels={{.ID}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}" style="color: {{.ForegroundColor}}; background-color: {{.Color}}">{{.Name}}</a>
<a class="ui label" href="{{$.Link}}?q={{$.Keyword}}&type={{$.ViewType}}&state={{$.State}}&labels={{.ID}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}" style="color: {{.ForegroundColor}}; background-color: {{.Color}}">{{.Name}}</a>
{{end}}
{{if .NumComments}}
@ -115,7 +116,7 @@
<p class="desc">
{{$.i18n.Tr "repo.issues.opened_by" $timeStr .Poster.HomeLink .Poster.Name | Safe}}
{{if .Milestone}}
<a class="milestone" href="{{$.Link}}?type={{$.ViewType}}&state={{$.State}}&labels={{$.SelectLabels}}&milestone={{.Milestone.ID}}&assignee={{$.AssigneeID}}">
<a class="milestone" href="{{$.Link}}?q={{$.Keyword}}&type={{$.ViewType}}&state={{$.State}}&labels={{$.SelectLabels}}&milestone={{.Milestone.ID}}&assignee={{$.AssigneeID}}">
<span class="octicon octicon-milestone"></span> {{.Milestone.Name}}
</a>
{{end}}
@ -132,17 +133,17 @@
{{if gt .TotalPages 1}}
<div class="center page buttons">
<div class="ui borderless pagination menu">
<a class="{{if not .HasPrevious}}disabled{{end}} item" {{if .HasPrevious}}href="{{$.Link}}?type={{$.ViewType}}&sort={{$.SortType}}&state={{$.State}}&labels={{$.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}&page={{.Previous}}"{{end}}>
<a class="{{if not .HasPrevious}}disabled{{end}} item" {{if .HasPrevious}}href="{{$.Link}}?q={{$.Keyword}}&type={{$.ViewType}}&sort={{$.SortType}}&state={{$.State}}&labels={{$.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}&page={{.Previous}}"{{end}}>
<i class="left arrow icon"></i> {{$.i18n.Tr "repo.issues.previous"}}
</a>
{{range .Pages}}
{{if eq .Num -1}}
<a class="disabled item">...</a>
{{else}}
<a class="{{if .IsCurrent}}active{{end}} item" {{if not .IsCurrent}}href="{{$.Link}}?type={{$.ViewType}}&sort={{$.SortType}}&state={{$.State}}&labels={{$.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}&page={{.Num}}"{{end}}>{{.Num}}</a>
<a class="{{if .IsCurrent}}active{{end}} item" {{if not .IsCurrent}}href="{{$.Link}}?q={{$.Keyword}}&type={{$.ViewType}}&sort={{$.SortType}}&state={{$.State}}&labels={{$.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}&page={{.Num}}"{{end}}>{{.Num}}</a>
{{end}}
{{end}}
<a class="{{if not .HasNext}}disabled{{end}} item" {{if .HasNext}}href="{{$.Link}}?type={{$.ViewType}}&sort={{$.SortType}}&state={{$.State}}&labels={{$.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}&page={{.Next}}"{{end}}>
<a class="{{if not .HasNext}}disabled{{end}} item" {{if .HasNext}}href="{{$.Link}}?q={{$.Keyword}}&type={{$.ViewType}}&sort={{$.SortType}}&state={{$.State}}&labels={{$.SelectLabels}}&milestone={{$.MilestoneID}}&assignee={{$.AssigneeID}}&page={{.Next}}"{{end}}>
{{$.i18n.Tr "repo.issues.next"}}&nbsp;<i class="icon right arrow"></i>
</a>
</div>

@ -1,4 +1,4 @@
<div class="ui compact small menu">
<div class="ui compact left small menu">
<a class="{{if .PageIsLabels}}active{{end}} item" href="{{.RepoLink}}/labels">{{.i18n.Tr "repo.labels"}}</a>
<a class="{{if .PageIsMilestones}}active{{end}} item" href="{{.RepoLink}}/milestones">{{.i18n.Tr "repo.milestones"}}</a>
</div>

@ -0,0 +1,13 @@
<form class="ui form">
<div class="ui fluid action input">
<input type="hidden" name="type" value="{{$.ViewType}}"/>
<input type="hidden" name="state" value="{{$.State}}"/>
<input type="hidden" name="labels" value="{{.SelectLabels}}"/>
<input type="hidden" name="milestone" value="{{$.MilestoneID}}"/>
<input type="hidden" name="assignee" value="{{$.AssigneeID}}"/>
<div class="ui search action input">
<input name="q" value="{{.Keyword}}" placeholder="{{.i18n.Tr "explore.search"}}..." autofocus>
</div>
<button class="ui blue button" type="submit">{{.i18n.Tr "explore.search"}}</button>
</div>
</form>

@ -0,0 +1,16 @@
# Contributing to Bleve
We look forward to your contributions, but ask that you first review these guidelines.
### Sign the CLA
As Bleve is a Couchbase project we require contributors accept the [Couchbase Contributor License Agreement](http://review.couchbase.org/static/individual_agreement.html). To sign this agreement log into the Couchbase [code review tool](http://review.couchbase.org/). The Bleve project does not use this code review tool but it is still used to track acceptance of the contributor license agreements.
### Submitting a Pull Request
All types of contributions are welcome, but please keep the following in mind:
- If you're planning a large change, you should really discuss it in a github issue or on the google group first. This helps avoid duplicate effort and spending time on something that may not be merged.
- Existing tests should continue to pass, new tests for the contribution are nice to have.
- All code should have gone through `go fmt`
- All code should pass `go vet`

@ -0,0 +1,202 @@
Apache License
Version 2.0, January 2004
http://www.apache.org/licenses/
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
1. Definitions.
"License" shall mean the terms and conditions for use, reproduction,
and distribution as defined by Sections 1 through 9 of this document.
"Licensor" shall mean the copyright owner or entity authorized by
the copyright owner that is granting the License.
"Legal Entity" shall mean the union of the acting entity and all
other entities that control, are controlled by, or are under common
control with that entity. For the purposes of this definition,
"control" means (i) the power, direct or indirect, to cause the
direction or management of such entity, whether by contract or
otherwise, or (ii) ownership of fifty percent (50%) or more of the
outstanding shares, or (iii) beneficial ownership of such entity.
"You" (or "Your") shall mean an individual or Legal Entity
exercising permissions granted by this License.
"Source" form shall mean the preferred form for making modifications,
including but not limited to software source code, documentation
source, and configuration files.
"Object" form shall mean any form resulting from mechanical
transformation or translation of a Source form, including but
not limited to compiled object code, generated documentation,
and conversions to other media types.
"Work" shall mean the work of authorship, whether in Source or
Object form, made available under the License, as indicated by a
copyright notice that is included in or attached to the work
(an example is provided in the Appendix below).
"Derivative Works" shall mean any work, whether in Source or Object
form, that is based on (or derived from) the Work and for which the
editorial revisions, annotations, elaborations, or other modifications
represent, as a whole, an original work of authorship. For the purposes
of this License, Derivative Works shall not include works that remain
separable from, or merely link (or bind by name) to the interfaces of,
the Work and Derivative Works thereof.
"Contribution" shall mean any work of authorship, including
the original version of the Work and any modifications or additions
to that Work or Derivative Works thereof, that is intentionally
submitted to Licensor for inclusion in the Work by the copyright owner
or by an individual or Legal Entity authorized to submit on behalf of
the copyright owner. For the purposes of this definition, "submitted"
means any form of electronic, verbal, or written communication sent
to the Licensor or its representatives, including but not limited to
communication on electronic mailing lists, source code control systems,
and issue tracking systems that are managed by, or on behalf of, the
Licensor for the purpose of discussing and improving the Work, but
excluding communication that is conspicuously marked or otherwise
designated in writing by the copyright owner as "Not a Contribution."
"Contributor" shall mean Licensor and any individual or Legal Entity
on behalf of whom a Contribution has been received by Licensor and
subsequently incorporated within the Work.
2. Grant of Copyright License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
copyright license to reproduce, prepare Derivative Works of,
publicly display, publicly perform, sublicense, and distribute the
Work and such Derivative Works in Source or Object form.
3. Grant of Patent License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
(except as stated in this section) patent license to make, have made,
use, offer to sell, sell, import, and otherwise transfer the Work,
where such license applies only to those patent claims licensable
by such Contributor that are necessarily infringed by their
Contribution(s) alone or by combination of their Contribution(s)
with the Work to which such Contribution(s) was submitted. If You
institute patent litigation against any entity (including a
cross-claim or counterclaim in a lawsuit) alleging that the Work
or a Contribution incorporated within the Work constitutes direct
or contributory patent infringement, then any patent licenses
granted to You under this License for that Work shall terminate
as of the date such litigation is filed.
4. Redistribution. You may reproduce and distribute copies of the
Work or Derivative Works thereof in any medium, with or without
modifications, and in Source or Object form, provided that You
meet the following conditions:
(a) You must give any other recipients of the Work or
Derivative Works a copy of this License; and
(b) You must cause any modified files to carry prominent notices
stating that You changed the files; and
(c) You must retain, in the Source form of any Derivative Works
that You distribute, all copyright, patent, trademark, and
attribution notices from the Source form of the Work,
excluding those notices that do not pertain to any part of
the Derivative Works; and
(d) If the Work includes a "NOTICE" text file as part of its
distribution, then any Derivative Works that You distribute must
include a readable copy of the attribution notices contained
within such NOTICE file, excluding those notices that do not
pertain to any part of the Derivative Works, in at least one
of the following places: within a NOTICE text file distributed
as part of the Derivative Works; within the Source form or
documentation, if provided along with the Derivative Works; or,
within a display generated by the Derivative Works, if and
wherever such third-party notices normally appear. The contents
of the NOTICE file are for informational purposes only and
do not modify the License. You may add Your own attribution
notices within Derivative Works that You distribute, alongside
or as an addendum to the NOTICE text from the Work, provided
that such additional attribution notices cannot be construed
as modifying the License.
You may add Your own copyright statement to Your modifications and
may provide additional or different license terms and conditions
for use, reproduction, or distribution of Your modifications, or
for any such Derivative Works as a whole, provided Your use,
reproduction, and distribution of the Work otherwise complies with
the conditions stated in this License.
5. Submission of Contributions. Unless You explicitly state otherwise,
any Contribution intentionally submitted for inclusion in the Work
by You to the Licensor shall be under the terms and conditions of
this License, without any additional terms or conditions.
Notwithstanding the above, nothing herein shall supersede or modify
the terms of any separate license agreement you may have executed
with Licensor regarding such Contributions.
6. Trademarks. This License does not grant permission to use the trade
names, trademarks, service marks, or product names of the Licensor,
except as required for reasonable and customary use in describing the
origin of the Work and reproducing the content of the NOTICE file.
7. Disclaimer of Warranty. Unless required by applicable law or
agreed to in writing, Licensor provides the Work (and each
Contributor provides its Contributions) on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
implied, including, without limitation, any warranties or conditions
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
PARTICULAR PURPOSE. You are solely responsible for determining the
appropriateness of using or redistributing the Work and assume any
risks associated with Your exercise of permissions under this License.
8. Limitation of Liability. In no event and under no legal theory,
whether in tort (including negligence), contract, or otherwise,
unless required by applicable law (such as deliberate and grossly
negligent acts) or agreed to in writing, shall any Contributor be
liable to You for damages, including any direct, indirect, special,
incidental, or consequential damages of any character arising as a
result of this License or out of the use or inability to use the
Work (including but not limited to damages for loss of goodwill,
work stoppage, computer failure or malfunction, or any and all
other commercial damages or losses), even if such Contributor
has been advised of the possibility of such damages.
9. Accepting Warranty or Additional Liability. While redistributing
the Work or Derivative Works thereof, You may choose to offer,
and charge a fee for, acceptance of support, warranty, indemnity,
or other liability obligations and/or rights consistent with this
License. However, in accepting such obligations, You may act only
on Your own behalf and on Your sole responsibility, not on behalf
of any other Contributor, and only if You agree to indemnify,
defend, and hold each Contributor harmless for any liability
incurred by, or claims asserted against, such Contributor by reason
of your accepting any such warranty or additional liability.
END OF TERMS AND CONDITIONS
APPENDIX: How to apply the Apache License to your work.
To apply the Apache License to your work, attach the following
boilerplate notice, with the fields enclosed by brackets "[]"
replaced with your own identifying information. (Don't include
the brackets!) The text should be enclosed in the appropriate
comment syntax for the file format. We also recommend that a
file or class name and description of purpose be included on the
same "printed page" as the copyright notice for easier
identification within third-party archives.
Copyright [yyyy] [name of copyright owner]
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.

@ -0,0 +1,62 @@
# ![bleve](docs/bleve.png) bleve
[![Build Status](https://travis-ci.org/blevesearch/bleve.svg?branch=master)](https://travis-ci.org/blevesearch/bleve) [![Coverage Status](https://coveralls.io/repos/blevesearch/bleve/badge.png?branch=master)](https://coveralls.io/r/blevesearch/bleve?branch=master) [![GoDoc](https://godoc.org/github.com/blevesearch/bleve?status.svg)](https://godoc.org/github.com/blevesearch/bleve)
[![Join the chat at https://gitter.im/blevesearch/bleve](https://badges.gitter.im/Join%20Chat.svg)](https://gitter.im/blevesearch/bleve?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge)
[![codebeat](https://codebeat.co/badges/38a7cbc9-9cf5-41c0-a315-0746178230f4)](https://codebeat.co/projects/github-com-blevesearch-bleve)
[![Go Report Card](https://goreportcard.com/badge/blevesearch/bleve)](https://goreportcard.com/report/blevesearch/bleve)
modern text indexing in go - [blevesearch.com](http://www.blevesearch.com/)
Try out bleve live by [searching the bleve website](http://www.blevesearch.com/search/?q=bleve).
## Features
* Index any go data structure (including JSON)
* Intelligent defaults backed up by powerful configuration
* Supported field types:
* Text, Numeric, Date
* Supported query types:
* Term, Phrase, Match, Match Phrase, Prefix
* Conjunction, Disjunction, Boolean
* Numeric Range, Date Range
* Simple query [syntax](http://www.blevesearch.com/docs/Query-String-Query/) for human entry
* tf-idf Scoring
* Search result match highlighting
* Supports Aggregating Facets:
* Terms Facet
* Numeric Range Facet
* Date Range Facet
## Discussion
Discuss usage and development of bleve in the [google group](https://groups.google.com/forum/#!forum/bleve).
## Indexing
message := struct{
Id string
From string
Body string
}{
Id: "example",
From: "marty.schoch@gmail.com",
Body: "bleve indexing is easy",
}
mapping := bleve.NewIndexMapping()
index, err := bleve.New("example.bleve", mapping)
if err != nil {
panic(err)
}
index.Index(message.Id, message)
## Querying
index, _ := bleve.Open("example.bleve")
query := bleve.NewQueryStringQuery("bleve")
searchRequest := bleve.NewSearchRequest(query)
searchResult, _ := index.Search(searchRequest)
## License
Apache License Version 2.0

@ -0,0 +1,46 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package simple
import (
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/analysis/token/lowercase"
"github.com/blevesearch/bleve/analysis/tokenizer/letter"
"github.com/blevesearch/bleve/registry"
)
const Name = "simple"
func AnalyzerConstructor(config map[string]interface{}, cache *registry.Cache) (*analysis.Analyzer, error) {
tokenizer, err := cache.TokenizerNamed(letter.Name)
if err != nil {
return nil, err
}
toLowerFilter, err := cache.TokenFilterNamed(lowercase.Name)
if err != nil {
return nil, err
}
rv := analysis.Analyzer{
Tokenizer: tokenizer,
TokenFilters: []analysis.TokenFilter{
toLowerFilter,
},
}
return &rv, nil
}
func init() {
registry.RegisterAnalyzer(Name, AnalyzerConstructor)
}

@ -0,0 +1,52 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package standard
import (
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/analysis/lang/en"
"github.com/blevesearch/bleve/analysis/token/lowercase"
"github.com/blevesearch/bleve/analysis/tokenizer/unicode"
"github.com/blevesearch/bleve/registry"
)
const Name = "standard"
func AnalyzerConstructor(config map[string]interface{}, cache *registry.Cache) (*analysis.Analyzer, error) {
tokenizer, err := cache.TokenizerNamed(unicode.Name)
if err != nil {
return nil, err
}
toLowerFilter, err := cache.TokenFilterNamed(lowercase.Name)
if err != nil {
return nil, err
}
stopEnFilter, err := cache.TokenFilterNamed(en.StopName)
if err != nil {
return nil, err
}
rv := analysis.Analyzer{
Tokenizer: tokenizer,
TokenFilters: []analysis.TokenFilter{
toLowerFilter,
stopEnFilter,
},
}
return &rv, nil
}
func init() {
registry.RegisterAnalyzer(Name, AnalyzerConstructor)
}

@ -0,0 +1,64 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package flexible
import (
"fmt"
"time"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/registry"
)
const Name = "flexiblego"
type DateTimeParser struct {
layouts []string
}
func New(layouts []string) *DateTimeParser {
return &DateTimeParser{
layouts: layouts,
}
}
func (p *DateTimeParser) ParseDateTime(input string) (time.Time, error) {
for _, layout := range p.layouts {
rv, err := time.Parse(layout, input)
if err == nil {
return rv, nil
}
}
return time.Time{}, analysis.ErrInvalidDateTime
}
func DateTimeParserConstructor(config map[string]interface{}, cache *registry.Cache) (analysis.DateTimeParser, error) {
layouts, ok := config["layouts"].([]interface{})
if !ok {
return nil, fmt.Errorf("must specify layouts")
}
var layoutStrs []string
for _, layout := range layouts {
layoutStr, ok := layout.(string)
if ok {
layoutStrs = append(layoutStrs, layoutStr)
}
}
return New(layoutStrs), nil
}
func init() {
registry.RegisterDateTimeParser(Name, DateTimeParserConstructor)
}

@ -0,0 +1,45 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package optional
import (
"time"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/analysis/datetime/flexible"
"github.com/blevesearch/bleve/registry"
)
const Name = "dateTimeOptional"
const rfc3339NoTimezone = "2006-01-02T15:04:05"
const rfc3339NoTimezoneNoT = "2006-01-02 15:04:05"
const rfc3339NoTime = "2006-01-02"
var layouts = []string{
time.RFC3339Nano,
time.RFC3339,
rfc3339NoTimezone,
rfc3339NoTimezoneNoT,
rfc3339NoTime,
}
func DateTimeParserConstructor(config map[string]interface{}, cache *registry.Cache) (analysis.DateTimeParser, error) {
return flexible.New(layouts), nil
}
func init() {
registry.RegisterDateTimeParser(Name, DateTimeParserConstructor)
}

@ -0,0 +1,111 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package analysis
// TokenLocation represents one occurrence of a term at a particular location in
// a field. Start, End and Position have the same meaning as in analysis.Token.
// Field and ArrayPositions identify the field value in the source document.
// See document.Field for details.
type TokenLocation struct {
Field string
ArrayPositions []uint64
Start int
End int
Position int
}
// TokenFreq represents all the occurrences of a term in all fields of a
// document.
type TokenFreq struct {
Term []byte
Locations []*TokenLocation
frequency int
}
func (tf *TokenFreq) Frequency() int {
return tf.frequency
}
// TokenFrequencies maps document terms to their combined frequencies from all
// fields.
type TokenFrequencies map[string]*TokenFreq
func (tfs TokenFrequencies) MergeAll(remoteField string, other TokenFrequencies) {
// walk the new token frequencies
for tfk, tf := range other {
// set the remoteField value in incoming token freqs
for _, l := range tf.Locations {
l.Field = remoteField
}
existingTf, exists := tfs[tfk]
if exists {
existingTf.Locations = append(existingTf.Locations, tf.Locations...)
existingTf.frequency = existingTf.frequency + tf.frequency
} else {
tfs[tfk] = &TokenFreq{
Term: tf.Term,
frequency: tf.frequency,
Locations: make([]*TokenLocation, len(tf.Locations)),
}
copy(tfs[tfk].Locations, tf.Locations)
}
}
}
func TokenFrequency(tokens TokenStream, arrayPositions []uint64, includeTermVectors bool) TokenFrequencies {
rv := make(map[string]*TokenFreq, len(tokens))
if includeTermVectors {
tls := make([]TokenLocation, len(tokens))
tlNext := 0
for _, token := range tokens {
tls[tlNext] = TokenLocation{
ArrayPositions: arrayPositions,
Start: token.Start,
End: token.End,
Position: token.Position,
}
curr, ok := rv[string(token.Term)]
if ok {
curr.Locations = append(curr.Locations, &tls[tlNext])
curr.frequency++
} else {
rv[string(token.Term)] = &TokenFreq{
Term: token.Term,
Locations: []*TokenLocation{&tls[tlNext]},
frequency: 1,
}
}
tlNext++
}
} else {
for _, token := range tokens {
curr, exists := rv[string(token.Term)]
if exists {
curr.frequency++
} else {
rv[string(token.Term)] = &TokenFreq{
Term: token.Term,
frequency: 1,
}
}
}
}
return rv
}

@ -0,0 +1,70 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
// Package en implements an analyzer with reasonable defaults for processing
// English text.
//
// It strips possessive suffixes ('s), transforms tokens to lower case,
// removes stopwords from a built-in list, and applies porter stemming.
//
// The built-in stopwords list is defined in EnglishStopWords.
package en
import (
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/registry"
"github.com/blevesearch/bleve/analysis/token/lowercase"
"github.com/blevesearch/bleve/analysis/token/porter"
"github.com/blevesearch/bleve/analysis/tokenizer/unicode"
)
const AnalyzerName = "en"
func AnalyzerConstructor(config map[string]interface{}, cache *registry.Cache) (*analysis.Analyzer, error) {
tokenizer, err := cache.TokenizerNamed(unicode.Name)
if err != nil {
return nil, err
}
possEnFilter, err := cache.TokenFilterNamed(PossessiveName)
if err != nil {
return nil, err
}
toLowerFilter, err := cache.TokenFilterNamed(lowercase.Name)
if err != nil {
return nil, err
}
stopEnFilter, err := cache.TokenFilterNamed(StopName)
if err != nil {
return nil, err
}
stemmerEnFilter, err := cache.TokenFilterNamed(porter.Name)
if err != nil {
return nil, err
}
rv := analysis.Analyzer{
Tokenizer: tokenizer,
TokenFilters: []analysis.TokenFilter{
possEnFilter,
toLowerFilter,
stopEnFilter,
stemmerEnFilter,
},
}
return &rv, nil
}
func init() {
registry.RegisterAnalyzer(AnalyzerName, AnalyzerConstructor)
}

@ -0,0 +1,67 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package en
import (
"unicode/utf8"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/registry"
)
// PossessiveName is the name PossessiveFilter is registered as
// in the bleve registry.
const PossessiveName = "possessive_en"
const rightSingleQuotationMark = '’'
const apostrophe = '\''
const fullWidthApostrophe = '''
const apostropheChars = rightSingleQuotationMark + apostrophe + fullWidthApostrophe
// PossessiveFilter implements a TokenFilter which
// strips the English possessive suffix ('s) from tokens.
// It handle a variety of apostrophe types, is case-insensitive
// and doesn't distinguish between possessive and contraction.
// (ie "She's So Rad" becomes "She So Rad")
type PossessiveFilter struct {
}
func NewPossessiveFilter() *PossessiveFilter {
return &PossessiveFilter{}
}
func (s *PossessiveFilter) Filter(input analysis.TokenStream) analysis.TokenStream {
for _, token := range input {
lastRune, lastRuneSize := utf8.DecodeLastRune(token.Term)
if lastRune == 's' || lastRune == 'S' {
nextLastRune, nextLastRuneSize := utf8.DecodeLastRune(token.Term[:len(token.Term)-lastRuneSize])
if nextLastRune == rightSingleQuotationMark ||
nextLastRune == apostrophe ||
nextLastRune == fullWidthApostrophe {
token.Term = token.Term[:len(token.Term)-lastRuneSize-nextLastRuneSize]
}
}
}
return input
}
func PossessiveFilterConstructor(config map[string]interface{}, cache *registry.Cache) (analysis.TokenFilter, error) {
return NewPossessiveFilter(), nil
}
func init() {
registry.RegisterTokenFilter(PossessiveName, PossessiveFilterConstructor)
}

@ -0,0 +1,33 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package en
import (
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/analysis/token/stop"
"github.com/blevesearch/bleve/registry"
)
func StopTokenFilterConstructor(config map[string]interface{}, cache *registry.Cache) (analysis.TokenFilter, error) {
tokenMap, err := cache.TokenMapNamed(StopName)
if err != nil {
return nil, err
}
return stop.NewStopTokensFilter(tokenMap), nil
}
func init() {
registry.RegisterTokenFilter(StopName, StopTokenFilterConstructor)
}

@ -0,0 +1,344 @@
package en
import (
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/registry"
)
const StopName = "stop_en"
// EnglishStopWords is the built-in list of stopwords used by the "stop_en" TokenFilter.
//
// this content was obtained from:
// lucene-4.7.2/analysis/common/src/resources/org/apache/lucene/analysis/snowball/
// ` was changed to ' to allow for literal string
var EnglishStopWords = []byte(` | From svn.tartarus.org/snowball/trunk/website/algorithms/english/stop.txt
| This file is distributed under the BSD License.
| See http://snowball.tartarus.org/license.php
| Also see http://www.opensource.org/licenses/bsd-license.html
| - Encoding was converted to UTF-8.
| - This notice was added.
|
| NOTE: To use this file with StopFilterFactory, you must specify format="snowball"
| An English stop word list. Comments begin with vertical bar. Each stop
| word is at the start of a line.
| Many of the forms below are quite rare (e.g. "yourselves") but included for
| completeness.
| PRONOUNS FORMS
| 1st person sing
i | subject, always in upper case of course
me | object
my | possessive adjective
| the possessive pronoun 'mine' is best suppressed, because of the
| sense of coal-mine etc.
myself | reflexive
| 1st person plural
we | subject
| us | object
| care is required here because US = United States. It is usually
| safe to remove it if it is in lower case.
our | possessive adjective
ours | possessive pronoun
ourselves | reflexive
| second person (archaic 'thou' forms not included)
you | subject and object
your | possessive adjective
yours | possessive pronoun
yourself | reflexive (singular)
yourselves | reflexive (plural)
| third person singular
he | subject
him | object
his | possessive adjective and pronoun
himself | reflexive
she | subject
her | object and possessive adjective
hers | possessive pronoun
herself | reflexive
it | subject and object
its | possessive adjective
itself | reflexive
| third person plural
they | subject
them | object
their | possessive adjective
theirs | possessive pronoun
themselves | reflexive
| other forms (demonstratives, interrogatives)
what
which
who
whom
this
that
these
those
| VERB FORMS (using F.R. Palmer's nomenclature)
| BE
am | 1st person, present
is | -s form (3rd person, present)
are | present
was | 1st person, past
were | past
be | infinitive
been | past participle
being | -ing form
| HAVE
have | simple
has | -s form
had | past
having | -ing form
| DO
do | simple
does | -s form
did | past
doing | -ing form
| The forms below are, I believe, best omitted, because of the significant
| homonym forms:
| He made a WILL
| old tin CAN
| merry month of MAY
| a smell of MUST
| fight the good fight with all thy MIGHT
| would, could, should, ought might however be included
| | AUXILIARIES
| | WILL
|will
would
| | SHALL
|shall
should
| | CAN
|can
could
| | MAY
|may
|might
| | MUST
|must
| | OUGHT
ought
| COMPOUND FORMS, increasingly encountered nowadays in 'formal' writing
| pronoun + verb
i'm
you're
he's
she's
it's
we're
they're
i've
you've
we've
they've
i'd
you'd
he'd
she'd
we'd
they'd
i'll
you'll
he'll
she'll
we'll
they'll
| verb + negation
isn't
aren't
wasn't
weren't
hasn't
haven't
hadn't
doesn't
don't
didn't
| auxiliary + negation
won't
wouldn't
shan't
shouldn't
can't
cannot
couldn't
mustn't
| miscellaneous forms
let's
that's
who's
what's
here's
there's
when's
where's
why's
how's
| rarer forms
| daren't needn't
| doubtful forms
| oughtn't mightn't
| ARTICLES
a
an
the
| THE REST (Overlap among prepositions, conjunctions, adverbs etc is so
| high, that classification is pointless.)
and
but
if
or
because
as
until
while
of
at
by
for
with
about
against
between
into
through
during
before
after
above
below
to
from
up
down
in
out
on
off
over
under
again
further
then
once
here
there
when
where
why
how
all
any
both
each
few
more
most
other
some
such
no
nor
not
only
own
same
so
than
too
very
| Just for the record, the following words are among the commonest in English
| one
| every
| least
| less
| many
| now
| ever
| never
| say
| says
| said
| also
| get
| go
| goes
| just
| made
| make
| put
| see
| seen
| whether
| like
| well
| back
| even
| still
| way
| take
| since
| another
| however
| two
| three
| four
| five
| first
| second
| new
| old
| high
| long
`)
func TokenMapConstructor(config map[string]interface{}, cache *registry.Cache) (analysis.TokenMap, error) {
rv := analysis.NewTokenMap()
err := rv.LoadBytes(EnglishStopWords)
return rv, err
}
func init() {
registry.RegisterTokenMap(StopName, TokenMapConstructor)
}

@ -0,0 +1,7 @@
# full line comment
marty
steve # trailing comment
| different format of comment
dustin
siri | different style trailing comment
multiple words with different whitespace

@ -0,0 +1,105 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
// Package lowercase implements a TokenFilter which converts
// tokens to lower case according to unicode rules.
package lowercase
import (
"bytes"
"unicode"
"unicode/utf8"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/registry"
)
// Name is the name used to register LowerCaseFilter in the bleve registry
const Name = "to_lower"
type LowerCaseFilter struct {
}
func NewLowerCaseFilter() *LowerCaseFilter {
return &LowerCaseFilter{}
}
func (f *LowerCaseFilter) Filter(input analysis.TokenStream) analysis.TokenStream {
for _, token := range input {
token.Term = toLowerDeferredCopy(token.Term)
}
return input
}
func LowerCaseFilterConstructor(config map[string]interface{}, cache *registry.Cache) (analysis.TokenFilter, error) {
return NewLowerCaseFilter(), nil
}
func init() {
registry.RegisterTokenFilter(Name, LowerCaseFilterConstructor)
}
// toLowerDeferredCopy will function exactly like
// bytes.ToLower() only it will reuse (overwrite)
// the original byte array when possible
// NOTE: because its possible that the lower-case
// form of a rune has a different utf-8 encoded
// length, in these cases a new byte array is allocated
func toLowerDeferredCopy(s []byte) []byte {
j := 0
for i := 0; i < len(s); {
wid := 1
r := rune(s[i])
if r >= utf8.RuneSelf {
r, wid = utf8.DecodeRune(s[i:])
}
l := unicode.ToLower(r)
// If the rune is already lowercased, just move to the
// next rune.
if l == r {
i += wid
j += wid
continue
}
// Handles the Unicode edge-case where the last
// rune in a word on the greek Σ needs to be converted
// differently.
if l == 'σ' && i+2 == len(s) {
l = 'ς'
}
lwid := utf8.RuneLen(l)
if lwid > wid {
// utf-8 encoded replacement is wider
// for now, punt and defer
// to bytes.ToLower() for the remainder
// only known to happen with chars
// Rune Ⱥ(570) width 2 - Lower ⱥ(11365) width 3
// Rune Ⱦ(574) width 2 - Lower ⱦ(11366) width 3
rest := bytes.ToLower(s[i:])
rv := make([]byte, j+len(rest))
copy(rv[:j], s[:j])
copy(rv[j:], rest)
return rv
} else {
utf8.EncodeRune(s[j:], l)
}
i += wid
j += lwid
}
return s[:j]
}

@ -0,0 +1,53 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package porter
import (
"bytes"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/registry"
"github.com/blevesearch/go-porterstemmer"
)
const Name = "stemmer_porter"
type PorterStemmer struct {
}
func NewPorterStemmer() *PorterStemmer {
return &PorterStemmer{}
}
func (s *PorterStemmer) Filter(input analysis.TokenStream) analysis.TokenStream {
for _, token := range input {
// if it is not a protected keyword, stem it
if !token.KeyWord {
termRunes := bytes.Runes(token.Term)
stemmedRunes := porterstemmer.StemWithoutLowerCasing(termRunes)
token.Term = analysis.BuildTermFromRunes(stemmedRunes)
}
}
return input
}
func PorterStemmerConstructor(config map[string]interface{}, cache *registry.Cache) (analysis.TokenFilter, error) {
return NewPorterStemmer(), nil
}
func init() {
registry.RegisterTokenFilter(Name, PorterStemmerConstructor)
}

@ -0,0 +1,70 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
// Package stop implements a TokenFilter removing tokens found in
// a TokenMap.
//
// It constructor takes the following arguments:
//
// "stop_token_map" (string): the name of the token map identifying tokens to
// remove.
package stop
import (
"fmt"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/registry"
)
const Name = "stop_tokens"
type StopTokensFilter struct {
stopTokens analysis.TokenMap
}
func NewStopTokensFilter(stopTokens analysis.TokenMap) *StopTokensFilter {
return &StopTokensFilter{
stopTokens: stopTokens,
}
}
func (f *StopTokensFilter) Filter(input analysis.TokenStream) analysis.TokenStream {
j := 0
for _, token := range input {
_, isStopToken := f.stopTokens[string(token.Term)]
if !isStopToken {
input[j] = token
j++
}
}
return input[:j]
}
func StopTokensFilterConstructor(config map[string]interface{}, cache *registry.Cache) (analysis.TokenFilter, error) {
stopTokenMapName, ok := config["stop_token_map"].(string)
if !ok {
return nil, fmt.Errorf("must specify stop_token_map")
}
stopTokenMap, err := cache.TokenMapNamed(stopTokenMapName)
if err != nil {
return nil, fmt.Errorf("error building stop words filter: %v", err)
}
return NewStopTokensFilter(stopTokenMap), nil
}
func init() {
registry.RegisterTokenFilter(Name, StopTokensFilterConstructor)
}

@ -0,0 +1,76 @@
// Copyright (c) 2016 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package character
import (
"unicode/utf8"
"github.com/blevesearch/bleve/analysis"
)
type IsTokenRune func(r rune) bool
type CharacterTokenizer struct {
isTokenRun IsTokenRune
}
func NewCharacterTokenizer(f IsTokenRune) *CharacterTokenizer {
return &CharacterTokenizer{
isTokenRun: f,
}
}
func (c *CharacterTokenizer) Tokenize(input []byte) analysis.TokenStream {
rv := make(analysis.TokenStream, 0, 1024)
offset := 0
start := 0
end := 0
count := 0
for currRune, size := utf8.DecodeRune(input[offset:]); currRune != utf8.RuneError; currRune, size = utf8.DecodeRune(input[offset:]) {
isToken := c.isTokenRun(currRune)
if isToken {
end = offset + size
} else {
if end-start > 0 {
// build token
rv = append(rv, &analysis.Token{
Term: input[start:end],
Start: start,
End: end,
Position: count + 1,
Type: analysis.AlphaNumeric,
})
count++
}
start = offset + size
end = start
}
offset += size
}
// if we ended in the middle of a token, finish it
if end-start > 0 {
// build token
rv = append(rv, &analysis.Token{
Term: input[start:end],
Start: start,
End: end,
Position: count + 1,
Type: analysis.AlphaNumeric,
})
}
return rv
}

@ -0,0 +1,33 @@
// Copyright (c) 2016 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package letter
import (
"unicode"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/analysis/tokenizer/character"
"github.com/blevesearch/bleve/registry"
)
const Name = "letter"
func TokenizerConstructor(config map[string]interface{}, cache *registry.Cache) (analysis.Tokenizer, error) {
return character.NewCharacterTokenizer(unicode.IsLetter), nil
}
func init() {
registry.RegisterTokenizer(Name, TokenizerConstructor)
}

@ -0,0 +1,131 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package unicode
import (
"github.com/blevesearch/segment"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/registry"
)
const Name = "unicode"
type UnicodeTokenizer struct {
}
func NewUnicodeTokenizer() *UnicodeTokenizer {
return &UnicodeTokenizer{}
}
func (rt *UnicodeTokenizer) Tokenize(input []byte) analysis.TokenStream {
rvx := make([]analysis.TokenStream, 0, 10) // When rv gets full, append to rvx.
rv := make(analysis.TokenStream, 0, 1)
ta := []analysis.Token(nil)
taNext := 0
segmenter := segment.NewWordSegmenterDirect(input)
start := 0
pos := 1
guessRemaining := func(end int) int {
avgSegmentLen := end / (len(rv) + 1)
if avgSegmentLen < 1 {
avgSegmentLen = 1
}
remainingLen := len(input) - end
return remainingLen / avgSegmentLen
}
for segmenter.Segment() {
segmentBytes := segmenter.Bytes()
end := start + len(segmentBytes)
if segmenter.Type() != segment.None {
if taNext >= len(ta) {
remainingSegments := guessRemaining(end)
if remainingSegments > 1000 {
remainingSegments = 1000
}
if remainingSegments < 1 {
remainingSegments = 1
}
ta = make([]analysis.Token, remainingSegments)
taNext = 0
}
token := &ta[taNext]
taNext++
token.Term = segmentBytes
token.Start = start
token.End = end
token.Position = pos
token.Type = convertType(segmenter.Type())
if len(rv) >= cap(rv) { // When rv is full, save it into rvx.
rvx = append(rvx, rv)
rvCap := cap(rv) * 2
if rvCap > 256 {
rvCap = 256
}
rv = make(analysis.TokenStream, 0, rvCap) // Next rv cap is bigger.
}
rv = append(rv, token)
pos++
}
start = end
}
if len(rvx) > 0 {
n := len(rv)
for _, r := range rvx {
n += len(r)
}
rall := make(analysis.TokenStream, 0, n)
for _, r := range rvx {
rall = append(rall, r...)
}
return append(rall, rv...)
}
return rv
}
func UnicodeTokenizerConstructor(config map[string]interface{}, cache *registry.Cache) (analysis.Tokenizer, error) {
return NewUnicodeTokenizer(), nil
}
func init() {
registry.RegisterTokenizer(Name, UnicodeTokenizerConstructor)
}
func convertType(segmentWordType int) analysis.TokenType {
switch segmentWordType {
case segment.Ideo:
return analysis.Ideographic
case segment.Kana:
return analysis.Ideographic
case segment.Number:
return analysis.Numeric
}
return analysis.AlphaNumeric
}

@ -0,0 +1,76 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package analysis
import (
"bufio"
"bytes"
"io"
"io/ioutil"
"strings"
)
type TokenMap map[string]bool
func NewTokenMap() TokenMap {
return make(TokenMap, 0)
}
// LoadFile reads in a list of tokens from a text file,
// one per line.
// Comments are supported using `#` or `|`
func (t TokenMap) LoadFile(filename string) error {
data, err := ioutil.ReadFile(filename)
if err != nil {
return err
}
return t.LoadBytes(data)
}
// LoadBytes reads in a list of tokens from memory,
// one per line.
// Comments are supported using `#` or `|`
func (t TokenMap) LoadBytes(data []byte) error {
bytesReader := bytes.NewReader(data)
bufioReader := bufio.NewReader(bytesReader)
line, err := bufioReader.ReadString('\n')
for err == nil {
t.LoadLine(line)
line, err = bufioReader.ReadString('\n')
}
// if the err was EOF we still need to process the last value
if err == io.EOF {
t.LoadLine(line)
return nil
}
return err
}
func (t TokenMap) LoadLine(line string) {
// find the start of a comment, if any
startComment := strings.IndexAny(line, "#|")
if startComment >= 0 {
line = line[:startComment]
}
tokens := strings.Fields(line)
for _, token := range tokens {
t.AddToken(token)
}
}
func (t TokenMap) AddToken(token string) {
t[token] = true
}

@ -0,0 +1,103 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package analysis
import (
"fmt"
"time"
)
type CharFilter interface {
Filter([]byte) []byte
}
type TokenType int
const (
AlphaNumeric TokenType = iota
Ideographic
Numeric
DateTime
Shingle
Single
Double
Boolean
)
// Token represents one occurrence of a term at a particular location in a
// field.
type Token struct {
// Start specifies the byte offset of the beginning of the term in the
// field.
Start int `json:"start"`
// End specifies the byte offset of the end of the term in the field.
End int `json:"end"`
Term []byte `json:"term"`
// Position specifies the 1-based index of the token in the sequence of
// occurrences of its term in the field.
Position int `json:"position"`
Type TokenType `json:"type"`
KeyWord bool `json:"keyword"`
}
func (t *Token) String() string {
return fmt.Sprintf("Start: %d End: %d Position: %d Token: %s Type: %d", t.Start, t.End, t.Position, string(t.Term), t.Type)
}
type TokenStream []*Token
// A Tokenizer splits an input string into tokens, the usual behaviour being to
// map words to tokens.
type Tokenizer interface {
Tokenize([]byte) TokenStream
}
// A TokenFilter adds, transforms or removes tokens from a token stream.
type TokenFilter interface {
Filter(TokenStream) TokenStream
}
type Analyzer struct {
CharFilters []CharFilter
Tokenizer Tokenizer
TokenFilters []TokenFilter
}
func (a *Analyzer) Analyze(input []byte) TokenStream {
if a.CharFilters != nil {
for _, cf := range a.CharFilters {
input = cf.Filter(input)
}
}
tokens := a.Tokenizer.Tokenize(input)
if a.TokenFilters != nil {
for _, tf := range a.TokenFilters {
tokens = tf.Filter(tokens)
}
}
return tokens
}
var ErrInvalidDateTime = fmt.Errorf("unable to parse datetime with any of the layouts")
type DateTimeParser interface {
ParseDateTime(string) (time.Time, error)
}
type ByteArrayConverter interface {
Convert([]byte) (interface{}, error)
}

@ -0,0 +1,92 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package analysis
import (
"bytes"
"unicode/utf8"
)
func DeleteRune(in []rune, pos int) []rune {
if pos >= len(in) {
return in
}
copy(in[pos:], in[pos+1:])
return in[:len(in)-1]
}
func InsertRune(in []rune, pos int, r rune) []rune {
// create a new slice 1 rune larger
rv := make([]rune, len(in)+1)
// copy the characters before the insert pos
copy(rv[0:pos], in[0:pos])
// set the inserted rune
rv[pos] = r
// copy the characters after the insert pos
copy(rv[pos+1:], in[pos:])
return rv
}
// BuildTermFromRunesOptimistic will build a term from the provided runes
// AND optimistically attempt to encode into the provided buffer
// if at any point it appears the buffer is too small, a new buffer is
// allocated and that is used instead
// this should be used in cases where frequently the new term is the same
// length or shorter than the original term (in number of bytes)
func BuildTermFromRunesOptimistic(buf []byte, runes []rune) []byte {
rv := buf
used := 0
for _, r := range runes {
nextLen := utf8.RuneLen(r)
if used+nextLen > len(rv) {
// alloc new buf
buf = make([]byte, len(runes)*utf8.UTFMax)
// copy work we've already done
copy(buf, rv[:used])
rv = buf
}
written := utf8.EncodeRune(rv[used:], r)
used += written
}
return rv[:used]
}
func BuildTermFromRunes(runes []rune) []byte {
return BuildTermFromRunesOptimistic(make([]byte, len(runes)*utf8.UTFMax), runes)
}
func TruncateRunes(input []byte, num int) []byte {
runes := bytes.Runes(input)
runes = runes[:len(runes)-num]
out := BuildTermFromRunes(runes)
return out
}
func RunesEndsWith(input []rune, suffix string) bool {
inputLen := len(input)
suffixRunes := []rune(suffix)
suffixLen := len(suffixRunes)
if suffixLen > inputLen {
return false
}
for i := suffixLen - 1; i >= 0; i-- {
if input[inputLen-(suffixLen-i)] != suffixRunes[i] {
return false
}
}
return true
}

@ -0,0 +1,88 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package bleve
import (
"expvar"
"io/ioutil"
"log"
"time"
"github.com/blevesearch/bleve/index"
"github.com/blevesearch/bleve/index/store/gtreap"
"github.com/blevesearch/bleve/index/upsidedown"
"github.com/blevesearch/bleve/registry"
"github.com/blevesearch/bleve/search/highlight/highlighter/html"
)
var bleveExpVar = expvar.NewMap("bleve")
type configuration struct {
Cache *registry.Cache
DefaultHighlighter string
DefaultKVStore string
DefaultMemKVStore string
DefaultIndexType string
SlowSearchLogThreshold time.Duration
analysisQueue *index.AnalysisQueue
}
func (c *configuration) SetAnalysisQueueSize(n int) {
c.analysisQueue = index.NewAnalysisQueue(n)
}
func newConfiguration() *configuration {
return &configuration{
Cache: registry.NewCache(),
analysisQueue: index.NewAnalysisQueue(4),
}
}
// Config contains library level configuration
var Config *configuration
func init() {
bootStart := time.Now()
// build the default configuration
Config = newConfiguration()
// set the default highlighter
Config.DefaultHighlighter = html.Name
// default kv store
Config.DefaultKVStore = ""
// default mem only kv store
Config.DefaultMemKVStore = gtreap.Name
// default index
Config.DefaultIndexType = upsidedown.Name
bootDuration := time.Since(bootStart)
bleveExpVar.Add("bootDuration", int64(bootDuration))
indexStats = NewIndexStats()
bleveExpVar.Set("indexes", indexStats)
initDisk()
}
var logger = log.New(ioutil.Discard, "bleve", log.LstdFlags)
// SetLog sets the logger used for logging
// by default log messages are sent to ioutil.Discard
func SetLog(l *log.Logger) {
logger = l
}

@ -0,0 +1,23 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
// +build appengine appenginevm
package bleve
// in the appengine environment we cannot support disk based indexes
// so we do no extra configuration in this method
func initDisk() {
}

@ -0,0 +1,25 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
// +build !appengine,!appenginevm
package bleve
import "github.com/blevesearch/bleve/index/store/boltdb"
// in normal environments we configure boltdb as the default storage
func initDisk() {
// default kv store
Config.DefaultKVStore = boltdb.Name
}

@ -0,0 +1,38 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
/*
Package bleve is a library for indexing and searching text.
Example Opening New Index, Indexing Data
message := struct{
Id: "example"
From: "marty.schoch@gmail.com",
Body: "bleve indexing is easy",
}
mapping := bleve.NewIndexMapping()
index, _ := bleve.New("example.bleve", mapping)
index.Index(message.Id, message)
Example Opening Existing Index, Searching Data
index, _ := bleve.Open("example.bleve")
query := bleve.NewQueryStringQuery("bleve")
searchRequest := bleve.NewSearchRequest(query)
searchResult, _ := index.Search(searchRequest)
*/
package bleve

@ -0,0 +1,75 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package document
import "fmt"
type Document struct {
ID string `json:"id"`
Fields []Field `json:"fields"`
CompositeFields []*CompositeField
Number uint64 `json:"-"`
}
func NewDocument(id string) *Document {
return &Document{
ID: id,
Fields: make([]Field, 0),
CompositeFields: make([]*CompositeField, 0),
}
}
func (d *Document) AddField(f Field) *Document {
switch f := f.(type) {
case *CompositeField:
d.CompositeFields = append(d.CompositeFields, f)
default:
d.Fields = append(d.Fields, f)
}
return d
}
func (d *Document) GoString() string {
fields := ""
for i, field := range d.Fields {
if i != 0 {
fields += ", "
}
fields += fmt.Sprintf("%#v", field)
}
compositeFields := ""
for i, field := range d.CompositeFields {
if i != 0 {
compositeFields += ", "
}
compositeFields += fmt.Sprintf("%#v", field)
}
return fmt.Sprintf("&document.Document{ID:%s, Fields: %s, CompositeFields: %s}", d.ID, fields, compositeFields)
}
func (d *Document) NumPlainTextBytes() uint64 {
rv := uint64(0)
for _, field := range d.Fields {
rv += field.NumPlainTextBytes()
}
for _, compositeField := range d.CompositeFields {
for _, field := range d.Fields {
if compositeField.includesField(field.Name()) {
rv += field.NumPlainTextBytes()
}
}
}
return rv
}

@ -0,0 +1,39 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package document
import (
"github.com/blevesearch/bleve/analysis"
)
type Field interface {
// Name returns the path of the field from the root DocumentMapping.
// A root field path is "field", a subdocument field is "parent.field".
Name() string
// ArrayPositions returns the intermediate document and field indices
// required to resolve the field value in the document. For example, if the
// field path is "doc1.doc2.field" where doc1 and doc2 are slices or
// arrays, ArrayPositions returns 2 indices used to resolve "doc2" value in
// "doc1", then "field" in "doc2".
ArrayPositions() []uint64
Options() IndexingOptions
Analyze() (int, analysis.TokenFrequencies)
Value() []byte
// NumPlainTextBytes should return the number of plain text bytes
// that this field represents - this is a common metric for tracking
// the rate of indexing
NumPlainTextBytes() uint64
}

@ -0,0 +1,107 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package document
import (
"fmt"
"github.com/blevesearch/bleve/analysis"
)
const DefaultBooleanIndexingOptions = StoreField | IndexField
type BooleanField struct {
name string
arrayPositions []uint64
options IndexingOptions
value []byte
numPlainTextBytes uint64
}
func (b *BooleanField) Name() string {
return b.name
}
func (b *BooleanField) ArrayPositions() []uint64 {
return b.arrayPositions
}
func (b *BooleanField) Options() IndexingOptions {
return b.options
}
func (b *BooleanField) Analyze() (int, analysis.TokenFrequencies) {
tokens := make(analysis.TokenStream, 0)
tokens = append(tokens, &analysis.Token{
Start: 0,
End: len(b.value),
Term: b.value,
Position: 1,
Type: analysis.Boolean,
})
fieldLength := len(tokens)
tokenFreqs := analysis.TokenFrequency(tokens, b.arrayPositions, b.options.IncludeTermVectors())
return fieldLength, tokenFreqs
}
func (b *BooleanField) Value() []byte {
return b.value
}
func (b *BooleanField) Boolean() (bool, error) {
if len(b.value) == 1 {
return b.value[0] == 'T', nil
}
return false, fmt.Errorf("boolean field has %d bytes", len(b.value))
}
func (b *BooleanField) GoString() string {
return fmt.Sprintf("&document.BooleanField{Name:%s, Options: %s, Value: %s}", b.name, b.options, b.value)
}
func (b *BooleanField) NumPlainTextBytes() uint64 {
return b.numPlainTextBytes
}
func NewBooleanFieldFromBytes(name string, arrayPositions []uint64, value []byte) *BooleanField {
return &BooleanField{
name: name,
arrayPositions: arrayPositions,
value: value,
options: DefaultNumericIndexingOptions,
numPlainTextBytes: uint64(len(value)),
}
}
func NewBooleanField(name string, arrayPositions []uint64, b bool) *BooleanField {
return NewBooleanFieldWithIndexingOptions(name, arrayPositions, b, DefaultNumericIndexingOptions)
}
func NewBooleanFieldWithIndexingOptions(name string, arrayPositions []uint64, b bool, options IndexingOptions) *BooleanField {
numPlainTextBytes := 5
v := []byte("F")
if b {
numPlainTextBytes = 4
v = []byte("T")
}
return &BooleanField{
name: name,
arrayPositions: arrayPositions,
value: v,
options: options,
numPlainTextBytes: uint64(numPlainTextBytes),
}
}

@ -0,0 +1,99 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package document
import (
"github.com/blevesearch/bleve/analysis"
)
const DefaultCompositeIndexingOptions = IndexField
type CompositeField struct {
name string
includedFields map[string]bool
excludedFields map[string]bool
defaultInclude bool
options IndexingOptions
totalLength int
compositeFrequencies analysis.TokenFrequencies
}
func NewCompositeField(name string, defaultInclude bool, include []string, exclude []string) *CompositeField {
return NewCompositeFieldWithIndexingOptions(name, defaultInclude, include, exclude, DefaultCompositeIndexingOptions)
}
func NewCompositeFieldWithIndexingOptions(name string, defaultInclude bool, include []string, exclude []string, options IndexingOptions) *CompositeField {
rv := &CompositeField{
name: name,
options: options,
defaultInclude: defaultInclude,
includedFields: make(map[string]bool, len(include)),
excludedFields: make(map[string]bool, len(exclude)),
compositeFrequencies: make(analysis.TokenFrequencies),
}
for _, i := range include {
rv.includedFields[i] = true
}
for _, e := range exclude {
rv.excludedFields[e] = true
}
return rv
}
func (c *CompositeField) Name() string {
return c.name
}
func (c *CompositeField) ArrayPositions() []uint64 {
return []uint64{}
}
func (c *CompositeField) Options() IndexingOptions {
return c.options
}
func (c *CompositeField) Analyze() (int, analysis.TokenFrequencies) {
return c.totalLength, c.compositeFrequencies
}
func (c *CompositeField) Value() []byte {
return []byte{}
}
func (c *CompositeField) NumPlainTextBytes() uint64 {
return 0
}
func (c *CompositeField) includesField(field string) bool {
shouldInclude := c.defaultInclude
_, fieldShouldBeIncluded := c.includedFields[field]
if fieldShouldBeIncluded {
shouldInclude = true
}
_, fieldShouldBeExcluded := c.excludedFields[field]
if fieldShouldBeExcluded {
shouldInclude = false
}
return shouldInclude
}
func (c *CompositeField) Compose(field string, length int, freq analysis.TokenFrequencies) {
if c.includesField(field) {
c.totalLength += length
c.compositeFrequencies.MergeAll(field, freq)
}
}

@ -0,0 +1,144 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package document
import (
"fmt"
"math"
"time"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/numeric"
)
const DefaultDateTimeIndexingOptions = StoreField | IndexField
const DefaultDateTimePrecisionStep uint = 4
var MinTimeRepresentable = time.Unix(0, math.MinInt64)
var MaxTimeRepresentable = time.Unix(0, math.MaxInt64)
type DateTimeField struct {
name string
arrayPositions []uint64
options IndexingOptions
value numeric.PrefixCoded
numPlainTextBytes uint64
}
func (n *DateTimeField) Name() string {
return n.name
}
func (n *DateTimeField) ArrayPositions() []uint64 {
return n.arrayPositions
}
func (n *DateTimeField) Options() IndexingOptions {
return n.options
}
func (n *DateTimeField) Analyze() (int, analysis.TokenFrequencies) {
tokens := make(analysis.TokenStream, 0)
tokens = append(tokens, &analysis.Token{
Start: 0,
End: len(n.value),
Term: n.value,
Position: 1,
Type: analysis.DateTime,
})
original, err := n.value.Int64()
if err == nil {
shift := DefaultDateTimePrecisionStep
for shift < 64 {
shiftEncoded, err := numeric.NewPrefixCodedInt64(original, shift)
if err != nil {
break
}
token := analysis.Token{
Start: 0,
End: len(shiftEncoded),
Term: shiftEncoded,
Position: 1,
Type: analysis.DateTime,
}
tokens = append(tokens, &token)
shift += DefaultDateTimePrecisionStep
}
}
fieldLength := len(tokens)
tokenFreqs := analysis.TokenFrequency(tokens, n.arrayPositions, n.options.IncludeTermVectors())
return fieldLength, tokenFreqs
}
func (n *DateTimeField) Value() []byte {
return n.value
}
func (n *DateTimeField) DateTime() (time.Time, error) {
i64, err := n.value.Int64()
if err != nil {
return time.Time{}, err
}
return time.Unix(0, i64).UTC(), nil
}
func (n *DateTimeField) GoString() string {
return fmt.Sprintf("&document.DateField{Name:%s, Options: %s, Value: %s}", n.name, n.options, n.value)
}
func (n *DateTimeField) NumPlainTextBytes() uint64 {
return n.numPlainTextBytes
}
func NewDateTimeFieldFromBytes(name string, arrayPositions []uint64, value []byte) *DateTimeField {
return &DateTimeField{
name: name,
arrayPositions: arrayPositions,
value: value,
options: DefaultDateTimeIndexingOptions,
numPlainTextBytes: uint64(len(value)),
}
}
func NewDateTimeField(name string, arrayPositions []uint64, dt time.Time) (*DateTimeField, error) {
return NewDateTimeFieldWithIndexingOptions(name, arrayPositions, dt, DefaultDateTimeIndexingOptions)
}
func NewDateTimeFieldWithIndexingOptions(name string, arrayPositions []uint64, dt time.Time, options IndexingOptions) (*DateTimeField, error) {
if canRepresent(dt) {
dtInt64 := dt.UnixNano()
prefixCoded := numeric.MustNewPrefixCodedInt64(dtInt64, 0)
return &DateTimeField{
name: name,
arrayPositions: arrayPositions,
value: prefixCoded,
options: options,
// not correct, just a place holder until we revisit how fields are
// represented and can fix this better
numPlainTextBytes: uint64(8),
}, nil
}
return nil, fmt.Errorf("cannot represent %s in this type", dt)
}
func canRepresent(dt time.Time) bool {
if dt.Before(MinTimeRepresentable) || dt.After(MaxTimeRepresentable) {
return false
}
return true
}

@ -0,0 +1,130 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package document
import (
"fmt"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/numeric"
)
const DefaultNumericIndexingOptions = StoreField | IndexField
const DefaultPrecisionStep uint = 4
type NumericField struct {
name string
arrayPositions []uint64
options IndexingOptions
value numeric.PrefixCoded
numPlainTextBytes uint64
}
func (n *NumericField) Name() string {
return n.name
}
func (n *NumericField) ArrayPositions() []uint64 {
return n.arrayPositions
}
func (n *NumericField) Options() IndexingOptions {
return n.options
}
func (n *NumericField) Analyze() (int, analysis.TokenFrequencies) {
tokens := make(analysis.TokenStream, 0)
tokens = append(tokens, &analysis.Token{
Start: 0,
End: len(n.value),
Term: n.value,
Position: 1,
Type: analysis.Numeric,
})
original, err := n.value.Int64()
if err == nil {
shift := DefaultPrecisionStep
for shift < 64 {
shiftEncoded, err := numeric.NewPrefixCodedInt64(original, shift)
if err != nil {
break
}
token := analysis.Token{
Start: 0,
End: len(shiftEncoded),
Term: shiftEncoded,
Position: 1,
Type: analysis.Numeric,
}
tokens = append(tokens, &token)
shift += DefaultPrecisionStep
}
}
fieldLength := len(tokens)
tokenFreqs := analysis.TokenFrequency(tokens, n.arrayPositions, n.options.IncludeTermVectors())
return fieldLength, tokenFreqs
}
func (n *NumericField) Value() []byte {
return n.value
}
func (n *NumericField) Number() (float64, error) {
i64, err := n.value.Int64()
if err != nil {
return 0.0, err
}
return numeric.Int64ToFloat64(i64), nil
}
func (n *NumericField) GoString() string {
return fmt.Sprintf("&document.NumericField{Name:%s, Options: %s, Value: %s}", n.name, n.options, n.value)
}
func (n *NumericField) NumPlainTextBytes() uint64 {
return n.numPlainTextBytes
}
func NewNumericFieldFromBytes(name string, arrayPositions []uint64, value []byte) *NumericField {
return &NumericField{
name: name,
arrayPositions: arrayPositions,
value: value,
options: DefaultNumericIndexingOptions,
numPlainTextBytes: uint64(len(value)),
}
}
func NewNumericField(name string, arrayPositions []uint64, number float64) *NumericField {
return NewNumericFieldWithIndexingOptions(name, arrayPositions, number, DefaultNumericIndexingOptions)
}
func NewNumericFieldWithIndexingOptions(name string, arrayPositions []uint64, number float64, options IndexingOptions) *NumericField {
numberInt64 := numeric.Float64ToInt64(number)
prefixCoded := numeric.MustNewPrefixCodedInt64(numberInt64, 0)
return &NumericField{
name: name,
arrayPositions: arrayPositions,
value: prefixCoded,
options: options,
// not correct, just a place holder until we revisit how fields are
// represented and can fix this better
numPlainTextBytes: uint64(8),
}
}

@ -0,0 +1,119 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package document
import (
"fmt"
"github.com/blevesearch/bleve/analysis"
)
const DefaultTextIndexingOptions = IndexField
type TextField struct {
name string
arrayPositions []uint64
options IndexingOptions
analyzer *analysis.Analyzer
value []byte
numPlainTextBytes uint64
}
func (t *TextField) Name() string {
return t.name
}
func (t *TextField) ArrayPositions() []uint64 {
return t.arrayPositions
}
func (t *TextField) Options() IndexingOptions {
return t.options
}
func (t *TextField) Analyze() (int, analysis.TokenFrequencies) {
var tokens analysis.TokenStream
if t.analyzer != nil {
bytesToAnalyze := t.Value()
if t.options.IsStored() {
// need to copy
bytesCopied := make([]byte, len(bytesToAnalyze))
copy(bytesCopied, bytesToAnalyze)
bytesToAnalyze = bytesCopied
}
tokens = t.analyzer.Analyze(bytesToAnalyze)
} else {
tokens = analysis.TokenStream{
&analysis.Token{
Start: 0,
End: len(t.value),
Term: t.value,
Position: 1,
Type: analysis.AlphaNumeric,
},
}
}
fieldLength := len(tokens) // number of tokens in this doc field
tokenFreqs := analysis.TokenFrequency(tokens, t.arrayPositions, t.options.IncludeTermVectors())
return fieldLength, tokenFreqs
}
func (t *TextField) Value() []byte {
return t.value
}
func (t *TextField) GoString() string {
return fmt.Sprintf("&document.TextField{Name:%s, Options: %s, Analyzer: %v, Value: %s, ArrayPositions: %v}", t.name, t.options, t.analyzer, t.value, t.arrayPositions)
}
func (t *TextField) NumPlainTextBytes() uint64 {
return t.numPlainTextBytes
}
func NewTextField(name string, arrayPositions []uint64, value []byte) *TextField {
return NewTextFieldWithIndexingOptions(name, arrayPositions, value, DefaultTextIndexingOptions)
}
func NewTextFieldWithIndexingOptions(name string, arrayPositions []uint64, value []byte, options IndexingOptions) *TextField {
return &TextField{
name: name,
arrayPositions: arrayPositions,
options: options,
value: value,
numPlainTextBytes: uint64(len(value)),
}
}
func NewTextFieldWithAnalyzer(name string, arrayPositions []uint64, value []byte, analyzer *analysis.Analyzer) *TextField {
return &TextField{
name: name,
arrayPositions: arrayPositions,
options: DefaultTextIndexingOptions,
analyzer: analyzer,
value: value,
numPlainTextBytes: uint64(len(value)),
}
}
func NewTextFieldCustom(name string, arrayPositions []uint64, value []byte, options IndexingOptions, analyzer *analysis.Analyzer) *TextField {
return &TextField{
name: name,
arrayPositions: arrayPositions,
options: options,
analyzer: analyzer,
value: value,
numPlainTextBytes: uint64(len(value)),
}
}

@ -0,0 +1,55 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package document
type IndexingOptions int
const (
IndexField IndexingOptions = 1 << iota
StoreField
IncludeTermVectors
)
func (o IndexingOptions) IsIndexed() bool {
return o&IndexField != 0
}
func (o IndexingOptions) IsStored() bool {
return o&StoreField != 0
}
func (o IndexingOptions) IncludeTermVectors() bool {
return o&IncludeTermVectors != 0
}
func (o IndexingOptions) String() string {
rv := ""
if o.IsIndexed() {
rv += "INDEXED"
}
if o.IsStored() {
if rv != "" {
rv += ", "
}
rv += "STORE"
}
if o.IncludeTermVectors() {
if rv != "" {
rv += ", "
}
rv += "TV"
}
return rv
}

@ -0,0 +1,52 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package bleve
// Constant Error values which can be compared to determine the type of error
const (
ErrorIndexPathExists Error = iota
ErrorIndexPathDoesNotExist
ErrorIndexMetaMissing
ErrorIndexMetaCorrupt
ErrorUnknownStorageType
ErrorIndexClosed
ErrorAliasMulti
ErrorAliasEmpty
ErrorUnknownIndexType
ErrorEmptyID
ErrorIndexReadInconsistency
)
// Error represents a more strongly typed bleve error for detecting
// and handling specific types of errors.
type Error int
func (e Error) Error() string {
return errorMessages[e]
}
var errorMessages = map[Error]string{
ErrorIndexPathExists: "cannot create new index, path already exists",
ErrorIndexPathDoesNotExist: "cannot open index, path does not exist",
ErrorIndexMetaMissing: "cannot open index, metadata missing",
ErrorIndexMetaCorrupt: "cannot open index, metadata corrupt",
ErrorUnknownStorageType: "unknown storage type",
ErrorIndexClosed: "index is closed",
ErrorAliasMulti: "cannot perform single index operation on multiple index alias",
ErrorAliasEmpty: "cannot perform operation on empty alias",
ErrorUnknownIndexType: "unknown index type",
ErrorEmptyID: "document ID cannot be empty",
ErrorIndexReadInconsistency: "index read inconsistency detected",
}

@ -0,0 +1,243 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package bleve
import (
"github.com/blevesearch/bleve/document"
"github.com/blevesearch/bleve/index"
"github.com/blevesearch/bleve/index/store"
"github.com/blevesearch/bleve/mapping"
"golang.org/x/net/context"
)
// A Batch groups together multiple Index and Delete
// operations you would like performed at the same
// time. The Batch structure is NOT thread-safe.
// You should only perform operations on a batch
// from a single thread at a time. Once batch
// execution has started, you may not modify it.
type Batch struct {
index Index
internal *index.Batch
}
// Index adds the specified index operation to the
// batch. NOTE: the bleve Index is not updated
// until the batch is executed.
func (b *Batch) Index(id string, data interface{}) error {
if id == "" {
return ErrorEmptyID
}
doc := document.NewDocument(id)
err := b.index.Mapping().MapDocument(doc, data)
if err != nil {
return err
}
b.internal.Update(doc)
return nil
}
// Delete adds the specified delete operation to the
// batch. NOTE: the bleve Index is not updated until
// the batch is executed.
func (b *Batch) Delete(id string) {
if id != "" {
b.internal.Delete(id)
}
}
// SetInternal adds the specified set internal
// operation to the batch. NOTE: the bleve Index is
// not updated until the batch is executed.
func (b *Batch) SetInternal(key, val []byte) {
b.internal.SetInternal(key, val)
}
// SetInternal adds the specified delete internal
// operation to the batch. NOTE: the bleve Index is
// not updated until the batch is executed.
func (b *Batch) DeleteInternal(key []byte) {
b.internal.DeleteInternal(key)
}
// Size returns the total number of operations inside the batch
// including normal index operations and internal operations.
func (b *Batch) Size() int {
return len(b.internal.IndexOps) + len(b.internal.InternalOps)
}
// String prints a user friendly string representation of what
// is inside this batch.
func (b *Batch) String() string {
return b.internal.String()
}
// Reset returns a Batch to the empty state so that it can
// be re-used in the future.
func (b *Batch) Reset() {
b.internal.Reset()
}
// An Index implements all the indexing and searching
// capabilities of bleve. An Index can be created
// using the New() and Open() methods.
//
// Index() takes an input value, deduces a DocumentMapping for its type,
// assigns string paths to its fields or values then applies field mappings on
// them.
//
// The DocumentMapping used to index a value is deduced by the following rules:
// 1) If value implements Classifier interface, resolve the mapping from Type().
// 2) If value has a string field or value at IndexMapping.TypeField.
// (defaulting to "_type"), use it to resolve the mapping. Fields addressing
// is described below.
// 3) If IndexMapping.DefaultType is registered, return it.
// 4) Return IndexMapping.DefaultMapping.
//
// Each field or nested field of the value is identified by a string path, then
// mapped to one or several FieldMappings which extract the result for analysis.
//
// Struct values fields are identified by their "json:" tag, or by their name.
// Nested fields are identified by prefixing with their parent identifier,
// separated by a dot.
//
// Map values entries are identified by their string key. Entries not indexed
// by strings are ignored. Entry values are identified recursively like struct
// fields.
//
// Slice and array values are identified by their field name. Their elements
// are processed sequentially with the same FieldMapping.
//
// String, float64 and time.Time values are identified by their field name.
// Other types are ignored.
//
// Each value identifier is decomposed in its parts and recursively address
// SubDocumentMappings in the tree starting at the root DocumentMapping. If a
// mapping is found, all its FieldMappings are applied to the value. If no
// mapping is found and the root DocumentMapping is dynamic, default mappings
// are used based on value type and IndexMapping default configurations.
//
// Finally, mapped values are analyzed, indexed or stored. See
// FieldMapping.Analyzer to know how an analyzer is resolved for a given field.
//
// Examples:
//
// type Date struct {
// Day string `json:"day"`
// Month string
// Year string
// }
//
// type Person struct {
// FirstName string `json:"first_name"`
// LastName string
// BirthDate Date `json:"birth_date"`
// }
//
// A Person value FirstName is mapped by the SubDocumentMapping at
// "first_name". Its LastName is mapped by the one at "LastName". The day of
// BirthDate is mapped to the SubDocumentMapping "day" of the root
// SubDocumentMapping "birth_date". It will appear as the "birth_date.day"
// field in the index. The month is mapped to "birth_date.Month".
type Index interface {
// Index analyzes, indexes or stores mapped data fields. Supplied
// identifier is bound to analyzed data and will be retrieved by search
// requests. See Index interface documentation for details about mapping
// rules.
Index(id string, data interface{}) error
Delete(id string) error
NewBatch() *Batch
Batch(b *Batch) error
// Document returns specified document or nil if the document is not
// indexed or stored.
Document(id string) (*document.Document, error)
// DocCount returns the number of documents in the index.
DocCount() (uint64, error)
Search(req *SearchRequest) (*SearchResult, error)
SearchInContext(ctx context.Context, req *SearchRequest) (*SearchResult, error)
Fields() ([]string, error)
FieldDict(field string) (index.FieldDict, error)
FieldDictRange(field string, startTerm []byte, endTerm []byte) (index.FieldDict, error)
FieldDictPrefix(field string, termPrefix []byte) (index.FieldDict, error)
Close() error
Mapping() mapping.IndexMapping
Stats() *IndexStat
StatsMap() map[string]interface{}
GetInternal(key []byte) ([]byte, error)
SetInternal(key, val []byte) error
DeleteInternal(key []byte) error
// Name returns the name of the index (by default this is the path)
Name() string
// SetName lets you assign your own logical name to this index
SetName(string)
// Advanced returns the indexer and data store, exposing lower level
// methods to enumerate records and access data.
Advanced() (index.Index, store.KVStore, error)
}
// New index at the specified path, must not exist.
// The provided mapping will be used for all
// Index/Search operations.
func New(path string, mapping mapping.IndexMapping) (Index, error) {
return newIndexUsing(path, mapping, Config.DefaultIndexType, Config.DefaultKVStore, nil)
}
// NewMemOnly creates a memory-only index.
// The contents of the index is NOT persisted,
// and will be lost once closed.
// The provided mapping will be used for all
// Index/Search operations.
func NewMemOnly(mapping mapping.IndexMapping) (Index, error) {
return newIndexUsing("", mapping, Config.DefaultIndexType, Config.DefaultMemKVStore, nil)
}
// NewUsing creates index at the specified path,
// which must not already exist.
// The provided mapping will be used for all
// Index/Search operations.
// The specified index type will be used.
// The specified kvstore implementation will be used
// and the provided kvconfig will be passed to its
// constructor. Note that currently the values of kvconfig must
// be able to be marshaled and unmarshaled using the encoding/json library (used
// when reading/writing the index metadata file).
func NewUsing(path string, mapping mapping.IndexMapping, indexType string, kvstore string, kvconfig map[string]interface{}) (Index, error) {
return newIndexUsing(path, mapping, indexType, kvstore, kvconfig)
}
// Open index at the specified path, must exist.
// The mapping used when it was created will be used for all Index/Search operations.
func Open(path string) (Index, error) {
return openIndexUsing(path, nil)
}
// OpenUsing opens index at the specified path, must exist.
// The mapping used when it was created will be used for all Index/Search operations.
// The provided runtimeConfig can override settings
// persisted when the kvstore was created.
func OpenUsing(path string, runtimeConfig map[string]interface{}) (Index, error) {
return openIndexUsing(path, runtimeConfig)
}

@ -0,0 +1,83 @@
// Copyright (c) 2015 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package index
import "github.com/blevesearch/bleve/document"
type IndexRow interface {
KeySize() int
KeyTo([]byte) (int, error)
Key() []byte
ValueSize() int
ValueTo([]byte) (int, error)
Value() []byte
}
type AnalysisResult struct {
DocID string
Rows []IndexRow
}
type AnalysisWork struct {
i Index
d *document.Document
rc chan *AnalysisResult
}
func NewAnalysisWork(i Index, d *document.Document, rc chan *AnalysisResult) *AnalysisWork {
return &AnalysisWork{
i: i,
d: d,
rc: rc,
}
}
type AnalysisQueue struct {
queue chan *AnalysisWork
done chan struct{}
}
func (q *AnalysisQueue) Queue(work *AnalysisWork) {
q.queue <- work
}
func (q *AnalysisQueue) Close() {
close(q.done)
}
func NewAnalysisQueue(numWorkers int) *AnalysisQueue {
rv := AnalysisQueue{
queue: make(chan *AnalysisWork),
done: make(chan struct{}),
}
for i := 0; i < numWorkers; i++ {
go AnalysisWorker(rv)
}
return &rv
}
func AnalysisWorker(q AnalysisQueue) {
// read work off the queue
for {
select {
case <-q.done:
return
case w := <-q.queue:
r := w.i.Analyze(w.d)
w.rc <- r
}
}
}

@ -0,0 +1,88 @@
// Copyright (c) 2015 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package index
import (
"sync"
)
type FieldCache struct {
fieldIndexes map[string]uint16
indexFields []string
lastFieldIndex int
mutex sync.RWMutex
}
func NewFieldCache() *FieldCache {
return &FieldCache{
fieldIndexes: make(map[string]uint16),
lastFieldIndex: -1,
}
}
func (f *FieldCache) AddExisting(field string, index uint16) {
f.mutex.Lock()
f.addLOCKED(field, index)
f.mutex.Unlock()
}
func (f *FieldCache) addLOCKED(field string, index uint16) uint16 {
f.fieldIndexes[field] = index
if len(f.indexFields) < int(index)+1 {
prevIndexFields := f.indexFields
f.indexFields = make([]string, int(index)+16)
copy(f.indexFields, prevIndexFields)
}
f.indexFields[int(index)] = field
if int(index) > f.lastFieldIndex {
f.lastFieldIndex = int(index)
}
return index
}
// FieldNamed returns the index of the field, and whether or not it existed
// before this call. if createIfMissing is true, and new field index is assigned
// but the second return value will still be false
func (f *FieldCache) FieldNamed(field string, createIfMissing bool) (uint16, bool) {
f.mutex.RLock()
if index, ok := f.fieldIndexes[field]; ok {
f.mutex.RUnlock()
return index, true
} else if !createIfMissing {
f.mutex.RUnlock()
return 0, false
}
// trade read lock for write lock
f.mutex.RUnlock()
f.mutex.Lock()
// need to check again with write lock
if index, ok := f.fieldIndexes[field]; ok {
f.mutex.Unlock()
return index, true
}
// assign next field id
index := f.addLOCKED(field, uint16(f.lastFieldIndex+1))
f.mutex.Unlock()
return index, false
}
func (f *FieldCache) FieldIndexed(index uint16) (field string) {
f.mutex.RLock()
if int(index) < len(f.indexFields) {
field = f.indexFields[int(index)]
}
f.mutex.RUnlock()
return field
}

@ -0,0 +1,239 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package index
import (
"bytes"
"encoding/json"
"fmt"
"github.com/blevesearch/bleve/document"
"github.com/blevesearch/bleve/index/store"
)
var ErrorUnknownStorageType = fmt.Errorf("unknown storage type")
type Index interface {
Open() error
Close() error
Update(doc *document.Document) error
Delete(id string) error
Batch(batch *Batch) error
SetInternal(key, val []byte) error
DeleteInternal(key []byte) error
// Reader returns a low-level accessor on the index data. Close it to
// release associated resources.
Reader() (IndexReader, error)
Stats() json.Marshaler
StatsMap() map[string]interface{}
Analyze(d *document.Document) *AnalysisResult
Advanced() (store.KVStore, error)
}
type IndexReader interface {
TermFieldReader(term []byte, field string, includeFreq, includeNorm, includeTermVectors bool) (TermFieldReader, error)
// DocIDReader returns an iterator over all doc ids
// The caller must close returned instance to release associated resources.
DocIDReaderAll() (DocIDReader, error)
DocIDReaderOnly(ids []string) (DocIDReader, error)
FieldDict(field string) (FieldDict, error)
// FieldDictRange is currently defined to include the start and end terms
FieldDictRange(field string, startTerm []byte, endTerm []byte) (FieldDict, error)
FieldDictPrefix(field string, termPrefix []byte) (FieldDict, error)
Document(id string) (*document.Document, error)
DocumentFieldTerms(id IndexInternalID, fields []string) (FieldTerms, error)
Fields() ([]string, error)
GetInternal(key []byte) ([]byte, error)
DocCount() (uint64, error)
ExternalID(id IndexInternalID) (string, error)
InternalID(id string) (IndexInternalID, error)
DumpAll() chan interface{}
DumpDoc(id string) chan interface{}
DumpFields() chan interface{}
Close() error
}
// FieldTerms contains the terms used by a document, keyed by field
type FieldTerms map[string][]string
// FieldsNotYetCached returns a list of fields not yet cached out of a larger list of fields
func (f FieldTerms) FieldsNotYetCached(fields []string) []string {
rv := make([]string, 0, len(fields))
for _, field := range fields {
if _, ok := f[field]; !ok {
rv = append(rv, field)
}
}
return rv
}
// Merge will combine two FieldTerms
// it assumes that the terms lists are complete (thus do not need to be merged)
// field terms from the other list always replace the ones in the receiver
func (f FieldTerms) Merge(other FieldTerms) {
for field, terms := range other {
f[field] = terms
}
}
type TermFieldVector struct {
Field string
ArrayPositions []uint64
Pos uint64
Start uint64
End uint64
}
// IndexInternalID is an opaque document identifier interal to the index impl
type IndexInternalID []byte
func (id IndexInternalID) Equals(other IndexInternalID) bool {
return id.Compare(other) == 0
}
func (id IndexInternalID) Compare(other IndexInternalID) int {
return bytes.Compare(id, other)
}
type TermFieldDoc struct {
Term string
ID IndexInternalID
Freq uint64
Norm float64
Vectors []*TermFieldVector
}
// Reset allows an already allocated TermFieldDoc to be reused
func (tfd *TermFieldDoc) Reset() *TermFieldDoc {
// remember the []byte used for the ID
id := tfd.ID
// idiom to copy over from empty TermFieldDoc (0 allocations)
*tfd = TermFieldDoc{}
// reuse the []byte already allocated (and reset len to 0)
tfd.ID = id[:0]
return tfd
}
// TermFieldReader is the interface exposing the enumeration of documents
// containing a given term in a given field. Documents are returned in byte
// lexicographic order over their identifiers.
type TermFieldReader interface {
// Next returns the next document containing the term in this field, or nil
// when it reaches the end of the enumeration. The preAlloced TermFieldDoc
// is optional, and when non-nil, will be used instead of allocating memory.
Next(preAlloced *TermFieldDoc) (*TermFieldDoc, error)
// Advance resets the enumeration at specified document or its immediate
// follower.
Advance(ID IndexInternalID, preAlloced *TermFieldDoc) (*TermFieldDoc, error)
// Count returns the number of documents contains the term in this field.
Count() uint64
Close() error
}
type DictEntry struct {
Term string
Count uint64
}
type FieldDict interface {
Next() (*DictEntry, error)
Close() error
}
// DocIDReader is the interface exposing enumeration of documents identifiers.
// Close the reader to release associated resources.
type DocIDReader interface {
// Next returns the next document internal identifier in the natural
// index order, nil when the end of the sequence is reached.
Next() (IndexInternalID, error)
// Advance resets the iteration to the first internal identifier greater than
// or equal to ID. If ID is smaller than the start of the range, the iteration
// will start there instead. If ID is greater than or equal to the end of
// the range, Next() call will return io.EOF.
Advance(ID IndexInternalID) (IndexInternalID, error)
Close() error
}
type Batch struct {
IndexOps map[string]*document.Document
InternalOps map[string][]byte
}
func NewBatch() *Batch {
return &Batch{
IndexOps: make(map[string]*document.Document),
InternalOps: make(map[string][]byte),
}
}
func (b *Batch) Update(doc *document.Document) {
b.IndexOps[doc.ID] = doc
}
func (b *Batch) Delete(id string) {
b.IndexOps[id] = nil
}
func (b *Batch) SetInternal(key, val []byte) {
b.InternalOps[string(key)] = val
}
func (b *Batch) DeleteInternal(key []byte) {
b.InternalOps[string(key)] = nil
}
func (b *Batch) String() string {
rv := fmt.Sprintf("Batch (%d ops, %d internal ops)\n", len(b.IndexOps), len(b.InternalOps))
for k, v := range b.IndexOps {
if v != nil {
rv += fmt.Sprintf("\tINDEX - '%s'\n", k)
} else {
rv += fmt.Sprintf("\tDELETE - '%s'\n", k)
}
}
for k, v := range b.InternalOps {
if v != nil {
rv += fmt.Sprintf("\tSET INTERNAL - '%s'\n", k)
} else {
rv += fmt.Sprintf("\tDELETE INTERNAL - '%s'\n", k)
}
}
return rv
}
func (b *Batch) Reset() {
b.IndexOps = make(map[string]*document.Document)
b.InternalOps = make(map[string][]byte)
}

@ -0,0 +1,62 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package store
type op struct {
K []byte
V []byte
}
type EmulatedBatch struct {
Ops []*op
Merger *EmulatedMerge
}
func NewEmulatedBatch(mo MergeOperator) *EmulatedBatch {
return &EmulatedBatch{
Ops: make([]*op, 0, 1000),
Merger: NewEmulatedMerge(mo),
}
}
func (b *EmulatedBatch) Set(key, val []byte) {
ck := make([]byte, len(key))
copy(ck, key)
cv := make([]byte, len(val))
copy(cv, val)
b.Ops = append(b.Ops, &op{ck, cv})
}
func (b *EmulatedBatch) Delete(key []byte) {
ck := make([]byte, len(key))
copy(ck, key)
b.Ops = append(b.Ops, &op{ck, nil})
}
func (b *EmulatedBatch) Merge(key, val []byte) {
ck := make([]byte, len(key))
copy(ck, key)
cv := make([]byte, len(val))
copy(cv, val)
b.Merger.Merge(key, val)
}
func (b *EmulatedBatch) Reset() {
b.Ops = b.Ops[:0]
}
func (b *EmulatedBatch) Close() error {
return nil
}

@ -0,0 +1,85 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package boltdb
import (
"bytes"
"github.com/boltdb/bolt"
)
type Iterator struct {
store *Store
tx *bolt.Tx
cursor *bolt.Cursor
prefix []byte
start []byte
end []byte
valid bool
key []byte
val []byte
}
func (i *Iterator) updateValid() {
i.valid = (i.key != nil)
if i.valid {
if i.prefix != nil {
i.valid = bytes.HasPrefix(i.key, i.prefix)
} else if i.end != nil {
i.valid = bytes.Compare(i.key, i.end) < 0
}
}
}
func (i *Iterator) Seek(k []byte) {
if i.start != nil && bytes.Compare(k, i.start) < 0 {
k = i.start
}
if i.prefix != nil && !bytes.HasPrefix(k, i.prefix) {
if bytes.Compare(k, i.prefix) < 0 {
k = i.prefix
} else {
i.valid = false
return
}
}
i.key, i.val = i.cursor.Seek(k)
i.updateValid()
}
func (i *Iterator) Next() {
i.key, i.val = i.cursor.Next()
i.updateValid()
}
func (i *Iterator) Current() ([]byte, []byte, bool) {
return i.key, i.val, i.valid
}
func (i *Iterator) Key() []byte {
return i.key
}
func (i *Iterator) Value() []byte {
return i.val
}
func (i *Iterator) Valid() bool {
return i.valid
}
func (i *Iterator) Close() error {
return nil
}

@ -0,0 +1,73 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package boltdb
import (
"github.com/blevesearch/bleve/index/store"
"github.com/boltdb/bolt"
)
type Reader struct {
store *Store
tx *bolt.Tx
bucket *bolt.Bucket
}
func (r *Reader) Get(key []byte) ([]byte, error) {
var rv []byte
v := r.bucket.Get(key)
if v != nil {
rv = make([]byte, len(v))
copy(rv, v)
}
return rv, nil
}
func (r *Reader) MultiGet(keys [][]byte) ([][]byte, error) {
return store.MultiGet(r, keys)
}
func (r *Reader) PrefixIterator(prefix []byte) store.KVIterator {
cursor := r.bucket.Cursor()
rv := &Iterator{
store: r.store,
tx: r.tx,
cursor: cursor,
prefix: prefix,
}
rv.Seek(prefix)
return rv
}
func (r *Reader) RangeIterator(start, end []byte) store.KVIterator {
cursor := r.bucket.Cursor()
rv := &Iterator{
store: r.store,
tx: r.tx,
cursor: cursor,
start: start,
end: end,
}
rv.Seek(start)
return rv
}
func (r *Reader) Close() error {
return r.tx.Rollback()
}

@ -0,0 +1,26 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package boltdb
import "encoding/json"
type stats struct {
s *Store
}
func (s *stats) MarshalJSON() ([]byte, error) {
bs := s.s.db.Stats()
return json.Marshal(bs)
}

@ -0,0 +1,175 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
// Package boltdb implements a store.KVStore on top of BoltDB. It supports the
// following options:
//
// "bucket" (string): the name of BoltDB bucket to use, defaults to "bleve".
//
// "nosync" (bool): if true, set boltdb.DB.NoSync to true. It speeds up index
// operations in exchange of losing integrity guarantees if indexation aborts
// without closing the index. Use it when rebuilding indexes from zero.
package boltdb
import (
"bytes"
"encoding/json"
"fmt"
"os"
"github.com/blevesearch/bleve/index/store"
"github.com/blevesearch/bleve/registry"
"github.com/boltdb/bolt"
)
const (
Name = "boltdb"
defaultCompactBatchSize = 100
)
type Store struct {
path string
bucket string
db *bolt.DB
noSync bool
fillPercent float64
mo store.MergeOperator
}
func New(mo store.MergeOperator, config map[string]interface{}) (store.KVStore, error) {
path, ok := config["path"].(string)
if !ok {
return nil, fmt.Errorf("must specify path")
}
if path == "" {
return nil, os.ErrInvalid
}
bucket, ok := config["bucket"].(string)
if !ok {
bucket = "bleve"
}
noSync, _ := config["nosync"].(bool)
fillPercent, ok := config["fillPercent"].(float64)
if !ok {
fillPercent = bolt.DefaultFillPercent
}
bo := &bolt.Options{}
ro, ok := config["read_only"].(bool)
if ok {
bo.ReadOnly = ro
}
db, err := bolt.Open(path, 0600, bo)
if err != nil {
return nil, err
}
db.NoSync = noSync
if !bo.ReadOnly {
err = db.Update(func(tx *bolt.Tx) error {
_, err := tx.CreateBucketIfNotExists([]byte(bucket))
return err
})
if err != nil {
return nil, err
}
}
rv := Store{
path: path,
bucket: bucket,
db: db,
mo: mo,
noSync: noSync,
fillPercent: fillPercent,
}
return &rv, nil
}
func (bs *Store) Close() error {
return bs.db.Close()
}
func (bs *Store) Reader() (store.KVReader, error) {
tx, err := bs.db.Begin(false)
if err != nil {
return nil, err
}
return &Reader{
store: bs,
tx: tx,
bucket: tx.Bucket([]byte(bs.bucket)),
}, nil
}
func (bs *Store) Writer() (store.KVWriter, error) {
return &Writer{
store: bs,
}, nil
}
func (bs *Store) Stats() json.Marshaler {
return &stats{
s: bs,
}
}
// CompactWithBatchSize removes DictionaryTerm entries with a count of zero (in batchSize batches)
// Removing entries is a workaround for github issue #374.
func (bs *Store) CompactWithBatchSize(batchSize int) error {
for {
cnt := 0
err := bs.db.Batch(func(tx *bolt.Tx) error {
c := tx.Bucket([]byte(bs.bucket)).Cursor()
prefix := []byte("d")
for k, v := c.Seek(prefix); bytes.HasPrefix(k, prefix); k, v = c.Next() {
if bytes.Equal(v, []byte{0}) {
cnt++
if err := c.Delete(); err != nil {
return err
}
if cnt == batchSize {
break
}
}
}
return nil
})
if err != nil {
return err
}
if cnt == 0 {
break
}
}
return nil
}
// Compact calls CompactWithBatchSize with a default batch size of 100. This is a workaround
// for github issue #374.
func (bs *Store) Compact() error {
return bs.CompactWithBatchSize(defaultCompactBatchSize)
}
func init() {
registry.RegisterKVStore(Name, New)
}

@ -0,0 +1,95 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package boltdb
import (
"fmt"
"github.com/blevesearch/bleve/index/store"
)
type Writer struct {
store *Store
}
func (w *Writer) NewBatch() store.KVBatch {
return store.NewEmulatedBatch(w.store.mo)
}
func (w *Writer) NewBatchEx(options store.KVBatchOptions) ([]byte, store.KVBatch, error) {
return make([]byte, options.TotalBytes), w.NewBatch(), nil
}
func (w *Writer) ExecuteBatch(batch store.KVBatch) (err error) {
emulatedBatch, ok := batch.(*store.EmulatedBatch)
if !ok {
return fmt.Errorf("wrong type of batch")
}
tx, err := w.store.db.Begin(true)
if err != nil {
return
}
// defer function to ensure that once started,
// we either Commit tx or Rollback
defer func() {
// if nothing went wrong, commit
if err == nil {
// careful to catch error here too
err = tx.Commit()
} else {
// caller should see error that caused abort,
// not success or failure of Rollback itself
_ = tx.Rollback()
}
}()
bucket := tx.Bucket([]byte(w.store.bucket))
bucket.FillPercent = w.store.fillPercent
for k, mergeOps := range emulatedBatch.Merger.Merges {
kb := []byte(k)
existingVal := bucket.Get(kb)
mergedVal, fullMergeOk := w.store.mo.FullMerge(kb, existingVal, mergeOps)
if !fullMergeOk {
err = fmt.Errorf("merge operator returned failure")
return
}
err = bucket.Put(kb, mergedVal)
if err != nil {
return
}
}
for _, op := range emulatedBatch.Ops {
if op.V != nil {
err = bucket.Put(op.K, op.V)
if err != nil {
return
}
} else {
err = bucket.Delete(op.K)
if err != nil {
return
}
}
}
return
}
func (w *Writer) Close() error {
return nil
}

@ -0,0 +1,152 @@
// Copyright (c) 2015 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
// Package gtreap provides an in-memory implementation of the
// KVStore interfaces using the gtreap balanced-binary treap,
// copy-on-write data structure.
package gtreap
import (
"bytes"
"sync"
"github.com/steveyen/gtreap"
)
type Iterator struct {
t *gtreap.Treap
m sync.Mutex
cancelCh chan struct{}
nextCh chan *Item
curr *Item
currOk bool
prefix []byte
start []byte
end []byte
}
func (w *Iterator) Seek(k []byte) {
if w.start != nil && bytes.Compare(k, w.start) < 0 {
k = w.start
}
if w.prefix != nil && !bytes.HasPrefix(k, w.prefix) {
if bytes.Compare(k, w.prefix) < 0 {
k = w.prefix
} else {
var end []byte
for i := len(w.prefix) - 1; i >= 0; i-- {
c := w.prefix[i]
if c < 0xff {
end = make([]byte, i+1)
copy(end, w.prefix)
end[i] = c + 1
break
}
}
k = end
}
}
w.restart(&Item{k: k})
}
func (w *Iterator) restart(start *Item) *Iterator {
cancelCh := make(chan struct{})
nextCh := make(chan *Item, 1)
w.m.Lock()
if w.cancelCh != nil {
close(w.cancelCh)
}
w.cancelCh = cancelCh
w.nextCh = nextCh
w.curr = nil
w.currOk = false
w.m.Unlock()
go func() {
if start != nil {
w.t.VisitAscend(start, func(itm gtreap.Item) bool {
select {
case <-cancelCh:
return false
case nextCh <- itm.(*Item):
return true
}
})
}
close(nextCh)
}()
w.Next()
return w
}
func (w *Iterator) Next() {
w.m.Lock()
nextCh := w.nextCh
w.m.Unlock()
w.curr, w.currOk = <-nextCh
}
func (w *Iterator) Current() ([]byte, []byte, bool) {
w.m.Lock()
defer w.m.Unlock()
if !w.currOk || w.curr == nil {
return nil, nil, false
}
if w.prefix != nil && !bytes.HasPrefix(w.curr.k, w.prefix) {
return nil, nil, false
} else if w.end != nil && bytes.Compare(w.curr.k, w.end) >= 0 {
return nil, nil, false
}
return w.curr.k, w.curr.v, w.currOk
}
func (w *Iterator) Key() []byte {
k, _, ok := w.Current()
if !ok {
return nil
}
return k
}
func (w *Iterator) Value() []byte {
_, v, ok := w.Current()
if !ok {
return nil
}
return v
}
func (w *Iterator) Valid() bool {
_, _, ok := w.Current()
return ok
}
func (w *Iterator) Close() error {
w.m.Lock()
if w.cancelCh != nil {
close(w.cancelCh)
}
w.cancelCh = nil
w.nextCh = nil
w.curr = nil
w.currOk = false
w.m.Unlock()
return nil
}

@ -0,0 +1,66 @@
// Copyright (c) 2015 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
// Package gtreap provides an in-memory implementation of the
// KVStore interfaces using the gtreap balanced-binary treap,
// copy-on-write data structure.
package gtreap
import (
"github.com/blevesearch/bleve/index/store"
"github.com/steveyen/gtreap"
)
type Reader struct {
t *gtreap.Treap
}
func (w *Reader) Get(k []byte) (v []byte, err error) {
var rv []byte
itm := w.t.Get(&Item{k: k})
if itm != nil {
rv = make([]byte, len(itm.(*Item).v))
copy(rv, itm.(*Item).v)
return rv, nil
}
return nil, nil
}
func (r *Reader) MultiGet(keys [][]byte) ([][]byte, error) {
return store.MultiGet(r, keys)
}
func (w *Reader) PrefixIterator(k []byte) store.KVIterator {
rv := Iterator{
t: w.t,
prefix: k,
}
rv.restart(&Item{k: k})
return &rv
}
func (w *Reader) RangeIterator(start, end []byte) store.KVIterator {
rv := Iterator{
t: w.t,
start: start,
end: end,
}
rv.restart(&Item{k: start})
return &rv
}
func (w *Reader) Close() error {
return nil
}

@ -0,0 +1,82 @@
// Copyright (c) 2015 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
// Package gtreap provides an in-memory implementation of the
// KVStore interfaces using the gtreap balanced-binary treap,
// copy-on-write data structure.
package gtreap
import (
"bytes"
"fmt"
"os"
"sync"
"github.com/blevesearch/bleve/index/store"
"github.com/blevesearch/bleve/registry"
"github.com/steveyen/gtreap"
)
const Name = "gtreap"
type Store struct {
m sync.Mutex
t *gtreap.Treap
mo store.MergeOperator
}
type Item struct {
k []byte
v []byte
}
func itemCompare(a, b interface{}) int {
return bytes.Compare(a.(*Item).k, b.(*Item).k)
}
func New(mo store.MergeOperator, config map[string]interface{}) (store.KVStore, error) {
path, ok := config["path"].(string)
if !ok {
return nil, fmt.Errorf("must specify path")
}
if path != "" {
return nil, os.ErrInvalid
}
rv := Store{
t: gtreap.NewTreap(itemCompare),
mo: mo,
}
return &rv, nil
}
func (s *Store) Close() error {
return nil
}
func (s *Store) Reader() (store.KVReader, error) {
s.m.Lock()
t := s.t
s.m.Unlock()
return &Reader{t: t}, nil
}
func (s *Store) Writer() (store.KVWriter, error) {
return &Writer{s: s}, nil
}
func init() {
registry.RegisterKVStore(Name, New)
}

@ -0,0 +1,76 @@
// Copyright (c) 2015 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
// Package gtreap provides an in-memory implementation of the
// KVStore interfaces using the gtreap balanced-binary treap,
// copy-on-write data structure.
package gtreap
import (
"fmt"
"math/rand"
"github.com/blevesearch/bleve/index/store"
)
type Writer struct {
s *Store
}
func (w *Writer) NewBatch() store.KVBatch {
return store.NewEmulatedBatch(w.s.mo)
}
func (w *Writer) NewBatchEx(options store.KVBatchOptions) ([]byte, store.KVBatch, error) {
return make([]byte, options.TotalBytes), w.NewBatch(), nil
}
func (w *Writer) ExecuteBatch(batch store.KVBatch) error {
emulatedBatch, ok := batch.(*store.EmulatedBatch)
if !ok {
return fmt.Errorf("wrong type of batch")
}
w.s.m.Lock()
for k, mergeOps := range emulatedBatch.Merger.Merges {
kb := []byte(k)
var existingVal []byte
existingItem := w.s.t.Get(&Item{k: kb})
if existingItem != nil {
existingVal = w.s.t.Get(&Item{k: kb}).(*Item).v
}
mergedVal, fullMergeOk := w.s.mo.FullMerge(kb, existingVal, mergeOps)
if !fullMergeOk {
return fmt.Errorf("merge operator returned failure")
}
w.s.t = w.s.t.Upsert(&Item{k: kb, v: mergedVal}, rand.Int())
}
for _, op := range emulatedBatch.Ops {
if op.V != nil {
w.s.t = w.s.t.Upsert(&Item{k: op.K, v: op.V}, rand.Int())
} else {
w.s.t = w.s.t.Delete(&Item{k: op.K})
}
}
w.s.m.Unlock()
return nil
}
func (w *Writer) Close() error {
w.s = nil
return nil
}

@ -0,0 +1,174 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package store
import "encoding/json"
// KVStore is an abstraction for working with KV stores. Note that
// in order to be used with the bleve.registry, it must also implement
// a constructor function of the registry.KVStoreConstructor type.
type KVStore interface {
// Writer returns a KVWriter which can be used to
// make changes to the KVStore. If a writer cannot
// be obtained a non-nil error is returned.
Writer() (KVWriter, error)
// Reader returns a KVReader which can be used to
// read data from the KVStore. If a reader cannot
// be obtained a non-nil error is returned.
Reader() (KVReader, error)
// Close closes the KVStore
Close() error
}
// KVReader is an abstraction of an **ISOLATED** reader
// In this context isolated is defined to mean that
// writes/deletes made after the KVReader is opened
// are not observed.
// Because there is usually a cost associated with
// keeping isolated readers active, users should
// close them as soon as they are no longer needed.
type KVReader interface {
// Get returns the value associated with the key
// If the key does not exist, nil is returned.
// The caller owns the bytes returned.
Get(key []byte) ([]byte, error)
// MultiGet retrieves multiple values in one call.
MultiGet(keys [][]byte) ([][]byte, error)
// PrefixIterator returns a KVIterator that will
// visit all K/V pairs with the provided prefix
PrefixIterator(prefix []byte) KVIterator
// RangeIterator returns a KVIterator that will
// visit all K/V pairs >= start AND < end
RangeIterator(start, end []byte) KVIterator
// Close closes the iterator
Close() error
}
// KVIterator is an abstraction around key iteration
type KVIterator interface {
// Seek will advance the iterator to the specified key
Seek(key []byte)
// Next will advance the iterator to the next key
Next()
// Key returns the key pointed to by the iterator
// The bytes returned are **ONLY** valid until the next call to Seek/Next/Close
// Continued use after that requires that they be copied.
Key() []byte
// Value returns the value pointed to by the iterator
// The bytes returned are **ONLY** valid until the next call to Seek/Next/Close
// Continued use after that requires that they be copied.
Value() []byte
// Valid returns whether or not the iterator is in a valid state
Valid() bool
// Current returns Key(),Value(),Valid() in a single operation
Current() ([]byte, []byte, bool)
// Close closes the iterator
Close() error
}
// KVWriter is an abstraction for mutating the KVStore
// KVWriter does **NOT** enforce restrictions of a single writer
// if the underlying KVStore allows concurrent writes, the
// KVWriter interface should also do so, it is up to the caller
// to do this in a way that is safe and makes sense
type KVWriter interface {
// NewBatch returns a KVBatch for performing batch operations on this kvstore
NewBatch() KVBatch
// NewBatchEx returns a KVBatch and an associated byte array
// that's pre-sized based on the KVBatchOptions. The caller can
// use the returned byte array for keys and values associated with
// the batch. Once the batch is either executed or closed, the
// associated byte array should no longer be accessed by the
// caller.
NewBatchEx(KVBatchOptions) ([]byte, KVBatch, error)
// ExecuteBatch will execute the KVBatch, the provided KVBatch **MUST** have
// been created by the same KVStore (though not necessarily the same KVWriter)
// Batch execution is atomic, either all the operations or none will be performed
ExecuteBatch(batch KVBatch) error
// Close closes the writer
Close() error
}
// KVBatchOptions provides the KVWriter.NewBatchEx() method with batch
// preparation and preallocation information.
type KVBatchOptions struct {
// TotalBytes is the sum of key and value bytes needed by the
// caller for the entire batch. It affects the size of the
// returned byte array of KVWrite.NewBatchEx().
TotalBytes int
// NumSets is the number of Set() calls the caller will invoke on
// the KVBatch.
NumSets int
// NumDeletes is the number of Delete() calls the caller will invoke
// on the KVBatch.
NumDeletes int
// NumMerges is the number of Merge() calls the caller will invoke
// on the KVBatch.
NumMerges int
}
// KVBatch is an abstraction for making multiple KV mutations at once
type KVBatch interface {
// Set updates the key with the specified value
// both key and value []byte may be reused as soon as this call returns
Set(key, val []byte)
// Delete removes the specified key
// the key []byte may be reused as soon as this call returns
Delete(key []byte)
// Merge merges old value with the new value at the specified key
// as prescribed by the KVStores merge operator
// both key and value []byte may be reused as soon as this call returns
Merge(key, val []byte)
// Reset frees resources for this batch and allows reuse
Reset()
// Close frees resources
Close() error
}
// KVStoreStats is an optional interface that KVStores can implement
// if they're able to report any useful stats
type KVStoreStats interface {
// Stats returns a JSON serializable object representing stats for this KVStore
Stats() json.Marshaler
StatsMap() map[string]interface{}
}

@ -0,0 +1,64 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package store
// At the moment this happens to be the same interface as described by
// RocksDB, but this may not always be the case.
type MergeOperator interface {
// FullMerge the full sequence of operands on top of the existingValue
// if no value currently exists, existingValue is nil
// return the merged value, and success/failure
FullMerge(key, existingValue []byte, operands [][]byte) ([]byte, bool)
// Partially merge these two operands.
// If partial merge cannot be done, return nil,false, which will defer
// all processing until the FullMerge is done.
PartialMerge(key, leftOperand, rightOperand []byte) ([]byte, bool)
// Name returns an identifier for the operator
Name() string
}
type EmulatedMerge struct {
Merges map[string][][]byte
mo MergeOperator
}
func NewEmulatedMerge(mo MergeOperator) *EmulatedMerge {
return &EmulatedMerge{
Merges: make(map[string][][]byte),
mo: mo,
}
}
func (m *EmulatedMerge) Merge(key, val []byte) {
ops, ok := m.Merges[string(key)]
if ok && len(ops) > 0 {
last := ops[len(ops)-1]
mergedVal, partialMergeOk := m.mo.PartialMerge(key, last, val)
if partialMergeOk {
// replace last entry with the result of the merge
ops[len(ops)-1] = mergedVal
} else {
// could not partial merge, append this to the end
ops = append(ops, val)
}
} else {
ops = [][]byte{val}
}
m.Merges[string(key)] = ops
}

@ -0,0 +1,33 @@
// Copyright (c) 2016 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package store
// MultiGet is a helper function to retrieve mutiple keys from a
// KVReader, and might be used by KVStore implementations that don't
// have a native multi-get facility.
func MultiGet(kvreader KVReader, keys [][]byte) ([][]byte, error) {
vals := make([][]byte, 0, len(keys))
for i, key := range keys {
val, err := kvreader.Get(key)
if err != nil {
return nil, err
}
vals[i] = val
}
return vals, nil
}

@ -0,0 +1,110 @@
// Copyright (c) 2015 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package upsidedown
import (
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/document"
"github.com/blevesearch/bleve/index"
)
func (udc *UpsideDownCouch) Analyze(d *document.Document) *index.AnalysisResult {
rv := &index.AnalysisResult{
DocID: d.ID,
Rows: make([]index.IndexRow, 0, 100),
}
docIDBytes := []byte(d.ID)
// track our back index entries
backIndexStoredEntries := make([]*BackIndexStoreEntry, 0)
// information we collate as we merge fields with same name
fieldTermFreqs := make(map[uint16]analysis.TokenFrequencies)
fieldLengths := make(map[uint16]int)
fieldIncludeTermVectors := make(map[uint16]bool)
fieldNames := make(map[uint16]string)
analyzeField := func(field document.Field, storable bool) {
fieldIndex, newFieldRow := udc.fieldIndexOrNewRow(field.Name())
if newFieldRow != nil {
rv.Rows = append(rv.Rows, newFieldRow)
}
fieldNames[fieldIndex] = field.Name()
if field.Options().IsIndexed() {
fieldLength, tokenFreqs := field.Analyze()
existingFreqs := fieldTermFreqs[fieldIndex]
if existingFreqs == nil {
fieldTermFreqs[fieldIndex] = tokenFreqs
} else {
existingFreqs.MergeAll(field.Name(), tokenFreqs)
fieldTermFreqs[fieldIndex] = existingFreqs
}
fieldLengths[fieldIndex] += fieldLength
fieldIncludeTermVectors[fieldIndex] = field.Options().IncludeTermVectors()
}
if storable && field.Options().IsStored() {
rv.Rows, backIndexStoredEntries = udc.storeField(docIDBytes, field, fieldIndex, rv.Rows, backIndexStoredEntries)
}
}
// walk all the fields, record stored fields now
// place information about indexed fields into map
// this collates information across fields with
// same names (arrays)
for _, field := range d.Fields {
analyzeField(field, true)
}
if len(d.CompositeFields) > 0 {
for fieldIndex, tokenFreqs := range fieldTermFreqs {
// see if any of the composite fields need this
for _, compositeField := range d.CompositeFields {
compositeField.Compose(fieldNames[fieldIndex], fieldLengths[fieldIndex], tokenFreqs)
}
}
for _, compositeField := range d.CompositeFields {
analyzeField(compositeField, false)
}
}
rowsCapNeeded := len(rv.Rows) + 1
for _, tokenFreqs := range fieldTermFreqs {
rowsCapNeeded += len(tokenFreqs)
}
rv.Rows = append(make([]index.IndexRow, 0, rowsCapNeeded), rv.Rows...)
backIndexTermEntries := make([]*BackIndexTermEntry, 0, rowsCapNeeded)
// walk through the collated information and process
// once for each indexed field (unique name)
for fieldIndex, tokenFreqs := range fieldTermFreqs {
fieldLength := fieldLengths[fieldIndex]
includeTermVectors := fieldIncludeTermVectors[fieldIndex]
// encode this field
rv.Rows, backIndexTermEntries = udc.indexField(docIDBytes, includeTermVectors, fieldIndex, fieldLength, tokenFreqs, rv.Rows, backIndexTermEntries)
}
// build the back index row
backIndexRow := NewBackIndexRow(docIDBytes, backIndexTermEntries, backIndexStoredEntries)
rv.Rows = append(rv.Rows, backIndexRow)
return rv
}

@ -0,0 +1,8 @@
#!/bin/sh
BENCHMARKS=`grep "func Benchmark" *_test.go | sed 's/.*func //' | sed s/\(.*{//`
for BENCHMARK in $BENCHMARKS
do
go test -v -run=xxx -bench=^$BENCHMARK$ -benchtime=10s -tags 'forestdb leveldb' | grep -v ok | grep -v PASS
done

@ -0,0 +1,172 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package upsidedown
import (
"bytes"
"sort"
"github.com/blevesearch/bleve/index/store"
)
// the functions in this file are only intended to be used by
// the bleve_dump utility and the debug http handlers
// if your application relies on them, you're doing something wrong
// they may change or be removed at any time
func dumpPrefix(kvreader store.KVReader, rv chan interface{}, prefix []byte) {
start := prefix
if start == nil {
start = []byte{0}
}
it := kvreader.PrefixIterator(start)
defer func() {
cerr := it.Close()
if cerr != nil {
rv <- cerr
}
}()
key, val, valid := it.Current()
for valid {
ck := make([]byte, len(key))
copy(ck, key)
cv := make([]byte, len(val))
copy(cv, val)
row, err := ParseFromKeyValue(ck, cv)
if err != nil {
rv <- err
return
}
rv <- row
it.Next()
key, val, valid = it.Current()
}
}
func dumpRange(kvreader store.KVReader, rv chan interface{}, start, end []byte) {
it := kvreader.RangeIterator(start, end)
defer func() {
cerr := it.Close()
if cerr != nil {
rv <- cerr
}
}()
key, val, valid := it.Current()
for valid {
ck := make([]byte, len(key))
copy(ck, key)
cv := make([]byte, len(val))
copy(cv, val)
row, err := ParseFromKeyValue(ck, cv)
if err != nil {
rv <- err
return
}
rv <- row
it.Next()
key, val, valid = it.Current()
}
}
func (i *IndexReader) DumpAll() chan interface{} {
rv := make(chan interface{})
go func() {
defer close(rv)
dumpRange(i.kvreader, rv, nil, nil)
}()
return rv
}
func (i *IndexReader) DumpFields() chan interface{} {
rv := make(chan interface{})
go func() {
defer close(rv)
dumpPrefix(i.kvreader, rv, []byte{'f'})
}()
return rv
}
type keyset [][]byte
func (k keyset) Len() int { return len(k) }
func (k keyset) Swap(i, j int) { k[i], k[j] = k[j], k[i] }
func (k keyset) Less(i, j int) bool { return bytes.Compare(k[i], k[j]) < 0 }
// DumpDoc returns all rows in the index related to this doc id
func (i *IndexReader) DumpDoc(id string) chan interface{} {
idBytes := []byte(id)
rv := make(chan interface{})
go func() {
defer close(rv)
back, err := backIndexRowForDoc(i.kvreader, []byte(id))
if err != nil {
rv <- err
return
}
// no such doc
if back == nil {
return
}
// build sorted list of term keys
keys := make(keyset, 0)
for _, entry := range back.termEntries {
tfr := NewTermFrequencyRow([]byte(*entry.Term), uint16(*entry.Field), idBytes, 0, 0)
key := tfr.Key()
keys = append(keys, key)
}
sort.Sort(keys)
// first add all the stored rows
storedRowPrefix := NewStoredRow(idBytes, 0, []uint64{}, 'x', []byte{}).ScanPrefixForDoc()
dumpPrefix(i.kvreader, rv, storedRowPrefix)
// now walk term keys in order and add them as well
if len(keys) > 0 {
it := i.kvreader.RangeIterator(keys[0], nil)
defer func() {
cerr := it.Close()
if cerr != nil {
rv <- cerr
}
}()
for _, key := range keys {
it.Seek(key)
rkey, rval, valid := it.Current()
if !valid {
break
}
rck := make([]byte, len(rkey))
copy(rck, key)
rcv := make([]byte, len(rval))
copy(rcv, rval)
row, err := ParseFromKeyValue(rck, rcv)
if err != nil {
rv <- err
return
}
rv <- row
}
}
}()
return rv
}

@ -0,0 +1,78 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package upsidedown
import (
"fmt"
"github.com/blevesearch/bleve/index"
"github.com/blevesearch/bleve/index/store"
)
type UpsideDownCouchFieldDict struct {
indexReader *IndexReader
iterator store.KVIterator
dictRow *DictionaryRow
dictEntry *index.DictEntry
field uint16
}
func newUpsideDownCouchFieldDict(indexReader *IndexReader, field uint16, startTerm, endTerm []byte) (*UpsideDownCouchFieldDict, error) {
startKey := NewDictionaryRow(startTerm, field, 0).Key()
if endTerm == nil {
endTerm = []byte{ByteSeparator}
} else {
endTerm = incrementBytes(endTerm)
}
endKey := NewDictionaryRow(endTerm, field, 0).Key()
it := indexReader.kvreader.RangeIterator(startKey, endKey)
return &UpsideDownCouchFieldDict{
indexReader: indexReader,
iterator: it,
dictRow: &DictionaryRow{}, // Pre-alloced, reused row.
dictEntry: &index.DictEntry{}, // Pre-alloced, reused entry.
field: field,
}, nil
}
func (r *UpsideDownCouchFieldDict) Next() (*index.DictEntry, error) {
key, val, valid := r.iterator.Current()
if !valid {
return nil, nil
}
err := r.dictRow.parseDictionaryK(key)
if err != nil {
return nil, fmt.Errorf("unexpected error parsing dictionary row key: %v", err)
}
err = r.dictRow.parseDictionaryV(val)
if err != nil {
return nil, fmt.Errorf("unexpected error parsing dictionary row val: %v", err)
}
r.dictEntry.Term = string(r.dictRow.term)
r.dictEntry.Count = r.dictRow.count
// advance the iterator to the next term
r.iterator.Next()
return r.dictEntry, nil
}
func (r *UpsideDownCouchFieldDict) Close() error {
return r.iterator.Close()
}

@ -0,0 +1,189 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package upsidedown
import (
"github.com/blevesearch/bleve/document"
"github.com/blevesearch/bleve/index"
"github.com/blevesearch/bleve/index/store"
)
type IndexReader struct {
index *UpsideDownCouch
kvreader store.KVReader
docCount uint64
}
func (i *IndexReader) TermFieldReader(term []byte, fieldName string, includeFreq, includeNorm, includeTermVectors bool) (index.TermFieldReader, error) {
fieldIndex, fieldExists := i.index.fieldCache.FieldNamed(fieldName, false)
if fieldExists {
return newUpsideDownCouchTermFieldReader(i, term, uint16(fieldIndex), includeFreq, includeNorm, includeTermVectors)
}
return newUpsideDownCouchTermFieldReader(i, []byte{ByteSeparator}, ^uint16(0), includeFreq, includeNorm, includeTermVectors)
}
func (i *IndexReader) FieldDict(fieldName string) (index.FieldDict, error) {
return i.FieldDictRange(fieldName, nil, nil)
}
func (i *IndexReader) FieldDictRange(fieldName string, startTerm []byte, endTerm []byte) (index.FieldDict, error) {
fieldIndex, fieldExists := i.index.fieldCache.FieldNamed(fieldName, false)
if fieldExists {
return newUpsideDownCouchFieldDict(i, uint16(fieldIndex), startTerm, endTerm)
}
return newUpsideDownCouchFieldDict(i, ^uint16(0), []byte{ByteSeparator}, []byte{})
}
func (i *IndexReader) FieldDictPrefix(fieldName string, termPrefix []byte) (index.FieldDict, error) {
return i.FieldDictRange(fieldName, termPrefix, termPrefix)
}
func (i *IndexReader) DocIDReaderAll() (index.DocIDReader, error) {
return newUpsideDownCouchDocIDReader(i)
}
func (i *IndexReader) DocIDReaderOnly(ids []string) (index.DocIDReader, error) {
return newUpsideDownCouchDocIDReaderOnly(i, ids)
}
func (i *IndexReader) Document(id string) (doc *document.Document, err error) {
// first hit the back index to confirm doc exists
var backIndexRow *BackIndexRow
backIndexRow, err = backIndexRowForDoc(i.kvreader, []byte(id))
if err != nil {
return
}
if backIndexRow == nil {
return
}
doc = document.NewDocument(id)
storedRow := NewStoredRow([]byte(id), 0, []uint64{}, 'x', nil)
storedRowScanPrefix := storedRow.ScanPrefixForDoc()
it := i.kvreader.PrefixIterator(storedRowScanPrefix)
defer func() {
if cerr := it.Close(); err == nil && cerr != nil {
err = cerr
}
}()
key, val, valid := it.Current()
for valid {
safeVal := make([]byte, len(val))
copy(safeVal, val)
var row *StoredRow
row, err = NewStoredRowKV(key, safeVal)
if err != nil {
doc = nil
return
}
if row != nil {
fieldName := i.index.fieldCache.FieldIndexed(row.field)
field := decodeFieldType(row.typ, fieldName, row.arrayPositions, row.value)
if field != nil {
doc.AddField(field)
}
}
it.Next()
key, val, valid = it.Current()
}
return
}
func (i *IndexReader) DocumentFieldTerms(id index.IndexInternalID, fields []string) (index.FieldTerms, error) {
back, err := backIndexRowForDoc(i.kvreader, id)
if err != nil {
return nil, err
}
if back == nil {
return nil, nil
}
rv := make(index.FieldTerms, len(fields))
fieldsMap := make(map[uint16]string, len(fields))
for _, f := range fields {
id, ok := i.index.fieldCache.FieldNamed(f, false)
if ok {
fieldsMap[id] = f
}
}
for _, entry := range back.termEntries {
if field, ok := fieldsMap[uint16(*entry.Field)]; ok {
rv[field] = append(rv[field], *entry.Term)
}
}
return rv, nil
}
func (i *IndexReader) Fields() (fields []string, err error) {
fields = make([]string, 0)
it := i.kvreader.PrefixIterator([]byte{'f'})
defer func() {
if cerr := it.Close(); err == nil && cerr != nil {
err = cerr
}
}()
key, val, valid := it.Current()
for valid {
var row UpsideDownCouchRow
row, err = ParseFromKeyValue(key, val)
if err != nil {
fields = nil
return
}
if row != nil {
fieldRow, ok := row.(*FieldRow)
if ok {
fields = append(fields, fieldRow.name)
}
}
it.Next()
key, val, valid = it.Current()
}
return
}
func (i *IndexReader) GetInternal(key []byte) ([]byte, error) {
internalRow := NewInternalRow(key, nil)
return i.kvreader.Get(internalRow.Key())
}
func (i *IndexReader) DocCount() (uint64, error) {
return i.docCount, nil
}
func (i *IndexReader) Close() error {
return i.kvreader.Close()
}
func (i *IndexReader) ExternalID(id index.IndexInternalID) (string, error) {
return string(id), nil
}
func (i *IndexReader) InternalID(id string) (index.IndexInternalID, error) {
return index.IndexInternalID(id), nil
}
func incrementBytes(in []byte) []byte {
rv := make([]byte, len(in))
copy(rv, in)
for i := len(rv) - 1; i >= 0; i-- {
rv[i] = rv[i] + 1
if rv[i] != 0 {
// didn't overflow, so stop
break
}
}
return rv
}

@ -0,0 +1,325 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package upsidedown
import (
"bytes"
"sort"
"sync/atomic"
"github.com/blevesearch/bleve/index"
"github.com/blevesearch/bleve/index/store"
)
type UpsideDownCouchTermFieldReader struct {
count uint64
indexReader *IndexReader
iterator store.KVIterator
term []byte
tfrNext *TermFrequencyRow
keyBuf []byte
field uint16
}
func newUpsideDownCouchTermFieldReader(indexReader *IndexReader, term []byte, field uint16, includeFreq, includeNorm, includeTermVectors bool) (*UpsideDownCouchTermFieldReader, error) {
dictionaryRow := NewDictionaryRow(term, field, 0)
val, err := indexReader.kvreader.Get(dictionaryRow.Key())
if err != nil {
return nil, err
}
if val == nil {
atomic.AddUint64(&indexReader.index.stats.termSearchersStarted, uint64(1))
return &UpsideDownCouchTermFieldReader{
count: 0,
term: term,
tfrNext: &TermFrequencyRow{},
field: field,
}, nil
}
err = dictionaryRow.parseDictionaryV(val)
if err != nil {
return nil, err
}
tfr := NewTermFrequencyRow(term, field, []byte{}, 0, 0)
it := indexReader.kvreader.PrefixIterator(tfr.Key())
atomic.AddUint64(&indexReader.index.stats.termSearchersStarted, uint64(1))
return &UpsideDownCouchTermFieldReader{
indexReader: indexReader,
iterator: it,
count: dictionaryRow.count,
term: term,
field: field,
}, nil
}
func (r *UpsideDownCouchTermFieldReader) Count() uint64 {
return r.count
}
func (r *UpsideDownCouchTermFieldReader) Next(preAlloced *index.TermFieldDoc) (*index.TermFieldDoc, error) {
if r.iterator != nil {
// We treat tfrNext also like an initialization flag, which
// tells us whether we need to invoke the underlying
// iterator.Next(). The first time, don't call iterator.Next().
if r.tfrNext != nil {
r.iterator.Next()
} else {
r.tfrNext = &TermFrequencyRow{}
}
key, val, valid := r.iterator.Current()
if valid {
tfr := r.tfrNext
err := tfr.parseKDoc(key, r.term)
if err != nil {
return nil, err
}
err = tfr.parseV(val)
if err != nil {
return nil, err
}
rv := preAlloced
if rv == nil {
rv = &index.TermFieldDoc{}
}
rv.ID = append(rv.ID, tfr.doc...)
rv.Freq = tfr.freq
rv.Norm = float64(tfr.norm)
if tfr.vectors != nil {
rv.Vectors = r.indexReader.index.termFieldVectorsFromTermVectors(tfr.vectors)
}
return rv, nil
}
}
return nil, nil
}
func (r *UpsideDownCouchTermFieldReader) Advance(docID index.IndexInternalID, preAlloced *index.TermFieldDoc) (rv *index.TermFieldDoc, err error) {
if r.iterator != nil {
if r.tfrNext == nil {
r.tfrNext = &TermFrequencyRow{}
}
tfr := InitTermFrequencyRow(r.tfrNext, r.term, r.field, docID, 0, 0)
r.keyBuf, err = tfr.KeyAppendTo(r.keyBuf[:0])
if err != nil {
return nil, err
}
r.iterator.Seek(r.keyBuf)
key, val, valid := r.iterator.Current()
if valid {
err := tfr.parseKDoc(key, r.term)
if err != nil {
return nil, err
}
err = tfr.parseV(val)
if err != nil {
return nil, err
}
rv = preAlloced
if rv == nil {
rv = &index.TermFieldDoc{}
}
rv.ID = append(rv.ID, tfr.doc...)
rv.Freq = tfr.freq
rv.Norm = float64(tfr.norm)
if tfr.vectors != nil {
rv.Vectors = r.indexReader.index.termFieldVectorsFromTermVectors(tfr.vectors)
}
return rv, nil
}
}
return nil, nil
}
func (r *UpsideDownCouchTermFieldReader) Close() error {
if r.indexReader != nil {
atomic.AddUint64(&r.indexReader.index.stats.termSearchersFinished, uint64(1))
}
if r.iterator != nil {
return r.iterator.Close()
}
return nil
}
type UpsideDownCouchDocIDReader struct {
indexReader *IndexReader
iterator store.KVIterator
only []string
onlyPos int
onlyMode bool
}
func newUpsideDownCouchDocIDReader(indexReader *IndexReader) (*UpsideDownCouchDocIDReader, error) {
startBytes := []byte{0x0}
endBytes := []byte{0xff}
bisr := NewBackIndexRow(startBytes, nil, nil)
bier := NewBackIndexRow(endBytes, nil, nil)
it := indexReader.kvreader.RangeIterator(bisr.Key(), bier.Key())
return &UpsideDownCouchDocIDReader{
indexReader: indexReader,
iterator: it,
}, nil
}
func newUpsideDownCouchDocIDReaderOnly(indexReader *IndexReader, ids []string) (*UpsideDownCouchDocIDReader, error) {
// ensure ids are sorted
sort.Strings(ids)
startBytes := []byte{0x0}
if len(ids) > 0 {
startBytes = []byte(ids[0])
}
endBytes := []byte{0xff}
if len(ids) > 0 {
endBytes = incrementBytes([]byte(ids[len(ids)-1]))
}
bisr := NewBackIndexRow(startBytes, nil, nil)
bier := NewBackIndexRow(endBytes, nil, nil)
it := indexReader.kvreader.RangeIterator(bisr.Key(), bier.Key())
return &UpsideDownCouchDocIDReader{
indexReader: indexReader,
iterator: it,
only: ids,
onlyMode: true,
}, nil
}
func (r *UpsideDownCouchDocIDReader) Next() (index.IndexInternalID, error) {
key, val, valid := r.iterator.Current()
if r.onlyMode {
var rv index.IndexInternalID
for valid && r.onlyPos < len(r.only) {
br, err := NewBackIndexRowKV(key, val)
if err != nil {
return nil, err
}
if !bytes.Equal(br.doc, []byte(r.only[r.onlyPos])) {
ok := r.nextOnly()
if !ok {
return nil, nil
}
r.iterator.Seek(NewBackIndexRow([]byte(r.only[r.onlyPos]), nil, nil).Key())
key, val, valid = r.iterator.Current()
continue
} else {
rv = append([]byte(nil), br.doc...)
break
}
}
if valid && r.onlyPos < len(r.only) {
ok := r.nextOnly()
if ok {
r.iterator.Seek(NewBackIndexRow([]byte(r.only[r.onlyPos]), nil, nil).Key())
}
return rv, nil
}
} else {
if valid {
br, err := NewBackIndexRowKV(key, val)
if err != nil {
return nil, err
}
rv := append([]byte(nil), br.doc...)
r.iterator.Next()
return rv, nil
}
}
return nil, nil
}
func (r *UpsideDownCouchDocIDReader) Advance(docID index.IndexInternalID) (index.IndexInternalID, error) {
if r.onlyMode {
r.onlyPos = sort.SearchStrings(r.only, string(docID))
if r.onlyPos >= len(r.only) {
// advanced to key after our last only key
return nil, nil
}
r.iterator.Seek(NewBackIndexRow([]byte(r.only[r.onlyPos]), nil, nil).Key())
key, val, valid := r.iterator.Current()
var rv index.IndexInternalID
for valid && r.onlyPos < len(r.only) {
br, err := NewBackIndexRowKV(key, val)
if err != nil {
return nil, err
}
if !bytes.Equal(br.doc, []byte(r.only[r.onlyPos])) {
// the only key we seek'd to didn't exist
// now look for the closest key that did exist in only
r.onlyPos = sort.SearchStrings(r.only, string(br.doc))
if r.onlyPos >= len(r.only) {
// advanced to key after our last only key
return nil, nil
}
// now seek to this new only key
r.iterator.Seek(NewBackIndexRow([]byte(r.only[r.onlyPos]), nil, nil).Key())
key, val, valid = r.iterator.Current()
continue
} else {
rv = append([]byte(nil), br.doc...)
break
}
}
if valid && r.onlyPos < len(r.only) {
ok := r.nextOnly()
if ok {
r.iterator.Seek(NewBackIndexRow([]byte(r.only[r.onlyPos]), nil, nil).Key())
}
return rv, nil
}
} else {
bir := NewBackIndexRow(docID, nil, nil)
r.iterator.Seek(bir.Key())
key, val, valid := r.iterator.Current()
if valid {
br, err := NewBackIndexRowKV(key, val)
if err != nil {
return nil, err
}
rv := append([]byte(nil), br.doc...)
r.iterator.Next()
return rv, nil
}
}
return nil, nil
}
func (r *UpsideDownCouchDocIDReader) Close() error {
return r.iterator.Close()
}
// move the r.only pos forward one, skipping duplicates
// return true if there is more data, or false if we got to the end of the list
func (r *UpsideDownCouchDocIDReader) nextOnly() bool {
// advance 1 position, until we see a different key
// it's already sorted, so this skips duplicates
start := r.onlyPos
r.onlyPos++
for r.onlyPos < len(r.only) && r.only[r.onlyPos] == r.only[start] {
start = r.onlyPos
r.onlyPos++
}
// inidicate if we got to the end of the list
return r.onlyPos < len(r.only)
}

@ -0,0 +1,853 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package upsidedown
import (
"bytes"
"encoding/binary"
"fmt"
"io"
"math"
"github.com/golang/protobuf/proto"
)
const ByteSeparator byte = 0xff
type UpsideDownCouchRowStream chan UpsideDownCouchRow
type UpsideDownCouchRow interface {
KeySize() int
KeyTo([]byte) (int, error)
Key() []byte
Value() []byte
ValueSize() int
ValueTo([]byte) (int, error)
}
func ParseFromKeyValue(key, value []byte) (UpsideDownCouchRow, error) {
if len(key) > 0 {
switch key[0] {
case 'v':
return NewVersionRowKV(key, value)
case 'f':
return NewFieldRowKV(key, value)
case 'd':
return NewDictionaryRowKV(key, value)
case 't':
return NewTermFrequencyRowKV(key, value)
case 'b':
return NewBackIndexRowKV(key, value)
case 's':
return NewStoredRowKV(key, value)
case 'i':
return NewInternalRowKV(key, value)
}
return nil, fmt.Errorf("Unknown field type '%s'", string(key[0]))
}
return nil, fmt.Errorf("Invalid empty key")
}
// VERSION
type VersionRow struct {
version uint8
}
func (v *VersionRow) Key() []byte {
return []byte{'v'}
}
func (v *VersionRow) KeySize() int {
return 1
}
func (v *VersionRow) KeyTo(buf []byte) (int, error) {
buf[0] = 'v'
return 1, nil
}
func (v *VersionRow) Value() []byte {
return []byte{byte(v.version)}
}
func (v *VersionRow) ValueSize() int {
return 1
}
func (v *VersionRow) ValueTo(buf []byte) (int, error) {
buf[0] = v.version
return 1, nil
}
func (v *VersionRow) String() string {
return fmt.Sprintf("Version: %d", v.version)
}
func NewVersionRow(version uint8) *VersionRow {
return &VersionRow{
version: version,
}
}
func NewVersionRowKV(key, value []byte) (*VersionRow, error) {
rv := VersionRow{}
buf := bytes.NewBuffer(value)
err := binary.Read(buf, binary.LittleEndian, &rv.version)
if err != nil {
return nil, err
}
return &rv, nil
}
// INTERNAL STORAGE
type InternalRow struct {
key []byte
val []byte
}
func (i *InternalRow) Key() []byte {
buf := make([]byte, i.KeySize())
size, _ := i.KeyTo(buf)
return buf[:size]
}
func (i *InternalRow) KeySize() int {
return len(i.key) + 1
}
func (i *InternalRow) KeyTo(buf []byte) (int, error) {
buf[0] = 'i'
actual := copy(buf[1:], i.key)
return 1 + actual, nil
}
func (i *InternalRow) Value() []byte {
return i.val
}
func (i *InternalRow) ValueSize() int {
return len(i.val)
}
func (i *InternalRow) ValueTo(buf []byte) (int, error) {
actual := copy(buf, i.val)
return actual, nil
}
func (i *InternalRow) String() string {
return fmt.Sprintf("InternalStore - Key: %s (% x) Val: %s (% x)", i.key, i.key, i.val, i.val)
}
func NewInternalRow(key, val []byte) *InternalRow {
return &InternalRow{
key: key,
val: val,
}
}
func NewInternalRowKV(key, value []byte) (*InternalRow, error) {
rv := InternalRow{}
rv.key = key[1:]
rv.val = value
return &rv, nil
}
// FIELD definition
type FieldRow struct {
index uint16
name string
}
func (f *FieldRow) Key() []byte {
buf := make([]byte, f.KeySize())
size, _ := f.KeyTo(buf)
return buf[:size]
}
func (f *FieldRow) KeySize() int {
return 3
}
func (f *FieldRow) KeyTo(buf []byte) (int, error) {
buf[0] = 'f'
binary.LittleEndian.PutUint16(buf[1:3], f.index)
return 3, nil
}
func (f *FieldRow) Value() []byte {
return append([]byte(f.name), ByteSeparator)
}
func (f *FieldRow) ValueSize() int {
return len(f.name) + 1
}
func (f *FieldRow) ValueTo(buf []byte) (int, error) {
size := copy(buf, f.name)
buf[size] = ByteSeparator
return size + 1, nil
}
func (f *FieldRow) String() string {
return fmt.Sprintf("Field: %d Name: %s", f.index, f.name)
}
func NewFieldRow(index uint16, name string) *FieldRow {
return &FieldRow{
index: index,
name: name,
}
}
func NewFieldRowKV(key, value []byte) (*FieldRow, error) {
rv := FieldRow{}
buf := bytes.NewBuffer(key)
_, err := buf.ReadByte() // type
if err != nil {
return nil, err
}
err = binary.Read(buf, binary.LittleEndian, &rv.index)
if err != nil {
return nil, err
}
buf = bytes.NewBuffer(value)
rv.name, err = buf.ReadString(ByteSeparator)
if err != nil {
return nil, err
}
rv.name = rv.name[:len(rv.name)-1] // trim off separator byte
return &rv, nil
}
// DICTIONARY
const DictionaryRowMaxValueSize = binary.MaxVarintLen64
type DictionaryRow struct {
term []byte
count uint64
field uint16
}
func (dr *DictionaryRow) Key() []byte {
buf := make([]byte, dr.KeySize())
size, _ := dr.KeyTo(buf)
return buf[:size]
}
func (dr *DictionaryRow) KeySize() int {
return len(dr.term) + 3
}
func (dr *DictionaryRow) KeyTo(buf []byte) (int, error) {
buf[0] = 'd'
binary.LittleEndian.PutUint16(buf[1:3], dr.field)
size := copy(buf[3:], dr.term)
return size + 3, nil
}
func (dr *DictionaryRow) Value() []byte {
buf := make([]byte, dr.ValueSize())
size, _ := dr.ValueTo(buf)
return buf[:size]
}
func (dr *DictionaryRow) ValueSize() int {
return DictionaryRowMaxValueSize
}
func (dr *DictionaryRow) ValueTo(buf []byte) (int, error) {
used := binary.PutUvarint(buf, dr.count)
return used, nil
}
func (dr *DictionaryRow) String() string {
return fmt.Sprintf("Dictionary Term: `%s` Field: %d Count: %d ", string(dr.term), dr.field, dr.count)
}
func NewDictionaryRow(term []byte, field uint16, count uint64) *DictionaryRow {
return &DictionaryRow{
term: term,
field: field,
count: count,
}
}
func NewDictionaryRowKV(key, value []byte) (*DictionaryRow, error) {
rv, err := NewDictionaryRowK(key)
if err != nil {
return nil, err
}
err = rv.parseDictionaryV(value)
if err != nil {
return nil, err
}
return rv, nil
}
func NewDictionaryRowK(key []byte) (*DictionaryRow, error) {
rv := &DictionaryRow{}
err := rv.parseDictionaryK(key)
if err != nil {
return nil, err
}
return rv, nil
}
func (dr *DictionaryRow) parseDictionaryK(key []byte) error {
dr.field = binary.LittleEndian.Uint16(key[1:3])
if dr.term != nil {
dr.term = dr.term[:0]
}
dr.term = append(dr.term, key[3:]...)
return nil
}
func (dr *DictionaryRow) parseDictionaryV(value []byte) error {
count, nread := binary.Uvarint(value)
if nread <= 0 {
return fmt.Errorf("DictionaryRow parse Uvarint error, nread: %d", nread)
}
dr.count = count
return nil
}
// TERM FIELD FREQUENCY
type TermVector struct {
field uint16
arrayPositions []uint64
pos uint64
start uint64
end uint64
}
func (tv *TermVector) String() string {
return fmt.Sprintf("Field: %d Pos: %d Start: %d End %d ArrayPositions: %#v", tv.field, tv.pos, tv.start, tv.end, tv.arrayPositions)
}
type TermFrequencyRow struct {
term []byte
doc []byte
freq uint64
vectors []*TermVector
norm float32
field uint16
}
func (tfr *TermFrequencyRow) Term() []byte {
return tfr.term
}
func (tfr *TermFrequencyRow) Freq() uint64 {
return tfr.freq
}
func (tfr *TermFrequencyRow) ScanPrefixForField() []byte {
buf := make([]byte, 3)
buf[0] = 't'
binary.LittleEndian.PutUint16(buf[1:3], tfr.field)
return buf
}
func (tfr *TermFrequencyRow) ScanPrefixForFieldTermPrefix() []byte {
buf := make([]byte, 3+len(tfr.term))
buf[0] = 't'
binary.LittleEndian.PutUint16(buf[1:3], tfr.field)
copy(buf[3:], tfr.term)
return buf
}
func (tfr *TermFrequencyRow) ScanPrefixForFieldTerm() []byte {
buf := make([]byte, 3+len(tfr.term)+1)
buf[0] = 't'
binary.LittleEndian.PutUint16(buf[1:3], tfr.field)
termLen := copy(buf[3:], tfr.term)
buf[3+termLen] = ByteSeparator
return buf
}
func (tfr *TermFrequencyRow) Key() []byte {
buf := make([]byte, tfr.KeySize())
size, _ := tfr.KeyTo(buf)
return buf[:size]
}
func (tfr *TermFrequencyRow) KeySize() int {
return 3 + len(tfr.term) + 1 + len(tfr.doc)
}
func (tfr *TermFrequencyRow) KeyTo(buf []byte) (int, error) {
buf[0] = 't'
binary.LittleEndian.PutUint16(buf[1:3], tfr.field)
termLen := copy(buf[3:], tfr.term)
buf[3+termLen] = ByteSeparator
docLen := copy(buf[3+termLen+1:], tfr.doc)
return 3 + termLen + 1 + docLen, nil
}
func (tfr *TermFrequencyRow) KeyAppendTo(buf []byte) ([]byte, error) {
keySize := tfr.KeySize()
if cap(buf) < keySize {
buf = make([]byte, keySize)
}
actualSize, err := tfr.KeyTo(buf[0:keySize])
return buf[0:actualSize], err
}
func (tfr *TermFrequencyRow) DictionaryRowKey() []byte {
dr := NewDictionaryRow(tfr.term, tfr.field, 0)
return dr.Key()
}
func (tfr *TermFrequencyRow) DictionaryRowKeySize() int {
dr := NewDictionaryRow(tfr.term, tfr.field, 0)
return dr.KeySize()
}
func (tfr *TermFrequencyRow) DictionaryRowKeyTo(buf []byte) (int, error) {
dr := NewDictionaryRow(tfr.term, tfr.field, 0)
return dr.KeyTo(buf)
}
func (tfr *TermFrequencyRow) Value() []byte {
buf := make([]byte, tfr.ValueSize())
size, _ := tfr.ValueTo(buf)
return buf[:size]
}
func (tfr *TermFrequencyRow) ValueSize() int {
bufLen := binary.MaxVarintLen64 + binary.MaxVarintLen64
for _, vector := range tfr.vectors {
bufLen += (binary.MaxVarintLen64 * 4) + (1+len(vector.arrayPositions))*binary.MaxVarintLen64
}
return bufLen
}
func (tfr *TermFrequencyRow) ValueTo(buf []byte) (int, error) {
used := binary.PutUvarint(buf[:binary.MaxVarintLen64], tfr.freq)
normuint32 := math.Float32bits(tfr.norm)
newbuf := buf[used : used+binary.MaxVarintLen64]
used += binary.PutUvarint(newbuf, uint64(normuint32))
for _, vector := range tfr.vectors {
used += binary.PutUvarint(buf[used:used+binary.MaxVarintLen64], uint64(vector.field))
used += binary.PutUvarint(buf[used:used+binary.MaxVarintLen64], vector.pos)
used += binary.PutUvarint(buf[used:used+binary.MaxVarintLen64], vector.start)
used += binary.PutUvarint(buf[used:used+binary.MaxVarintLen64], vector.end)
used += binary.PutUvarint(buf[used:used+binary.MaxVarintLen64], uint64(len(vector.arrayPositions)))
for _, arrayPosition := range vector.arrayPositions {
used += binary.PutUvarint(buf[used:used+binary.MaxVarintLen64], arrayPosition)
}
}
return used, nil
}
func (tfr *TermFrequencyRow) String() string {
return fmt.Sprintf("Term: `%s` Field: %d DocId: `%s` Frequency: %d Norm: %f Vectors: %v", string(tfr.term), tfr.field, string(tfr.doc), tfr.freq, tfr.norm, tfr.vectors)
}
func InitTermFrequencyRow(tfr *TermFrequencyRow, term []byte, field uint16, docID []byte, freq uint64, norm float32) *TermFrequencyRow {
tfr.term = term
tfr.field = field
tfr.doc = docID
tfr.freq = freq
tfr.norm = norm
return tfr
}
func NewTermFrequencyRow(term []byte, field uint16, docID []byte, freq uint64, norm float32) *TermFrequencyRow {
return &TermFrequencyRow{
term: term,
field: field,
doc: docID,
freq: freq,
norm: norm,
}
}
func NewTermFrequencyRowWithTermVectors(term []byte, field uint16, docID []byte, freq uint64, norm float32, vectors []*TermVector) *TermFrequencyRow {
return &TermFrequencyRow{
term: term,
field: field,
doc: docID,
freq: freq,
norm: norm,
vectors: vectors,
}
}
func NewTermFrequencyRowK(key []byte) (*TermFrequencyRow, error) {
rv := &TermFrequencyRow{}
err := rv.parseK(key)
if err != nil {
return nil, err
}
return rv, nil
}
func (tfr *TermFrequencyRow) parseK(key []byte) error {
keyLen := len(key)
if keyLen < 3 {
return fmt.Errorf("invalid term frequency key, no valid field")
}
tfr.field = binary.LittleEndian.Uint16(key[1:3])
termEndPos := bytes.IndexByte(key[3:], ByteSeparator)
if termEndPos < 0 {
return fmt.Errorf("invalid term frequency key, no byte separator terminating term")
}
tfr.term = key[3 : 3+termEndPos]
docLen := keyLen - (3 + termEndPos + 1)
if docLen < 1 {
return fmt.Errorf("invalid term frequency key, empty docid")
}
tfr.doc = key[3+termEndPos+1:]
return nil
}
func (tfr *TermFrequencyRow) parseKDoc(key []byte, term []byte) error {
tfr.doc = key[3+len(term)+1:]
if len(tfr.doc) <= 0 {
return fmt.Errorf("invalid term frequency key, empty docid")
}
return nil
}
func (tfr *TermFrequencyRow) parseV(value []byte) error {
var bytesRead int
tfr.freq, bytesRead = binary.Uvarint(value)
if bytesRead <= 0 {
return fmt.Errorf("invalid term frequency value, invalid frequency")
}
currOffset := bytesRead
var norm uint64
norm, bytesRead = binary.Uvarint(value[currOffset:])
if bytesRead <= 0 {
return fmt.Errorf("invalid term frequency value, no norm")
}
currOffset += bytesRead
tfr.norm = math.Float32frombits(uint32(norm))
tfr.vectors = nil
var field uint64
field, bytesRead = binary.Uvarint(value[currOffset:])
for bytesRead > 0 {
currOffset += bytesRead
tv := TermVector{}
tv.field = uint16(field)
// at this point we expect at least one term vector
if tfr.vectors == nil {
tfr.vectors = make([]*TermVector, 0)
}
tv.pos, bytesRead = binary.Uvarint(value[currOffset:])
if bytesRead <= 0 {
return fmt.Errorf("invalid term frequency value, vector contains no position")
}
currOffset += bytesRead
tv.start, bytesRead = binary.Uvarint(value[currOffset:])
if bytesRead <= 0 {
return fmt.Errorf("invalid term frequency value, vector contains no start")
}
currOffset += bytesRead
tv.end, bytesRead = binary.Uvarint(value[currOffset:])
if bytesRead <= 0 {
return fmt.Errorf("invalid term frequency value, vector contains no end")
}
currOffset += bytesRead
var arrayPositionsLen uint64 = 0
arrayPositionsLen, bytesRead = binary.Uvarint(value[currOffset:])
if bytesRead <= 0 {
return fmt.Errorf("invalid term frequency value, vector contains no arrayPositionLen")
}
currOffset += bytesRead
if arrayPositionsLen > 0 {
tv.arrayPositions = make([]uint64, arrayPositionsLen)
for i := 0; uint64(i) < arrayPositionsLen; i++ {
tv.arrayPositions[i], bytesRead = binary.Uvarint(value[currOffset:])
if bytesRead <= 0 {
return fmt.Errorf("invalid term frequency value, vector contains no arrayPosition of index %d", i)
}
currOffset += bytesRead
}
}
tfr.vectors = append(tfr.vectors, &tv)
// try to read next record (may not exist)
field, bytesRead = binary.Uvarint(value[currOffset:])
}
if len(value[currOffset:]) > 0 && bytesRead <= 0 {
return fmt.Errorf("invalid term frequency value, vector field invalid")
}
return nil
}
func NewTermFrequencyRowKV(key, value []byte) (*TermFrequencyRow, error) {
rv, err := NewTermFrequencyRowK(key)
if err != nil {
return nil, err
}
err = rv.parseV(value)
if err != nil {
return nil, err
}
return rv, nil
}
type BackIndexRow struct {
doc []byte
termEntries []*BackIndexTermEntry
storedEntries []*BackIndexStoreEntry
}
func (br *BackIndexRow) AllTermKeys() [][]byte {
if br == nil {
return nil
}
rv := make([][]byte, len(br.termEntries))
for i, termEntry := range br.termEntries {
termRow := NewTermFrequencyRow([]byte(termEntry.GetTerm()), uint16(termEntry.GetField()), br.doc, 0, 0)
rv[i] = termRow.Key()
}
return rv
}
func (br *BackIndexRow) AllStoredKeys() [][]byte {
if br == nil {
return nil
}
rv := make([][]byte, len(br.storedEntries))
for i, storedEntry := range br.storedEntries {
storedRow := NewStoredRow(br.doc, uint16(storedEntry.GetField()), storedEntry.GetArrayPositions(), 'x', []byte{})
rv[i] = storedRow.Key()
}
return rv
}
func (br *BackIndexRow) Key() []byte {
buf := make([]byte, br.KeySize())
size, _ := br.KeyTo(buf)
return buf[:size]
}
func (br *BackIndexRow) KeySize() int {
return len(br.doc) + 1
}
func (br *BackIndexRow) KeyTo(buf []byte) (int, error) {
buf[0] = 'b'
used := copy(buf[1:], br.doc)
return used + 1, nil
}
func (br *BackIndexRow) Value() []byte {
buf := make([]byte, br.ValueSize())
size, _ := br.ValueTo(buf)
return buf[:size]
}
func (br *BackIndexRow) ValueSize() int {
birv := &BackIndexRowValue{
TermEntries: br.termEntries,
StoredEntries: br.storedEntries,
}
return birv.Size()
}
func (br *BackIndexRow) ValueTo(buf []byte) (int, error) {
birv := &BackIndexRowValue{
TermEntries: br.termEntries,
StoredEntries: br.storedEntries,
}
return birv.MarshalTo(buf)
}
func (br *BackIndexRow) String() string {
return fmt.Sprintf("Backindex DocId: `%s` Term Entries: %v, Stored Entries: %v", string(br.doc), br.termEntries, br.storedEntries)
}
func NewBackIndexRow(docID []byte, entries []*BackIndexTermEntry, storedFields []*BackIndexStoreEntry) *BackIndexRow {
return &BackIndexRow{
doc: docID,
termEntries: entries,
storedEntries: storedFields,
}
}
func NewBackIndexRowKV(key, value []byte) (*BackIndexRow, error) {
rv := BackIndexRow{}
buf := bytes.NewBuffer(key)
_, err := buf.ReadByte() // type
if err != nil {
return nil, err
}
rv.doc, err = buf.ReadBytes(ByteSeparator)
if err == io.EOF && len(rv.doc) < 1 {
err = fmt.Errorf("invalid doc length 0 - % x", key)
}
if err != nil && err != io.EOF {
return nil, err
} else if err == nil {
rv.doc = rv.doc[:len(rv.doc)-1] // trim off separator byte
}
var birv BackIndexRowValue
err = proto.Unmarshal(value, &birv)
if err != nil {
return nil, err
}
rv.termEntries = birv.TermEntries
rv.storedEntries = birv.StoredEntries
return &rv, nil
}
// STORED
type StoredRow struct {
doc []byte
field uint16
arrayPositions []uint64
typ byte
value []byte
}
func (s *StoredRow) Key() []byte {
buf := make([]byte, s.KeySize())
size, _ := s.KeyTo(buf)
return buf[0:size]
}
func (s *StoredRow) KeySize() int {
return 1 + len(s.doc) + 1 + 2 + (binary.MaxVarintLen64 * len(s.arrayPositions))
}
func (s *StoredRow) KeyTo(buf []byte) (int, error) {
docLen := len(s.doc)
buf[0] = 's'
copy(buf[1:], s.doc)
buf[1+docLen] = ByteSeparator
binary.LittleEndian.PutUint16(buf[1+docLen+1:], s.field)
bytesUsed := 1 + docLen + 1 + 2
for _, arrayPosition := range s.arrayPositions {
varbytes := binary.PutUvarint(buf[bytesUsed:], arrayPosition)
bytesUsed += varbytes
}
return bytesUsed, nil
}
func (s *StoredRow) Value() []byte {
buf := make([]byte, s.ValueSize())
size, _ := s.ValueTo(buf)
return buf[:size]
}
func (s *StoredRow) ValueSize() int {
return len(s.value) + 1
}
func (s *StoredRow) ValueTo(buf []byte) (int, error) {
buf[0] = s.typ
used := copy(buf[1:], s.value)
return used + 1, nil
}
func (s *StoredRow) String() string {
return fmt.Sprintf("Document: %s Field %d, Array Positions: %v, Type: %s Value: %s", s.doc, s.field, s.arrayPositions, string(s.typ), s.value)
}
func (s *StoredRow) ScanPrefixForDoc() []byte {
docLen := len(s.doc)
buf := make([]byte, 1+docLen+1)
buf[0] = 's'
copy(buf[1:], s.doc)
buf[1+docLen] = ByteSeparator
return buf
}
func NewStoredRow(docID []byte, field uint16, arrayPositions []uint64, typ byte, value []byte) *StoredRow {
return &StoredRow{
doc: docID,
field: field,
arrayPositions: arrayPositions,
typ: typ,
value: value,
}
}
func NewStoredRowK(key []byte) (*StoredRow, error) {
rv := StoredRow{}
buf := bytes.NewBuffer(key)
_, err := buf.ReadByte() // type
if err != nil {
return nil, err
}
rv.doc, err = buf.ReadBytes(ByteSeparator)
if len(rv.doc) < 2 { // 1 for min doc id length, 1 for separator
err = fmt.Errorf("invalid doc length 0")
return nil, err
}
rv.doc = rv.doc[:len(rv.doc)-1] // trim off separator byte
err = binary.Read(buf, binary.LittleEndian, &rv.field)
if err != nil {
return nil, err
}
rv.arrayPositions = make([]uint64, 0)
nextArrayPos, err := binary.ReadUvarint(buf)
for err == nil {
rv.arrayPositions = append(rv.arrayPositions, nextArrayPos)
nextArrayPos, err = binary.ReadUvarint(buf)
}
return &rv, nil
}
func NewStoredRowKV(key, value []byte) (*StoredRow, error) {
rv, err := NewStoredRowK(key)
if err != nil {
return nil, err
}
rv.typ = value[0]
rv.value = value[1:]
return rv, nil
}

@ -0,0 +1,76 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package upsidedown
import (
"encoding/binary"
)
var mergeOperator upsideDownMerge
var dictionaryTermIncr []byte
var dictionaryTermDecr []byte
func init() {
dictionaryTermIncr = make([]byte, 8)
binary.LittleEndian.PutUint64(dictionaryTermIncr, uint64(1))
dictionaryTermDecr = make([]byte, 8)
var negOne = int64(-1)
binary.LittleEndian.PutUint64(dictionaryTermDecr, uint64(negOne))
}
type upsideDownMerge struct{}
func (m *upsideDownMerge) FullMerge(key, existingValue []byte, operands [][]byte) ([]byte, bool) {
// set up record based on key
dr, err := NewDictionaryRowK(key)
if err != nil {
return nil, false
}
if len(existingValue) > 0 {
// if existing value, parse it
err = dr.parseDictionaryV(existingValue)
if err != nil {
return nil, false
}
}
// now process operands
for _, operand := range operands {
next := int64(binary.LittleEndian.Uint64(operand))
if next < 0 && uint64(-next) > dr.count {
// subtracting next from existing would overflow
dr.count = 0
} else if next < 0 {
dr.count -= uint64(-next)
} else {
dr.count += uint64(next)
}
}
return dr.Value(), true
}
func (m *upsideDownMerge) PartialMerge(key, leftOperand, rightOperand []byte) ([]byte, bool) {
left := int64(binary.LittleEndian.Uint64(leftOperand))
right := int64(binary.LittleEndian.Uint64(rightOperand))
rv := make([]byte, 8)
binary.LittleEndian.PutUint64(rv, uint64(left+right))
return rv, true
}
func (m *upsideDownMerge) Name() string {
return "upsideDownMerge"
}

@ -0,0 +1,55 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package upsidedown
import (
"encoding/json"
"sync/atomic"
"github.com/blevesearch/bleve/index/store"
)
type indexStat struct {
updates, deletes, batches, errors uint64
analysisTime, indexTime uint64
termSearchersStarted uint64
termSearchersFinished uint64
numPlainTextBytesIndexed uint64
i *UpsideDownCouch
}
func (i *indexStat) statsMap() map[string]interface{} {
m := map[string]interface{}{}
m["updates"] = atomic.LoadUint64(&i.updates)
m["deletes"] = atomic.LoadUint64(&i.deletes)
m["batches"] = atomic.LoadUint64(&i.batches)
m["errors"] = atomic.LoadUint64(&i.errors)
m["analysis_time"] = atomic.LoadUint64(&i.analysisTime)
m["index_time"] = atomic.LoadUint64(&i.indexTime)
m["term_searchers_started"] = atomic.LoadUint64(&i.termSearchersStarted)
m["term_searchers_finished"] = atomic.LoadUint64(&i.termSearchersFinished)
m["num_plain_text_bytes_indexed"] = atomic.LoadUint64(&i.numPlainTextBytesIndexed)
if o, ok := i.i.store.(store.KVStoreStats); ok {
m["kv"] = o.StatsMap()
}
return m
}
func (i *indexStat) MarshalJSON() ([]byte, error) {
m := i.statsMap()
return json.Marshal(m)
}

File diff suppressed because it is too large Load Diff

@ -0,0 +1,684 @@
// Code generated by protoc-gen-gogo.
// source: upsidedown.proto
// DO NOT EDIT!
/*
Package upsidedown is a generated protocol buffer package.
It is generated from these files:
upsidedown.proto
It has these top-level messages:
BackIndexTermEntry
BackIndexStoreEntry
BackIndexRowValue
*/
package upsidedown
import proto "github.com/golang/protobuf/proto"
import math "math"
import io "io"
import fmt "fmt"
import github_com_golang_protobuf_proto "github.com/golang/protobuf/proto"
// Reference imports to suppress errors if they are not otherwise used.
var _ = proto.Marshal
var _ = math.Inf
type BackIndexTermEntry struct {
Term *string `protobuf:"bytes,1,req,name=term" json:"term,omitempty"`
Field *uint32 `protobuf:"varint,2,req,name=field" json:"field,omitempty"`
XXX_unrecognized []byte `json:"-"`
}
func (m *BackIndexTermEntry) Reset() { *m = BackIndexTermEntry{} }
func (m *BackIndexTermEntry) String() string { return proto.CompactTextString(m) }
func (*BackIndexTermEntry) ProtoMessage() {}
func (m *BackIndexTermEntry) GetTerm() string {
if m != nil && m.Term != nil {
return *m.Term
}
return ""
}
func (m *BackIndexTermEntry) GetField() uint32 {
if m != nil && m.Field != nil {
return *m.Field
}
return 0
}
type BackIndexStoreEntry struct {
Field *uint32 `protobuf:"varint,1,req,name=field" json:"field,omitempty"`
ArrayPositions []uint64 `protobuf:"varint,2,rep,name=arrayPositions" json:"arrayPositions,omitempty"`
XXX_unrecognized []byte `json:"-"`
}
func (m *BackIndexStoreEntry) Reset() { *m = BackIndexStoreEntry{} }
func (m *BackIndexStoreEntry) String() string { return proto.CompactTextString(m) }
func (*BackIndexStoreEntry) ProtoMessage() {}
func (m *BackIndexStoreEntry) GetField() uint32 {
if m != nil && m.Field != nil {
return *m.Field
}
return 0
}
func (m *BackIndexStoreEntry) GetArrayPositions() []uint64 {
if m != nil {
return m.ArrayPositions
}
return nil
}
type BackIndexRowValue struct {
TermEntries []*BackIndexTermEntry `protobuf:"bytes,1,rep,name=termEntries" json:"termEntries,omitempty"`
StoredEntries []*BackIndexStoreEntry `protobuf:"bytes,2,rep,name=storedEntries" json:"storedEntries,omitempty"`
XXX_unrecognized []byte `json:"-"`
}
func (m *BackIndexRowValue) Reset() { *m = BackIndexRowValue{} }
func (m *BackIndexRowValue) String() string { return proto.CompactTextString(m) }
func (*BackIndexRowValue) ProtoMessage() {}
func (m *BackIndexRowValue) GetTermEntries() []*BackIndexTermEntry {
if m != nil {
return m.TermEntries
}
return nil
}
func (m *BackIndexRowValue) GetStoredEntries() []*BackIndexStoreEntry {
if m != nil {
return m.StoredEntries
}
return nil
}
func (m *BackIndexTermEntry) Unmarshal(data []byte) error {
var hasFields [1]uint64
l := len(data)
iNdEx := 0
for iNdEx < l {
var wire uint64
for shift := uint(0); ; shift += 7 {
if iNdEx >= l {
return io.ErrUnexpectedEOF
}
b := data[iNdEx]
iNdEx++
wire |= (uint64(b) & 0x7F) << shift
if b < 0x80 {
break
}
}
fieldNum := int32(wire >> 3)
wireType := int(wire & 0x7)
switch fieldNum {
case 1:
if wireType != 2 {
return fmt.Errorf("proto: wrong wireType = %d for field Term", wireType)
}
var stringLen uint64
for shift := uint(0); ; shift += 7 {
if iNdEx >= l {
return io.ErrUnexpectedEOF
}
b := data[iNdEx]
iNdEx++
stringLen |= (uint64(b) & 0x7F) << shift
if b < 0x80 {
break
}
}
postIndex := iNdEx + int(stringLen)
if postIndex > l {
return io.ErrUnexpectedEOF
}
s := string(data[iNdEx:postIndex])
m.Term = &s
iNdEx = postIndex
hasFields[0] |= uint64(0x00000001)
case 2:
if wireType != 0 {
return fmt.Errorf("proto: wrong wireType = %d for field Field", wireType)
}
var v uint32
for shift := uint(0); ; shift += 7 {
if iNdEx >= l {
return io.ErrUnexpectedEOF
}
b := data[iNdEx]
iNdEx++
v |= (uint32(b) & 0x7F) << shift
if b < 0x80 {
break
}
}
m.Field = &v
hasFields[0] |= uint64(0x00000002)
default:
var sizeOfWire int
for {
sizeOfWire++
wire >>= 7
if wire == 0 {
break
}
}
iNdEx -= sizeOfWire
skippy, err := skipUpsidedown(data[iNdEx:])
if err != nil {
return err
}
if skippy < 0 {
return ErrInvalidLengthUpsidedown
}
if (iNdEx + skippy) > l {
return io.ErrUnexpectedEOF
}
m.XXX_unrecognized = append(m.XXX_unrecognized, data[iNdEx:iNdEx+skippy]...)
iNdEx += skippy
}
}
if hasFields[0]&uint64(0x00000001) == 0 {
return new(github_com_golang_protobuf_proto.RequiredNotSetError)
}
if hasFields[0]&uint64(0x00000002) == 0 {
return new(github_com_golang_protobuf_proto.RequiredNotSetError)
}
return nil
}
func (m *BackIndexStoreEntry) Unmarshal(data []byte) error {
var hasFields [1]uint64
l := len(data)
iNdEx := 0
for iNdEx < l {
var wire uint64
for shift := uint(0); ; shift += 7 {
if iNdEx >= l {
return io.ErrUnexpectedEOF
}
b := data[iNdEx]
iNdEx++
wire |= (uint64(b) & 0x7F) << shift
if b < 0x80 {
break
}
}
fieldNum := int32(wire >> 3)
wireType := int(wire & 0x7)
switch fieldNum {
case 1:
if wireType != 0 {
return fmt.Errorf("proto: wrong wireType = %d for field Field", wireType)
}
var v uint32
for shift := uint(0); ; shift += 7 {
if iNdEx >= l {
return io.ErrUnexpectedEOF
}
b := data[iNdEx]
iNdEx++
v |= (uint32(b) & 0x7F) << shift
if b < 0x80 {
break
}
}
m.Field = &v
hasFields[0] |= uint64(0x00000001)
case 2:
if wireType != 0 {
return fmt.Errorf("proto: wrong wireType = %d for field ArrayPositions", wireType)
}
var v uint64
for shift := uint(0); ; shift += 7 {
if iNdEx >= l {
return io.ErrUnexpectedEOF
}
b := data[iNdEx]
iNdEx++
v |= (uint64(b) & 0x7F) << shift
if b < 0x80 {
break
}
}
m.ArrayPositions = append(m.ArrayPositions, v)
default:
var sizeOfWire int
for {
sizeOfWire++
wire >>= 7
if wire == 0 {
break
}
}
iNdEx -= sizeOfWire
skippy, err := skipUpsidedown(data[iNdEx:])
if err != nil {
return err
}
if skippy < 0 {
return ErrInvalidLengthUpsidedown
}
if (iNdEx + skippy) > l {
return io.ErrUnexpectedEOF
}
m.XXX_unrecognized = append(m.XXX_unrecognized, data[iNdEx:iNdEx+skippy]...)
iNdEx += skippy
}
}
if hasFields[0]&uint64(0x00000001) == 0 {
return new(github_com_golang_protobuf_proto.RequiredNotSetError)
}
return nil
}
func (m *BackIndexRowValue) Unmarshal(data []byte) error {
l := len(data)
iNdEx := 0
for iNdEx < l {
var wire uint64
for shift := uint(0); ; shift += 7 {
if iNdEx >= l {
return io.ErrUnexpectedEOF
}
b := data[iNdEx]
iNdEx++
wire |= (uint64(b) & 0x7F) << shift
if b < 0x80 {
break
}
}
fieldNum := int32(wire >> 3)
wireType := int(wire & 0x7)
switch fieldNum {
case 1:
if wireType != 2 {
return fmt.Errorf("proto: wrong wireType = %d for field TermEntries", wireType)
}
var msglen int
for shift := uint(0); ; shift += 7 {
if iNdEx >= l {
return io.ErrUnexpectedEOF
}
b := data[iNdEx]
iNdEx++
msglen |= (int(b) & 0x7F) << shift
if b < 0x80 {
break
}
}
postIndex := iNdEx + msglen
if msglen < 0 {
return ErrInvalidLengthUpsidedown
}
if postIndex > l {
return io.ErrUnexpectedEOF
}
m.TermEntries = append(m.TermEntries, &BackIndexTermEntry{})
if err := m.TermEntries[len(m.TermEntries)-1].Unmarshal(data[iNdEx:postIndex]); err != nil {
return err
}
iNdEx = postIndex
case 2:
if wireType != 2 {
return fmt.Errorf("proto: wrong wireType = %d for field StoredEntries", wireType)
}
var msglen int
for shift := uint(0); ; shift += 7 {
if iNdEx >= l {
return io.ErrUnexpectedEOF
}
b := data[iNdEx]
iNdEx++
msglen |= (int(b) & 0x7F) << shift
if b < 0x80 {
break
}
}
postIndex := iNdEx + msglen
if msglen < 0 {
return ErrInvalidLengthUpsidedown
}
if postIndex > l {
return io.ErrUnexpectedEOF
}
m.StoredEntries = append(m.StoredEntries, &BackIndexStoreEntry{})
if err := m.StoredEntries[len(m.StoredEntries)-1].Unmarshal(data[iNdEx:postIndex]); err != nil {
return err
}
iNdEx = postIndex
default:
var sizeOfWire int
for {
sizeOfWire++
wire >>= 7
if wire == 0 {
break
}
}
iNdEx -= sizeOfWire
skippy, err := skipUpsidedown(data[iNdEx:])
if err != nil {
return err
}
if skippy < 0 {
return ErrInvalidLengthUpsidedown
}
if (iNdEx + skippy) > l {
return io.ErrUnexpectedEOF
}
m.XXX_unrecognized = append(m.XXX_unrecognized, data[iNdEx:iNdEx+skippy]...)
iNdEx += skippy
}
}
return nil
}
func skipUpsidedown(data []byte) (n int, err error) {
l := len(data)
iNdEx := 0
for iNdEx < l {
var wire uint64
for shift := uint(0); ; shift += 7 {
if iNdEx >= l {
return 0, io.ErrUnexpectedEOF
}
b := data[iNdEx]
iNdEx++
wire |= (uint64(b) & 0x7F) << shift
if b < 0x80 {
break
}
}
wireType := int(wire & 0x7)
switch wireType {
case 0:
for {
if iNdEx >= l {
return 0, io.ErrUnexpectedEOF
}
iNdEx++
if data[iNdEx-1] < 0x80 {
break
}
}
return iNdEx, nil
case 1:
iNdEx += 8
return iNdEx, nil
case 2:
var length int
for shift := uint(0); ; shift += 7 {
if iNdEx >= l {
return 0, io.ErrUnexpectedEOF
}
b := data[iNdEx]
iNdEx++
length |= (int(b) & 0x7F) << shift
if b < 0x80 {
break
}
}
iNdEx += length
if length < 0 {
return 0, ErrInvalidLengthUpsidedown
}
return iNdEx, nil
case 3:
for {
var innerWire uint64
var start int = iNdEx
for shift := uint(0); ; shift += 7 {
if iNdEx >= l {
return 0, io.ErrUnexpectedEOF
}
b := data[iNdEx]
iNdEx++
innerWire |= (uint64(b) & 0x7F) << shift
if b < 0x80 {
break
}
}
innerWireType := int(innerWire & 0x7)
if innerWireType == 4 {
break
}
next, err := skipUpsidedown(data[start:])
if err != nil {
return 0, err
}
iNdEx = start + next
}
return iNdEx, nil
case 4:
return iNdEx, nil
case 5:
iNdEx += 4
return iNdEx, nil
default:
return 0, fmt.Errorf("proto: illegal wireType %d", wireType)
}
}
panic("unreachable")
}
var (
ErrInvalidLengthUpsidedown = fmt.Errorf("proto: negative length found during unmarshaling")
)
func (m *BackIndexTermEntry) Size() (n int) {
var l int
_ = l
if m.Term != nil {
l = len(*m.Term)
n += 1 + l + sovUpsidedown(uint64(l))
}
if m.Field != nil {
n += 1 + sovUpsidedown(uint64(*m.Field))
}
if m.XXX_unrecognized != nil {
n += len(m.XXX_unrecognized)
}
return n
}
func (m *BackIndexStoreEntry) Size() (n int) {
var l int
_ = l
if m.Field != nil {
n += 1 + sovUpsidedown(uint64(*m.Field))
}
if len(m.ArrayPositions) > 0 {
for _, e := range m.ArrayPositions {
n += 1 + sovUpsidedown(uint64(e))
}
}
if m.XXX_unrecognized != nil {
n += len(m.XXX_unrecognized)
}
return n
}
func (m *BackIndexRowValue) Size() (n int) {
var l int
_ = l
if len(m.TermEntries) > 0 {
for _, e := range m.TermEntries {
l = e.Size()
n += 1 + l + sovUpsidedown(uint64(l))
}
}
if len(m.StoredEntries) > 0 {
for _, e := range m.StoredEntries {
l = e.Size()
n += 1 + l + sovUpsidedown(uint64(l))
}
}
if m.XXX_unrecognized != nil {
n += len(m.XXX_unrecognized)
}
return n
}
func sovUpsidedown(x uint64) (n int) {
for {
n++
x >>= 7
if x == 0 {
break
}
}
return n
}
func sozUpsidedown(x uint64) (n int) {
return sovUpsidedown(uint64((x << 1) ^ uint64((int64(x) >> 63))))
}
func (m *BackIndexTermEntry) Marshal() (data []byte, err error) {
size := m.Size()
data = make([]byte, size)
n, err := m.MarshalTo(data)
if err != nil {
return nil, err
}
return data[:n], nil
}
func (m *BackIndexTermEntry) MarshalTo(data []byte) (n int, err error) {
var i int
_ = i
var l int
_ = l
if m.Term == nil {
return 0, new(github_com_golang_protobuf_proto.RequiredNotSetError)
} else {
data[i] = 0xa
i++
i = encodeVarintUpsidedown(data, i, uint64(len(*m.Term)))
i += copy(data[i:], *m.Term)
}
if m.Field == nil {
return 0, new(github_com_golang_protobuf_proto.RequiredNotSetError)
} else {
data[i] = 0x10
i++
i = encodeVarintUpsidedown(data, i, uint64(*m.Field))
}
if m.XXX_unrecognized != nil {
i += copy(data[i:], m.XXX_unrecognized)
}
return i, nil
}
func (m *BackIndexStoreEntry) Marshal() (data []byte, err error) {
size := m.Size()
data = make([]byte, size)
n, err := m.MarshalTo(data)
if err != nil {
return nil, err
}
return data[:n], nil
}
func (m *BackIndexStoreEntry) MarshalTo(data []byte) (n int, err error) {
var i int
_ = i
var l int
_ = l
if m.Field == nil {
return 0, new(github_com_golang_protobuf_proto.RequiredNotSetError)
} else {
data[i] = 0x8
i++
i = encodeVarintUpsidedown(data, i, uint64(*m.Field))
}
if len(m.ArrayPositions) > 0 {
for _, num := range m.ArrayPositions {
data[i] = 0x10
i++
i = encodeVarintUpsidedown(data, i, uint64(num))
}
}
if m.XXX_unrecognized != nil {
i += copy(data[i:], m.XXX_unrecognized)
}
return i, nil
}
func (m *BackIndexRowValue) Marshal() (data []byte, err error) {
size := m.Size()
data = make([]byte, size)
n, err := m.MarshalTo(data)
if err != nil {
return nil, err
}
return data[:n], nil
}
func (m *BackIndexRowValue) MarshalTo(data []byte) (n int, err error) {
var i int
_ = i
var l int
_ = l
if len(m.TermEntries) > 0 {
for _, msg := range m.TermEntries {
data[i] = 0xa
i++
i = encodeVarintUpsidedown(data, i, uint64(msg.Size()))
n, err := msg.MarshalTo(data[i:])
if err != nil {
return 0, err
}
i += n
}
}
if len(m.StoredEntries) > 0 {
for _, msg := range m.StoredEntries {
data[i] = 0x12
i++
i = encodeVarintUpsidedown(data, i, uint64(msg.Size()))
n, err := msg.MarshalTo(data[i:])
if err != nil {
return 0, err
}
i += n
}
}
if m.XXX_unrecognized != nil {
i += copy(data[i:], m.XXX_unrecognized)
}
return i, nil
}
func encodeFixed64Upsidedown(data []byte, offset int, v uint64) int {
data[offset] = uint8(v)
data[offset+1] = uint8(v >> 8)
data[offset+2] = uint8(v >> 16)
data[offset+3] = uint8(v >> 24)
data[offset+4] = uint8(v >> 32)
data[offset+5] = uint8(v >> 40)
data[offset+6] = uint8(v >> 48)
data[offset+7] = uint8(v >> 56)
return offset + 8
}
func encodeFixed32Upsidedown(data []byte, offset int, v uint32) int {
data[offset] = uint8(v)
data[offset+1] = uint8(v >> 8)
data[offset+2] = uint8(v >> 16)
data[offset+3] = uint8(v >> 24)
return offset + 4
}
func encodeVarintUpsidedown(data []byte, offset int, v uint64) int {
for v >= 1<<7 {
data[offset] = uint8(v&0x7f | 0x80)
v >>= 7
offset++
}
data[offset] = uint8(v)
return offset + 1
}

@ -0,0 +1,14 @@
message BackIndexTermEntry {
required string term = 1;
required uint32 field = 2;
}
message BackIndexStoreEntry {
required uint32 field = 1;
repeated uint64 arrayPositions = 2;
}
message BackIndexRowValue {
repeated BackIndexTermEntry termEntries = 1;
repeated BackIndexStoreEntry storedEntries = 2;
}

@ -0,0 +1,37 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package bleve
// An IndexAlias is a wrapper around one or more
// Index objects. It has two distinct modes of
// operation.
// 1. When it points to a single index, ALL index
// operations are valid and will be passed through
// to the underlying index.
// 2. When it points to more than one index, the only
// valid operation is Search. In this case the
// search will be performed across all the
// underlying indexes and the results merged.
// Calls to Add/Remove/Swap the underlying indexes
// are atomic, so you can safely change the
// underlying Index objects while other components
// are performing operations.
type IndexAlias interface {
Index
Add(i ...Index)
Remove(i ...Index)
Swap(in, out []Index)
}

@ -0,0 +1,605 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package bleve
import (
"sort"
"sync"
"time"
"golang.org/x/net/context"
"github.com/blevesearch/bleve/document"
"github.com/blevesearch/bleve/index"
"github.com/blevesearch/bleve/index/store"
"github.com/blevesearch/bleve/mapping"
"github.com/blevesearch/bleve/search"
)
type indexAliasImpl struct {
name string
indexes []Index
mutex sync.RWMutex
open bool
}
// NewIndexAlias creates a new IndexAlias over the provided
// Index objects.
func NewIndexAlias(indexes ...Index) *indexAliasImpl {
return &indexAliasImpl{
name: "alias",
indexes: indexes,
open: true,
}
}
func (i *indexAliasImpl) isAliasToSingleIndex() error {
if len(i.indexes) < 1 {
return ErrorAliasEmpty
} else if len(i.indexes) > 1 {
return ErrorAliasMulti
}
return nil
}
func (i *indexAliasImpl) Index(id string, data interface{}) error {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return ErrorIndexClosed
}
err := i.isAliasToSingleIndex()
if err != nil {
return err
}
return i.indexes[0].Index(id, data)
}
func (i *indexAliasImpl) Delete(id string) error {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return ErrorIndexClosed
}
err := i.isAliasToSingleIndex()
if err != nil {
return err
}
return i.indexes[0].Delete(id)
}
func (i *indexAliasImpl) Batch(b *Batch) error {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return ErrorIndexClosed
}
err := i.isAliasToSingleIndex()
if err != nil {
return err
}
return i.indexes[0].Batch(b)
}
func (i *indexAliasImpl) Document(id string) (*document.Document, error) {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return nil, ErrorIndexClosed
}
err := i.isAliasToSingleIndex()
if err != nil {
return nil, err
}
return i.indexes[0].Document(id)
}
func (i *indexAliasImpl) DocCount() (uint64, error) {
i.mutex.RLock()
defer i.mutex.RUnlock()
rv := uint64(0)
if !i.open {
return 0, ErrorIndexClosed
}
for _, index := range i.indexes {
otherCount, err := index.DocCount()
if err == nil {
rv += otherCount
}
// tolerate errors to produce partial counts
}
return rv, nil
}
func (i *indexAliasImpl) Search(req *SearchRequest) (*SearchResult, error) {
return i.SearchInContext(context.Background(), req)
}
func (i *indexAliasImpl) SearchInContext(ctx context.Context, req *SearchRequest) (*SearchResult, error) {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return nil, ErrorIndexClosed
}
if len(i.indexes) < 1 {
return nil, ErrorAliasEmpty
}
// short circuit the simple case
if len(i.indexes) == 1 {
return i.indexes[0].SearchInContext(ctx, req)
}
return MultiSearch(ctx, req, i.indexes...)
}
func (i *indexAliasImpl) Fields() ([]string, error) {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return nil, ErrorIndexClosed
}
err := i.isAliasToSingleIndex()
if err != nil {
return nil, err
}
return i.indexes[0].Fields()
}
func (i *indexAliasImpl) FieldDict(field string) (index.FieldDict, error) {
i.mutex.RLock()
if !i.open {
i.mutex.RUnlock()
return nil, ErrorIndexClosed
}
err := i.isAliasToSingleIndex()
if err != nil {
i.mutex.RUnlock()
return nil, err
}
fieldDict, err := i.indexes[0].FieldDict(field)
if err != nil {
i.mutex.RUnlock()
return nil, err
}
return &indexAliasImplFieldDict{
index: i,
fieldDict: fieldDict,
}, nil
}
func (i *indexAliasImpl) FieldDictRange(field string, startTerm []byte, endTerm []byte) (index.FieldDict, error) {
i.mutex.RLock()
if !i.open {
i.mutex.RUnlock()
return nil, ErrorIndexClosed
}
err := i.isAliasToSingleIndex()
if err != nil {
i.mutex.RUnlock()
return nil, err
}
fieldDict, err := i.indexes[0].FieldDictRange(field, startTerm, endTerm)
if err != nil {
i.mutex.RUnlock()
return nil, err
}
return &indexAliasImplFieldDict{
index: i,
fieldDict: fieldDict,
}, nil
}
func (i *indexAliasImpl) FieldDictPrefix(field string, termPrefix []byte) (index.FieldDict, error) {
i.mutex.RLock()
if !i.open {
i.mutex.RUnlock()
return nil, ErrorIndexClosed
}
err := i.isAliasToSingleIndex()
if err != nil {
i.mutex.RUnlock()
return nil, err
}
fieldDict, err := i.indexes[0].FieldDictPrefix(field, termPrefix)
if err != nil {
i.mutex.RUnlock()
return nil, err
}
return &indexAliasImplFieldDict{
index: i,
fieldDict: fieldDict,
}, nil
}
func (i *indexAliasImpl) Close() error {
i.mutex.Lock()
defer i.mutex.Unlock()
i.open = false
return nil
}
func (i *indexAliasImpl) Mapping() mapping.IndexMapping {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return nil
}
err := i.isAliasToSingleIndex()
if err != nil {
return nil
}
return i.indexes[0].Mapping()
}
func (i *indexAliasImpl) Stats() *IndexStat {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return nil
}
err := i.isAliasToSingleIndex()
if err != nil {
return nil
}
return i.indexes[0].Stats()
}
func (i *indexAliasImpl) StatsMap() map[string]interface{} {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return nil
}
err := i.isAliasToSingleIndex()
if err != nil {
return nil
}
return i.indexes[0].StatsMap()
}
func (i *indexAliasImpl) GetInternal(key []byte) ([]byte, error) {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return nil, ErrorIndexClosed
}
err := i.isAliasToSingleIndex()
if err != nil {
return nil, err
}
return i.indexes[0].GetInternal(key)
}
func (i *indexAliasImpl) SetInternal(key, val []byte) error {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return ErrorIndexClosed
}
err := i.isAliasToSingleIndex()
if err != nil {
return err
}
return i.indexes[0].SetInternal(key, val)
}
func (i *indexAliasImpl) DeleteInternal(key []byte) error {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return ErrorIndexClosed
}
err := i.isAliasToSingleIndex()
if err != nil {
return err
}
return i.indexes[0].DeleteInternal(key)
}
func (i *indexAliasImpl) Advanced() (index.Index, store.KVStore, error) {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return nil, nil, ErrorIndexClosed
}
err := i.isAliasToSingleIndex()
if err != nil {
return nil, nil, err
}
return i.indexes[0].Advanced()
}
func (i *indexAliasImpl) Add(indexes ...Index) {
i.mutex.Lock()
defer i.mutex.Unlock()
i.indexes = append(i.indexes, indexes...)
}
func (i *indexAliasImpl) removeSingle(index Index) {
for pos, in := range i.indexes {
if in == index {
i.indexes = append(i.indexes[:pos], i.indexes[pos+1:]...)
break
}
}
}
func (i *indexAliasImpl) Remove(indexes ...Index) {
i.mutex.Lock()
defer i.mutex.Unlock()
for _, in := range indexes {
i.removeSingle(in)
}
}
func (i *indexAliasImpl) Swap(in, out []Index) {
i.mutex.Lock()
defer i.mutex.Unlock()
// add
i.indexes = append(i.indexes, in...)
// delete
for _, ind := range out {
i.removeSingle(ind)
}
}
// createChildSearchRequest creates a separate
// request from the original
// For now, avoid data race on req structure.
// TODO disable highlight/field load on child
// requests, and add code to do this only on
// the actual final results.
// Perhaps that part needs to be optional,
// could be slower in remote usages.
func createChildSearchRequest(req *SearchRequest) *SearchRequest {
rv := SearchRequest{
Query: req.Query,
Size: req.Size + req.From,
From: 0,
Highlight: req.Highlight,
Fields: req.Fields,
Facets: req.Facets,
Explain: req.Explain,
Sort: req.Sort,
}
return &rv
}
type asyncSearchResult struct {
Name string
Result *SearchResult
Err error
}
// MultiSearch executes a SearchRequest across multiple Index objects,
// then merges the results. The indexes must honor any ctx deadline.
func MultiSearch(ctx context.Context, req *SearchRequest, indexes ...Index) (*SearchResult, error) {
searchStart := time.Now()
asyncResults := make(chan *asyncSearchResult, len(indexes))
// run search on each index in separate go routine
var waitGroup sync.WaitGroup
var searchChildIndex = func(in Index, childReq *SearchRequest) {
rv := asyncSearchResult{Name: in.Name()}
rv.Result, rv.Err = in.SearchInContext(ctx, childReq)
asyncResults <- &rv
waitGroup.Done()
}
waitGroup.Add(len(indexes))
for _, in := range indexes {
go searchChildIndex(in, createChildSearchRequest(req))
}
// on another go routine, close after finished
go func() {
waitGroup.Wait()
close(asyncResults)
}()
var sr *SearchResult
indexErrors := make(map[string]error)
for asr := range asyncResults {
if asr.Err == nil {
if sr == nil {
// first result
sr = asr.Result
} else {
// merge with previous
sr.Merge(asr.Result)
}
} else {
indexErrors[asr.Name] = asr.Err
}
}
// merge just concatenated all the hits
// now lets clean it up
// handle case where no results were successful
if sr == nil {
sr = &SearchResult{
Status: &SearchStatus{
Errors: make(map[string]error),
},
}
}
// sort all hits with the requested order
if len(req.Sort) > 0 {
sorter := newMultiSearchHitSorter(req.Sort, sr.Hits)
sort.Sort(sorter)
}
// now skip over the correct From
if req.From > 0 && len(sr.Hits) > req.From {
sr.Hits = sr.Hits[req.From:]
} else if req.From > 0 {
sr.Hits = search.DocumentMatchCollection{}
}
// now trim to the correct size
if req.Size > 0 && len(sr.Hits) > req.Size {
sr.Hits = sr.Hits[0:req.Size]
}
// fix up facets
for name, fr := range req.Facets {
sr.Facets.Fixup(name, fr.Size)
}
// fix up original request
sr.Request = req
searchDuration := time.Since(searchStart)
sr.Took = searchDuration
// fix up errors
if len(indexErrors) > 0 {
if sr.Status.Errors == nil {
sr.Status.Errors = make(map[string]error)
}
for indexName, indexErr := range indexErrors {
sr.Status.Errors[indexName] = indexErr
sr.Status.Total++
sr.Status.Failed++
}
}
return sr, nil
}
func (i *indexAliasImpl) NewBatch() *Batch {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return nil
}
err := i.isAliasToSingleIndex()
if err != nil {
return nil
}
return i.indexes[0].NewBatch()
}
func (i *indexAliasImpl) Name() string {
return i.name
}
func (i *indexAliasImpl) SetName(name string) {
i.name = name
}
type indexAliasImplFieldDict struct {
index *indexAliasImpl
fieldDict index.FieldDict
}
func (f *indexAliasImplFieldDict) Next() (*index.DictEntry, error) {
return f.fieldDict.Next()
}
func (f *indexAliasImplFieldDict) Close() error {
defer f.index.mutex.RUnlock()
return f.fieldDict.Close()
}
type multiSearchHitSorter struct {
hits search.DocumentMatchCollection
sort search.SortOrder
cachedScoring []bool
cachedDesc []bool
}
func newMultiSearchHitSorter(sort search.SortOrder, hits search.DocumentMatchCollection) *multiSearchHitSorter {
return &multiSearchHitSorter{
sort: sort,
hits: hits,
cachedScoring: sort.CacheIsScore(),
cachedDesc: sort.CacheDescending(),
}
}
func (m *multiSearchHitSorter) Len() int { return len(m.hits) }
func (m *multiSearchHitSorter) Swap(i, j int) { m.hits[i], m.hits[j] = m.hits[j], m.hits[i] }
func (m *multiSearchHitSorter) Less(i, j int) bool {
c := m.sort.Compare(m.cachedScoring, m.cachedDesc, m.hits[i], m.hits[j])
return c < 0
}

@ -0,0 +1,729 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package bleve
import (
"encoding/json"
"fmt"
"os"
"sync"
"sync/atomic"
"time"
"golang.org/x/net/context"
"github.com/blevesearch/bleve/document"
"github.com/blevesearch/bleve/index"
"github.com/blevesearch/bleve/index/store"
"github.com/blevesearch/bleve/index/upsidedown"
"github.com/blevesearch/bleve/mapping"
"github.com/blevesearch/bleve/registry"
"github.com/blevesearch/bleve/search"
"github.com/blevesearch/bleve/search/collector"
"github.com/blevesearch/bleve/search/facet"
"github.com/blevesearch/bleve/search/highlight"
)
type indexImpl struct {
path string
name string
meta *indexMeta
i index.Index
m mapping.IndexMapping
mutex sync.RWMutex
open bool
stats *IndexStat
}
const storePath = "store"
var mappingInternalKey = []byte("_mapping")
func indexStorePath(path string) string {
return path + string(os.PathSeparator) + storePath
}
func newIndexUsing(path string, mapping mapping.IndexMapping, indexType string, kvstore string, kvconfig map[string]interface{}) (*indexImpl, error) {
// first validate the mapping
err := mapping.Validate()
if err != nil {
return nil, err
}
if kvconfig == nil {
kvconfig = map[string]interface{}{}
}
if kvstore == "" {
return nil, fmt.Errorf("bleve not configured for file based indexing")
}
rv := indexImpl{
path: path,
name: path,
m: mapping,
meta: newIndexMeta(indexType, kvstore, kvconfig),
}
rv.stats = &IndexStat{i: &rv}
// at this point there is hope that we can be successful, so save index meta
if path != "" {
err = rv.meta.Save(path)
if err != nil {
return nil, err
}
kvconfig["create_if_missing"] = true
kvconfig["error_if_exists"] = true
kvconfig["path"] = indexStorePath(path)
} else {
kvconfig["path"] = ""
}
// open the index
indexTypeConstructor := registry.IndexTypeConstructorByName(rv.meta.IndexType)
if indexTypeConstructor == nil {
return nil, ErrorUnknownIndexType
}
rv.i, err = indexTypeConstructor(rv.meta.Storage, kvconfig, Config.analysisQueue)
if err != nil {
return nil, err
}
err = rv.i.Open()
if err != nil {
if err == index.ErrorUnknownStorageType {
return nil, ErrorUnknownStorageType
}
return nil, err
}
// now persist the mapping
mappingBytes, err := json.Marshal(mapping)
if err != nil {
return nil, err
}
err = rv.i.SetInternal(mappingInternalKey, mappingBytes)
if err != nil {
return nil, err
}
// mark the index as open
rv.mutex.Lock()
defer rv.mutex.Unlock()
rv.open = true
indexStats.Register(&rv)
return &rv, nil
}
func openIndexUsing(path string, runtimeConfig map[string]interface{}) (rv *indexImpl, err error) {
rv = &indexImpl{
path: path,
name: path,
}
rv.stats = &IndexStat{i: rv}
rv.meta, err = openIndexMeta(path)
if err != nil {
return nil, err
}
// backwards compatibility if index type is missing
if rv.meta.IndexType == "" {
rv.meta.IndexType = upsidedown.Name
}
storeConfig := rv.meta.Config
if storeConfig == nil {
storeConfig = map[string]interface{}{}
}
storeConfig["path"] = indexStorePath(path)
storeConfig["create_if_missing"] = false
storeConfig["error_if_exists"] = false
for rck, rcv := range runtimeConfig {
storeConfig[rck] = rcv
}
// open the index
indexTypeConstructor := registry.IndexTypeConstructorByName(rv.meta.IndexType)
if indexTypeConstructor == nil {
return nil, ErrorUnknownIndexType
}
rv.i, err = indexTypeConstructor(rv.meta.Storage, storeConfig, Config.analysisQueue)
if err != nil {
return nil, err
}
err = rv.i.Open()
if err != nil {
if err == index.ErrorUnknownStorageType {
return nil, ErrorUnknownStorageType
}
return nil, err
}
// now load the mapping
indexReader, err := rv.i.Reader()
if err != nil {
return nil, err
}
defer func() {
if cerr := indexReader.Close(); cerr != nil && err == nil {
err = cerr
}
}()
mappingBytes, err := indexReader.GetInternal(mappingInternalKey)
if err != nil {
return nil, err
}
var im *mapping.IndexMappingImpl
err = json.Unmarshal(mappingBytes, &im)
if err != nil {
return nil, fmt.Errorf("error parsing mapping JSON: %v\nmapping contents:\n%s", err, string(mappingBytes))
}
// mark the index as open
rv.mutex.Lock()
defer rv.mutex.Unlock()
rv.open = true
// validate the mapping
err = im.Validate()
if err != nil {
// note even if the mapping is invalid
// we still return an open usable index
return rv, err
}
rv.m = im
indexStats.Register(rv)
return rv, err
}
// Advanced returns implementation internals
// necessary ONLY for advanced usage.
func (i *indexImpl) Advanced() (index.Index, store.KVStore, error) {
s, err := i.i.Advanced()
if err != nil {
return nil, nil, err
}
return i.i, s, nil
}
// Mapping returns the IndexMapping in use by this
// Index.
func (i *indexImpl) Mapping() mapping.IndexMapping {
return i.m
}
// Index the object with the specified identifier.
// The IndexMapping for this index will determine
// how the object is indexed.
func (i *indexImpl) Index(id string, data interface{}) (err error) {
if id == "" {
return ErrorEmptyID
}
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return ErrorIndexClosed
}
doc := document.NewDocument(id)
err = i.m.MapDocument(doc, data)
if err != nil {
return
}
err = i.i.Update(doc)
return
}
// Delete entries for the specified identifier from
// the index.
func (i *indexImpl) Delete(id string) (err error) {
if id == "" {
return ErrorEmptyID
}
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return ErrorIndexClosed
}
err = i.i.Delete(id)
return
}
// Batch executes multiple Index and Delete
// operations at the same time. There are often
// significant performance benefits when performing
// operations in a batch.
func (i *indexImpl) Batch(b *Batch) error {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return ErrorIndexClosed
}
return i.i.Batch(b.internal)
}
// Document is used to find the values of all the
// stored fields for a document in the index. These
// stored fields are put back into a Document object
// and returned.
func (i *indexImpl) Document(id string) (doc *document.Document, err error) {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return nil, ErrorIndexClosed
}
indexReader, err := i.i.Reader()
if err != nil {
return nil, err
}
defer func() {
if cerr := indexReader.Close(); err == nil && cerr != nil {
err = cerr
}
}()
doc, err = indexReader.Document(id)
if err != nil {
return nil, err
}
return doc, nil
}
// DocCount returns the number of documents in the
// index.
func (i *indexImpl) DocCount() (count uint64, err error) {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return 0, ErrorIndexClosed
}
// open a reader for this search
indexReader, err := i.i.Reader()
if err != nil {
return 0, fmt.Errorf("error opening index reader %v", err)
}
defer func() {
if cerr := indexReader.Close(); err == nil && cerr != nil {
err = cerr
}
}()
count, err = indexReader.DocCount()
return
}
// Search executes a search request operation.
// Returns a SearchResult object or an error.
func (i *indexImpl) Search(req *SearchRequest) (sr *SearchResult, err error) {
return i.SearchInContext(context.Background(), req)
}
// SearchInContext executes a search request operation within the provided
// Context. Returns a SearchResult object or an error.
func (i *indexImpl) SearchInContext(ctx context.Context, req *SearchRequest) (sr *SearchResult, err error) {
i.mutex.RLock()
defer i.mutex.RUnlock()
searchStart := time.Now()
if !i.open {
return nil, ErrorIndexClosed
}
collector := collector.NewTopNCollector(req.Size, req.From, req.Sort)
// open a reader for this search
indexReader, err := i.i.Reader()
if err != nil {
return nil, fmt.Errorf("error opening index reader %v", err)
}
defer func() {
if cerr := indexReader.Close(); err == nil && cerr != nil {
err = cerr
}
}()
searcher, err := req.Query.Searcher(indexReader, i.m, req.Explain)
if err != nil {
return nil, err
}
defer func() {
if serr := searcher.Close(); err == nil && serr != nil {
err = serr
}
}()
if req.Facets != nil {
facetsBuilder := search.NewFacetsBuilder(indexReader)
for facetName, facetRequest := range req.Facets {
if facetRequest.NumericRanges != nil {
// build numeric range facet
facetBuilder := facet.NewNumericFacetBuilder(facetRequest.Field, facetRequest.Size)
for _, nr := range facetRequest.NumericRanges {
facetBuilder.AddRange(nr.Name, nr.Min, nr.Max)
}
facetsBuilder.Add(facetName, facetBuilder)
} else if facetRequest.DateTimeRanges != nil {
// build date range facet
facetBuilder := facet.NewDateTimeFacetBuilder(facetRequest.Field, facetRequest.Size)
dateTimeParser := i.m.DateTimeParserNamed("")
for _, dr := range facetRequest.DateTimeRanges {
start, end := dr.ParseDates(dateTimeParser)
facetBuilder.AddRange(dr.Name, start, end)
}
facetsBuilder.Add(facetName, facetBuilder)
} else {
// build terms facet
facetBuilder := facet.NewTermsFacetBuilder(facetRequest.Field, facetRequest.Size)
facetsBuilder.Add(facetName, facetBuilder)
}
}
collector.SetFacetsBuilder(facetsBuilder)
}
err = collector.Collect(ctx, searcher, indexReader)
if err != nil {
return nil, err
}
hits := collector.Results()
var highlighter highlight.Highlighter
if req.Highlight != nil {
// get the right highlighter
highlighter, err = Config.Cache.HighlighterNamed(Config.DefaultHighlighter)
if err != nil {
return nil, err
}
if req.Highlight.Style != nil {
highlighter, err = Config.Cache.HighlighterNamed(*req.Highlight.Style)
if err != nil {
return nil, err
}
}
if highlighter == nil {
return nil, fmt.Errorf("no highlighter named `%s` registered", *req.Highlight.Style)
}
}
for _, hit := range hits {
if len(req.Fields) > 0 || highlighter != nil {
doc, err := indexReader.Document(hit.ID)
if err == nil && doc != nil {
if len(req.Fields) > 0 {
for _, f := range req.Fields {
for _, docF := range doc.Fields {
if f == "*" || docF.Name() == f {
var value interface{}
switch docF := docF.(type) {
case *document.TextField:
value = string(docF.Value())
case *document.NumericField:
num, err := docF.Number()
if err == nil {
value = num
}
case *document.DateTimeField:
datetime, err := docF.DateTime()
if err == nil {
value = datetime.Format(time.RFC3339)
}
case *document.BooleanField:
boolean, err := docF.Boolean()
if err == nil {
value = boolean
}
}
if value != nil {
hit.AddFieldValue(docF.Name(), value)
}
}
}
}
}
if highlighter != nil {
highlightFields := req.Highlight.Fields
if highlightFields == nil {
// add all fields with matches
highlightFields = make([]string, 0, len(hit.Locations))
for k := range hit.Locations {
highlightFields = append(highlightFields, k)
}
}
for _, hf := range highlightFields {
highlighter.BestFragmentsInField(hit, doc, hf, 1)
}
}
} else if doc == nil {
// unexpected case, a doc ID that was found as a search hit
// was unable to be found during document lookup
return nil, ErrorIndexReadInconsistency
}
}
if i.name != "" {
hit.Index = i.name
}
}
atomic.AddUint64(&i.stats.searches, 1)
searchDuration := time.Since(searchStart)
atomic.AddUint64(&i.stats.searchTime, uint64(searchDuration))
if Config.SlowSearchLogThreshold > 0 &&
searchDuration > Config.SlowSearchLogThreshold {
logger.Printf("slow search took %s - %v", searchDuration, req)
}
return &SearchResult{
Status: &SearchStatus{
Total: 1,
Failed: 0,
Successful: 1,
Errors: make(map[string]error),
},
Request: req,
Hits: hits,
Total: collector.Total(),
MaxScore: collector.MaxScore(),
Took: searchDuration,
Facets: collector.FacetResults(),
}, nil
}
// Fields returns the name of all the fields this
// Index has operated on.
func (i *indexImpl) Fields() (fields []string, err error) {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return nil, ErrorIndexClosed
}
indexReader, err := i.i.Reader()
if err != nil {
return nil, err
}
defer func() {
if cerr := indexReader.Close(); err == nil && cerr != nil {
err = cerr
}
}()
fields, err = indexReader.Fields()
if err != nil {
return nil, err
}
return fields, nil
}
func (i *indexImpl) FieldDict(field string) (index.FieldDict, error) {
i.mutex.RLock()
if !i.open {
i.mutex.RUnlock()
return nil, ErrorIndexClosed
}
indexReader, err := i.i.Reader()
if err != nil {
i.mutex.RUnlock()
return nil, err
}
fieldDict, err := indexReader.FieldDict(field)
if err != nil {
i.mutex.RUnlock()
return nil, err
}
return &indexImplFieldDict{
index: i,
indexReader: indexReader,
fieldDict: fieldDict,
}, nil
}
func (i *indexImpl) FieldDictRange(field string, startTerm []byte, endTerm []byte) (index.FieldDict, error) {
i.mutex.RLock()
if !i.open {
i.mutex.RUnlock()
return nil, ErrorIndexClosed
}
indexReader, err := i.i.Reader()
if err != nil {
i.mutex.RUnlock()
return nil, err
}
fieldDict, err := indexReader.FieldDictRange(field, startTerm, endTerm)
if err != nil {
i.mutex.RUnlock()
return nil, err
}
return &indexImplFieldDict{
index: i,
indexReader: indexReader,
fieldDict: fieldDict,
}, nil
}
func (i *indexImpl) FieldDictPrefix(field string, termPrefix []byte) (index.FieldDict, error) {
i.mutex.RLock()
if !i.open {
i.mutex.RUnlock()
return nil, ErrorIndexClosed
}
indexReader, err := i.i.Reader()
if err != nil {
i.mutex.RUnlock()
return nil, err
}
fieldDict, err := indexReader.FieldDictPrefix(field, termPrefix)
if err != nil {
i.mutex.RUnlock()
return nil, err
}
return &indexImplFieldDict{
index: i,
indexReader: indexReader,
fieldDict: fieldDict,
}, nil
}
func (i *indexImpl) Close() error {
i.mutex.Lock()
defer i.mutex.Unlock()
indexStats.UnRegister(i)
i.open = false
return i.i.Close()
}
func (i *indexImpl) Stats() *IndexStat {
return i.stats
}
func (i *indexImpl) StatsMap() map[string]interface{} {
return i.stats.statsMap()
}
func (i *indexImpl) GetInternal(key []byte) (val []byte, err error) {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return nil, ErrorIndexClosed
}
reader, err := i.i.Reader()
if err != nil {
return nil, err
}
defer func() {
if cerr := reader.Close(); err == nil && cerr != nil {
err = cerr
}
}()
val, err = reader.GetInternal(key)
if err != nil {
return nil, err
}
return val, nil
}
func (i *indexImpl) SetInternal(key, val []byte) error {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return ErrorIndexClosed
}
return i.i.SetInternal(key, val)
}
func (i *indexImpl) DeleteInternal(key []byte) error {
i.mutex.RLock()
defer i.mutex.RUnlock()
if !i.open {
return ErrorIndexClosed
}
return i.i.DeleteInternal(key)
}
// NewBatch creates a new empty batch.
func (i *indexImpl) NewBatch() *Batch {
return &Batch{
index: i,
internal: index.NewBatch(),
}
}
func (i *indexImpl) Name() string {
return i.name
}
func (i *indexImpl) SetName(name string) {
indexStats.UnRegister(i)
i.name = name
indexStats.Register(i)
}
type indexImplFieldDict struct {
index *indexImpl
indexReader index.IndexReader
fieldDict index.FieldDict
}
func (f *indexImplFieldDict) Next() (*index.DictEntry, error) {
return f.fieldDict.Next()
}
func (f *indexImplFieldDict) Close() error {
defer f.index.mutex.RUnlock()
err := f.fieldDict.Close()
if err != nil {
return err
}
return f.indexReader.Close()
}

@ -0,0 +1,96 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package bleve
import (
"encoding/json"
"io/ioutil"
"os"
"github.com/blevesearch/bleve/index/upsidedown"
)
const metaFilename = "index_meta.json"
type indexMeta struct {
Storage string `json:"storage"`
IndexType string `json:"index_type"`
Config map[string]interface{} `json:"config,omitempty"`
}
func newIndexMeta(indexType string, storage string, config map[string]interface{}) *indexMeta {
return &indexMeta{
IndexType: indexType,
Storage: storage,
Config: config,
}
}
func openIndexMeta(path string) (*indexMeta, error) {
if _, err := os.Stat(path); os.IsNotExist(err) {
return nil, ErrorIndexPathDoesNotExist
}
indexMetaPath := indexMetaPath(path)
metaBytes, err := ioutil.ReadFile(indexMetaPath)
if err != nil {
return nil, ErrorIndexMetaMissing
}
var im indexMeta
err = json.Unmarshal(metaBytes, &im)
if err != nil {
return nil, ErrorIndexMetaCorrupt
}
if im.IndexType == "" {
im.IndexType = upsidedown.Name
}
return &im, nil
}
func (i *indexMeta) Save(path string) (err error) {
indexMetaPath := indexMetaPath(path)
// ensure any necessary parent directories exist
err = os.MkdirAll(path, 0700)
if err != nil {
if os.IsExist(err) {
return ErrorIndexPathExists
}
return err
}
metaBytes, err := json.Marshal(i)
if err != nil {
return err
}
indexMetaFile, err := os.OpenFile(indexMetaPath, os.O_RDWR|os.O_CREATE|os.O_EXCL, 0666)
if err != nil {
if os.IsExist(err) {
return ErrorIndexPathExists
}
return err
}
defer func() {
if ierr := indexMetaFile.Close(); err == nil && ierr != nil {
err = ierr
}
}()
_, err = indexMetaFile.Write(metaBytes)
if err != nil {
return err
}
return nil
}
func indexMetaPath(path string) string {
return path + string(os.PathSeparator) + metaFilename
}

@ -0,0 +1,75 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package bleve
import (
"encoding/json"
"sync"
"sync/atomic"
)
type IndexStat struct {
searches uint64
searchTime uint64
i *indexImpl
}
func (is *IndexStat) statsMap() map[string]interface{} {
m := map[string]interface{}{}
m["index"] = is.i.i.StatsMap()
m["searches"] = atomic.LoadUint64(&is.searches)
m["search_time"] = atomic.LoadUint64(&is.searchTime)
return m
}
func (is *IndexStat) MarshalJSON() ([]byte, error) {
m := is.statsMap()
return json.Marshal(m)
}
type IndexStats struct {
indexes map[string]*IndexStat
mutex sync.RWMutex
}
func NewIndexStats() *IndexStats {
return &IndexStats{
indexes: make(map[string]*IndexStat),
}
}
func (i *IndexStats) Register(index Index) {
i.mutex.Lock()
defer i.mutex.Unlock()
i.indexes[index.Name()] = index.Stats()
}
func (i *IndexStats) UnRegister(index Index) {
i.mutex.Lock()
defer i.mutex.Unlock()
delete(i.indexes, index.Name())
}
func (i *IndexStats) String() string {
i.mutex.RLock()
defer i.mutex.RUnlock()
bytes, err := json.Marshal(i.indexes)
if err != nil {
return "error marshaling stats"
}
return string(bytes)
}
var indexStats *IndexStats

@ -0,0 +1,61 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package bleve
import "github.com/blevesearch/bleve/mapping"
// NewIndexMapping creates a new IndexMapping that will use all the default indexing rules
func NewIndexMapping() *mapping.IndexMappingImpl {
return mapping.NewIndexMapping()
}
// NewDocumentMapping returns a new document mapping
// with all the default values.
func NewDocumentMapping() *mapping.DocumentMapping {
return mapping.NewDocumentMapping()
}
// NewDocumentStaticMapping returns a new document
// mapping that will not automatically index parts
// of a document without an explicit mapping.
func NewDocumentStaticMapping() *mapping.DocumentMapping {
return mapping.NewDocumentStaticMapping()
}
// NewDocumentDisabledMapping returns a new document
// mapping that will not perform any indexing.
func NewDocumentDisabledMapping() *mapping.DocumentMapping {
return mapping.NewDocumentDisabledMapping()
}
// NewTextFieldMapping returns a default field mapping for text
func NewTextFieldMapping() *mapping.FieldMapping {
return mapping.NewTextFieldMapping()
}
// NewNumericFieldMapping returns a default field mapping for numbers
func NewNumericFieldMapping() *mapping.FieldMapping {
return mapping.NewNumericFieldMapping()
}
// NewDateTimeFieldMapping returns a default field mapping for dates
func NewDateTimeFieldMapping() *mapping.FieldMapping {
return mapping.NewDateTimeFieldMapping()
}
// NewBooleanFieldMapping returns a default field mapping for booleans
func NewBooleanFieldMapping() *mapping.FieldMapping {
return mapping.NewBooleanFieldMapping()
}

@ -0,0 +1,99 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package mapping
type customAnalysis struct {
CharFilters map[string]map[string]interface{} `json:"char_filters,omitempty"`
Tokenizers map[string]map[string]interface{} `json:"tokenizers,omitempty"`
TokenMaps map[string]map[string]interface{} `json:"token_maps,omitempty"`
TokenFilters map[string]map[string]interface{} `json:"token_filters,omitempty"`
Analyzers map[string]map[string]interface{} `json:"analyzers,omitempty"`
DateTimeParsers map[string]map[string]interface{} `json:"date_time_parsers,omitempty"`
}
func (c *customAnalysis) registerAll(i *IndexMappingImpl) error {
for name, config := range c.CharFilters {
_, err := i.cache.DefineCharFilter(name, config)
if err != nil {
return err
}
}
if len(c.Tokenizers) > 0 {
// put all the names in map tracking work to do
todo := map[string]struct{}{}
for name := range c.Tokenizers {
todo[name] = struct{}{}
}
registered := 1
errs := []error{}
// as long as we keep making progress, keep going
for len(todo) > 0 && registered > 0 {
registered = 0
errs = []error{}
for name := range todo {
config := c.Tokenizers[name]
_, err := i.cache.DefineTokenizer(name, config)
if err != nil {
errs = append(errs, err)
} else {
delete(todo, name)
registered++
}
}
}
if len(errs) > 0 {
return errs[0]
}
}
for name, config := range c.TokenMaps {
_, err := i.cache.DefineTokenMap(name, config)
if err != nil {
return err
}
}
for name, config := range c.TokenFilters {
_, err := i.cache.DefineTokenFilter(name, config)
if err != nil {
return err
}
}
for name, config := range c.Analyzers {
_, err := i.cache.DefineAnalyzer(name, config)
if err != nil {
return err
}
}
for name, config := range c.DateTimeParsers {
_, err := i.cache.DefineDateTimeParser(name, config)
if err != nil {
return err
}
}
return nil
}
func newCustomAnalysis() *customAnalysis {
rv := customAnalysis{
CharFilters: make(map[string]map[string]interface{}),
Tokenizers: make(map[string]map[string]interface{}),
TokenMaps: make(map[string]map[string]interface{}),
TokenFilters: make(map[string]map[string]interface{}),
Analyzers: make(map[string]map[string]interface{}),
DateTimeParsers: make(map[string]map[string]interface{}),
}
return &rv
}

@ -0,0 +1,490 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package mapping
import (
"encoding/json"
"fmt"
"reflect"
"time"
"github.com/blevesearch/bleve/registry"
)
// A DocumentMapping describes how a type of document
// should be indexed.
// As documents can be hierarchical, named sub-sections
// of documents are mapped using the same structure in
// the Properties field.
// Each value inside a document can be indexed 0 or more
// ways. These index entries are called fields and
// are stored in the Fields field.
// Entire sections of a document can be ignored or
// excluded by setting Enabled to false.
// If not explicitly mapped, default mapping operations
// are used. To disable this automatic handling, set
// Dynamic to false.
type DocumentMapping struct {
Enabled bool `json:"enabled"`
Dynamic bool `json:"dynamic"`
Properties map[string]*DocumentMapping `json:"properties,omitempty"`
Fields []*FieldMapping `json:"fields,omitempty"`
DefaultAnalyzer string `json:"default_analyzer"`
// StructTagKey overrides "json" when looking for field names in struct tags
StructTagKey string `json:"struct_tag_key,omitempty"`
}
func (dm *DocumentMapping) Validate(cache *registry.Cache) error {
var err error
if dm.DefaultAnalyzer != "" {
_, err := cache.AnalyzerNamed(dm.DefaultAnalyzer)
if err != nil {
return err
}
}
for _, property := range dm.Properties {
err = property.Validate(cache)
if err != nil {
return err
}
}
for _, field := range dm.Fields {
if field.Analyzer != "" {
_, err = cache.AnalyzerNamed(field.Analyzer)
if err != nil {
return err
}
}
if field.DateFormat != "" {
_, err = cache.DateTimeParserNamed(field.DateFormat)
if err != nil {
return err
}
}
switch field.Type {
case "text", "datetime", "number", "boolean":
default:
return fmt.Errorf("unknown field type: '%s'", field.Type)
}
}
return nil
}
// analyzerNameForPath attempts to first find the field
// described by this path, then returns the analyzer
// configured for that field
func (dm *DocumentMapping) analyzerNameForPath(path string) string {
field := dm.fieldDescribedByPath(path)
if field != nil {
return field.Analyzer
}
return ""
}
func (dm *DocumentMapping) fieldDescribedByPath(path string) *FieldMapping {
pathElements := decodePath(path)
if len(pathElements) > 1 {
// easy case, there is more than 1 path element remaining
// the next path element must match a property name
// at this level
for propName, subDocMapping := range dm.Properties {
if propName == pathElements[0] {
return subDocMapping.fieldDescribedByPath(encodePath(pathElements[1:]))
}
}
} else {
// just 1 path elememnt
// first look for property name with empty field
for propName, subDocMapping := range dm.Properties {
if propName == pathElements[0] {
// found property name match, now look at its fields
for _, field := range subDocMapping.Fields {
if field.Name == "" || field.Name == pathElements[0] {
// match
return field
}
}
}
}
// next, walk the properties again, looking for field overriding the name
for propName, subDocMapping := range dm.Properties {
if propName != pathElements[0] {
// property name isn't a match, but field name could override it
for _, field := range subDocMapping.Fields {
if field.Name == pathElements[0] {
return field
}
}
}
}
}
return nil
}
// documentMappingForPath only returns EXACT matches for a sub document
// or for an explicitly mapped field, if you want to find the
// closest document mapping to a field not explicitly mapped
// use closestDocMapping
func (dm *DocumentMapping) documentMappingForPath(path string) *DocumentMapping {
pathElements := decodePath(path)
current := dm
OUTER:
for i, pathElement := range pathElements {
for name, subDocMapping := range current.Properties {
if name == pathElement {
current = subDocMapping
continue OUTER
}
}
// no subDocMapping matches this pathElement
// only if this is the last element check for field name
if i == len(pathElements)-1 {
for _, field := range current.Fields {
if field.Name == pathElement {
break
}
}
}
return nil
}
return current
}
// closestDocMapping findest the most specific document mapping that matches
// part of the provided path
func (dm *DocumentMapping) closestDocMapping(path string) *DocumentMapping {
pathElements := decodePath(path)
current := dm
OUTER:
for _, pathElement := range pathElements {
for name, subDocMapping := range current.Properties {
if name == pathElement {
current = subDocMapping
continue OUTER
}
}
}
return current
}
// NewDocumentMapping returns a new document mapping
// with all the default values.
func NewDocumentMapping() *DocumentMapping {
return &DocumentMapping{
Enabled: true,
Dynamic: true,
}
}
// NewDocumentStaticMapping returns a new document
// mapping that will not automatically index parts
// of a document without an explicit mapping.
func NewDocumentStaticMapping() *DocumentMapping {
return &DocumentMapping{
Enabled: true,
}
}
// NewDocumentDisabledMapping returns a new document
// mapping that will not perform any indexing.
func NewDocumentDisabledMapping() *DocumentMapping {
return &DocumentMapping{}
}
// AddSubDocumentMapping adds the provided DocumentMapping as a sub-mapping
// for the specified named subsection.
func (dm *DocumentMapping) AddSubDocumentMapping(property string, sdm *DocumentMapping) {
if dm.Properties == nil {
dm.Properties = make(map[string]*DocumentMapping)
}
dm.Properties[property] = sdm
}
// AddFieldMappingsAt adds one or more FieldMappings
// at the named sub-document. If the named sub-document
// doesn't yet exist it is created for you.
// This is a convenience function to make most common
// mappings more concise.
// Otherwise, you would:
// subMapping := NewDocumentMapping()
// subMapping.AddFieldMapping(fieldMapping)
// parentMapping.AddSubDocumentMapping(property, subMapping)
func (dm *DocumentMapping) AddFieldMappingsAt(property string, fms ...*FieldMapping) {
if dm.Properties == nil {
dm.Properties = make(map[string]*DocumentMapping)
}
sdm, ok := dm.Properties[property]
if !ok {
sdm = NewDocumentMapping()
}
for _, fm := range fms {
sdm.AddFieldMapping(fm)
}
dm.Properties[property] = sdm
}
// AddFieldMapping adds the provided FieldMapping for this section
// of the document.
func (dm *DocumentMapping) AddFieldMapping(fm *FieldMapping) {
if dm.Fields == nil {
dm.Fields = make([]*FieldMapping, 0)
}
dm.Fields = append(dm.Fields, fm)
}
// UnmarshalJSON offers custom unmarshaling with optional strict validation
func (dm *DocumentMapping) UnmarshalJSON(data []byte) error {
var tmp map[string]json.RawMessage
err := json.Unmarshal(data, &tmp)
if err != nil {
return err
}
// set defaults for fields which might have been omitted
dm.Enabled = true
dm.Dynamic = true
var invalidKeys []string
for k, v := range tmp {
switch k {
case "enabled":
err := json.Unmarshal(v, &dm.Enabled)
if err != nil {
return err
}
case "dynamic":
err := json.Unmarshal(v, &dm.Dynamic)
if err != nil {
return err
}
case "default_analyzer":
err := json.Unmarshal(v, &dm.DefaultAnalyzer)
if err != nil {
return err
}
case "properties":
err := json.Unmarshal(v, &dm.Properties)
if err != nil {
return err
}
case "fields":
err := json.Unmarshal(v, &dm.Fields)
if err != nil {
return err
}
case "struct_tag_key":
err := json.Unmarshal(v, &dm.StructTagKey)
if err != nil {
return err
}
default:
invalidKeys = append(invalidKeys, k)
}
}
if MappingJSONStrict && len(invalidKeys) > 0 {
return fmt.Errorf("document mapping contains invalid keys: %v", invalidKeys)
}
return nil
}
func (dm *DocumentMapping) defaultAnalyzerName(path []string) string {
rv := ""
current := dm
for _, pathElement := range path {
var ok bool
current, ok = current.Properties[pathElement]
if !ok {
break
}
if current.DefaultAnalyzer != "" {
rv = current.DefaultAnalyzer
}
}
return rv
}
func (dm *DocumentMapping) walkDocument(data interface{}, path []string, indexes []uint64, context *walkContext) {
// allow default "json" tag to be overriden
structTagKey := dm.StructTagKey
if structTagKey == "" {
structTagKey = "json"
}
val := reflect.ValueOf(data)
typ := val.Type()
switch typ.Kind() {
case reflect.Map:
// FIXME can add support for other map keys in the future
if typ.Key().Kind() == reflect.String {
for _, key := range val.MapKeys() {
fieldName := key.String()
fieldVal := val.MapIndex(key).Interface()
dm.processProperty(fieldVal, append(path, fieldName), indexes, context)
}
}
case reflect.Struct:
for i := 0; i < val.NumField(); i++ {
field := typ.Field(i)
fieldName := field.Name
// anonymous fields of type struct can elide the type name
if field.Anonymous && field.Type.Kind() == reflect.Struct {
fieldName = ""
}
// if the field has a name under the specified tag, prefer that
tag := field.Tag.Get(structTagKey)
tagFieldName := parseTagName(tag)
if tagFieldName == "-" {
continue
}
// allow tag to set field name to empty, only if anonymous
if field.Tag != "" && (tagFieldName != "" || field.Anonymous) {
fieldName = tagFieldName
}
if val.Field(i).CanInterface() {
fieldVal := val.Field(i).Interface()
newpath := path
if fieldName != "" {
newpath = append(path, fieldName)
}
dm.processProperty(fieldVal, newpath, indexes, context)
}
}
case reflect.Slice, reflect.Array:
for i := 0; i < val.Len(); i++ {
if val.Index(i).CanInterface() {
fieldVal := val.Index(i).Interface()
dm.processProperty(fieldVal, path, append(indexes, uint64(i)), context)
}
}
case reflect.Ptr:
ptrElem := val.Elem()
if ptrElem.IsValid() && ptrElem.CanInterface() {
dm.processProperty(ptrElem.Interface(), path, indexes, context)
}
case reflect.String:
dm.processProperty(val.String(), path, indexes, context)
case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
dm.processProperty(float64(val.Int()), path, indexes, context)
case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64:
dm.processProperty(float64(val.Uint()), path, indexes, context)
case reflect.Float32, reflect.Float64:
dm.processProperty(float64(val.Float()), path, indexes, context)
case reflect.Bool:
dm.processProperty(val.Bool(), path, indexes, context)
}
}
func (dm *DocumentMapping) processProperty(property interface{}, path []string, indexes []uint64, context *walkContext) {
pathString := encodePath(path)
// look to see if there is a mapping for this field
subDocMapping := dm.documentMappingForPath(pathString)
closestDocMapping := dm.closestDocMapping(pathString)
// check to see if we even need to do further processing
if subDocMapping != nil && !subDocMapping.Enabled {
return
}
propertyValue := reflect.ValueOf(property)
if !propertyValue.IsValid() {
// cannot do anything with the zero value
return
}
propertyType := propertyValue.Type()
switch propertyType.Kind() {
case reflect.String:
propertyValueString := propertyValue.String()
if subDocMapping != nil {
// index by explicit mapping
for _, fieldMapping := range subDocMapping.Fields {
fieldMapping.processString(propertyValueString, pathString, path, indexes, context)
}
} else if closestDocMapping.Dynamic {
// automatic indexing behavior
// first see if it can be parsed by the default date parser
dateTimeParser := context.im.DateTimeParserNamed(context.im.DefaultDateTimeParser)
if dateTimeParser != nil {
parsedDateTime, err := dateTimeParser.ParseDateTime(propertyValueString)
if err != nil {
// index as text
fieldMapping := newTextFieldMappingDynamic(context.im)
fieldMapping.processString(propertyValueString, pathString, path, indexes, context)
} else {
// index as datetime
fieldMapping := newDateTimeFieldMappingDynamic(context.im)
fieldMapping.processTime(parsedDateTime, pathString, path, indexes, context)
}
}
}
case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
dm.processProperty(float64(propertyValue.Int()), path, indexes, context)
return
case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64:
dm.processProperty(float64(propertyValue.Uint()), path, indexes, context)
return
case reflect.Float64, reflect.Float32:
propertyValFloat := propertyValue.Float()
if subDocMapping != nil {
// index by explicit mapping
for _, fieldMapping := range subDocMapping.Fields {
fieldMapping.processFloat64(propertyValFloat, pathString, path, indexes, context)
}
} else if closestDocMapping.Dynamic {
// automatic indexing behavior
fieldMapping := newNumericFieldMappingDynamic(context.im)
fieldMapping.processFloat64(propertyValFloat, pathString, path, indexes, context)
}
case reflect.Bool:
propertyValBool := propertyValue.Bool()
if subDocMapping != nil {
// index by explicit mapping
for _, fieldMapping := range subDocMapping.Fields {
fieldMapping.processBoolean(propertyValBool, pathString, path, indexes, context)
}
} else if closestDocMapping.Dynamic {
// automatic indexing behavior
fieldMapping := newBooleanFieldMappingDynamic(context.im)
fieldMapping.processBoolean(propertyValBool, pathString, path, indexes, context)
}
case reflect.Struct:
switch property := property.(type) {
case time.Time:
// don't descend into the time struct
if subDocMapping != nil {
// index by explicit mapping
for _, fieldMapping := range subDocMapping.Fields {
fieldMapping.processTime(property, pathString, path, indexes, context)
}
} else if closestDocMapping.Dynamic {
fieldMapping := newDateTimeFieldMappingDynamic(context.im)
fieldMapping.processTime(property, pathString, path, indexes, context)
}
default:
dm.walkDocument(property, path, indexes, context)
}
default:
dm.walkDocument(property, path, indexes, context)
}
}

@ -0,0 +1,296 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package mapping
import (
"encoding/json"
"fmt"
"time"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/document"
)
// control the default behavior for dynamic fields (those not explicitly mapped)
var (
IndexDynamic = true
StoreDynamic = true
)
// A FieldMapping describes how a specific item
// should be put into the index.
type FieldMapping struct {
Name string `json:"name,omitempty"`
Type string `json:"type,omitempty"`
// Analyzer specifies the name of the analyzer to use for this field. If
// Analyzer is empty, traverse the DocumentMapping tree toward the root and
// pick the first non-empty DefaultAnalyzer found. If there is none, use
// the IndexMapping.DefaultAnalyzer.
Analyzer string `json:"analyzer,omitempty"`
// Store indicates whether to store field values in the index. Stored
// values can be retrieved from search results using SearchRequest.Fields.
Store bool `json:"store,omitempty"`
Index bool `json:"index,omitempty"`
// IncludeTermVectors, if true, makes terms occurrences to be recorded for
// this field. It includes the term position within the terms sequence and
// the term offsets in the source document field. Term vectors are required
// to perform phrase queries or terms highlighting in source documents.
IncludeTermVectors bool `json:"include_term_vectors,omitempty"`
IncludeInAll bool `json:"include_in_all,omitempty"`
DateFormat string `json:"date_format,omitempty"`
}
// NewTextFieldMapping returns a default field mapping for text
func NewTextFieldMapping() *FieldMapping {
return &FieldMapping{
Type: "text",
Store: true,
Index: true,
IncludeTermVectors: true,
IncludeInAll: true,
}
}
func newTextFieldMappingDynamic(im *IndexMappingImpl) *FieldMapping {
rv := NewTextFieldMapping()
rv.Store = im.StoreDynamic
rv.Index = im.IndexDynamic
return rv
}
// NewNumericFieldMapping returns a default field mapping for numbers
func NewNumericFieldMapping() *FieldMapping {
return &FieldMapping{
Type: "number",
Store: true,
Index: true,
IncludeInAll: true,
}
}
func newNumericFieldMappingDynamic(im *IndexMappingImpl) *FieldMapping {
rv := NewNumericFieldMapping()
rv.Store = im.StoreDynamic
rv.Index = im.IndexDynamic
return rv
}
// NewDateTimeFieldMapping returns a default field mapping for dates
func NewDateTimeFieldMapping() *FieldMapping {
return &FieldMapping{
Type: "datetime",
Store: true,
Index: true,
IncludeInAll: true,
}
}
func newDateTimeFieldMappingDynamic(im *IndexMappingImpl) *FieldMapping {
rv := NewDateTimeFieldMapping()
rv.Store = im.StoreDynamic
rv.Index = im.IndexDynamic
return rv
}
// NewBooleanFieldMapping returns a default field mapping for booleans
func NewBooleanFieldMapping() *FieldMapping {
return &FieldMapping{
Type: "boolean",
Store: true,
Index: true,
IncludeInAll: true,
}
}
func newBooleanFieldMappingDynamic(im *IndexMappingImpl) *FieldMapping {
rv := NewBooleanFieldMapping()
rv.Store = im.StoreDynamic
rv.Index = im.IndexDynamic
return rv
}
// Options returns the indexing options for this field.
func (fm *FieldMapping) Options() document.IndexingOptions {
var rv document.IndexingOptions
if fm.Store {
rv |= document.StoreField
}
if fm.Index {
rv |= document.IndexField
}
if fm.IncludeTermVectors {
rv |= document.IncludeTermVectors
}
return rv
}
func (fm *FieldMapping) processString(propertyValueString string, pathString string, path []string, indexes []uint64, context *walkContext) {
fieldName := getFieldName(pathString, path, fm)
options := fm.Options()
if fm.Type == "text" {
analyzer := fm.analyzerForField(path, context)
field := document.NewTextFieldCustom(fieldName, indexes, []byte(propertyValueString), options, analyzer)
context.doc.AddField(field)
if !fm.IncludeInAll {
context.excludedFromAll = append(context.excludedFromAll, fieldName)
}
} else if fm.Type == "datetime" {
dateTimeFormat := context.im.DefaultDateTimeParser
if fm.DateFormat != "" {
dateTimeFormat = fm.DateFormat
}
dateTimeParser := context.im.DateTimeParserNamed(dateTimeFormat)
if dateTimeParser != nil {
parsedDateTime, err := dateTimeParser.ParseDateTime(propertyValueString)
if err == nil {
fm.processTime(parsedDateTime, pathString, path, indexes, context)
}
}
}
}
func (fm *FieldMapping) processFloat64(propertyValFloat float64, pathString string, path []string, indexes []uint64, context *walkContext) {
fieldName := getFieldName(pathString, path, fm)
if fm.Type == "number" {
options := fm.Options()
field := document.NewNumericFieldWithIndexingOptions(fieldName, indexes, propertyValFloat, options)
context.doc.AddField(field)
if !fm.IncludeInAll {
context.excludedFromAll = append(context.excludedFromAll, fieldName)
}
}
}
func (fm *FieldMapping) processTime(propertyValueTime time.Time, pathString string, path []string, indexes []uint64, context *walkContext) {
fieldName := getFieldName(pathString, path, fm)
if fm.Type == "datetime" {
options := fm.Options()
field, err := document.NewDateTimeFieldWithIndexingOptions(fieldName, indexes, propertyValueTime, options)
if err == nil {
context.doc.AddField(field)
} else {
logger.Printf("could not build date %v", err)
}
if !fm.IncludeInAll {
context.excludedFromAll = append(context.excludedFromAll, fieldName)
}
}
}
func (fm *FieldMapping) processBoolean(propertyValueBool bool, pathString string, path []string, indexes []uint64, context *walkContext) {
fieldName := getFieldName(pathString, path, fm)
if fm.Type == "boolean" {
options := fm.Options()
field := document.NewBooleanFieldWithIndexingOptions(fieldName, indexes, propertyValueBool, options)
context.doc.AddField(field)
if !fm.IncludeInAll {
context.excludedFromAll = append(context.excludedFromAll, fieldName)
}
}
}
func (fm *FieldMapping) analyzerForField(path []string, context *walkContext) *analysis.Analyzer {
analyzerName := fm.Analyzer
if analyzerName == "" {
analyzerName = context.dm.defaultAnalyzerName(path)
if analyzerName == "" {
analyzerName = context.im.DefaultAnalyzer
}
}
return context.im.AnalyzerNamed(analyzerName)
}
func getFieldName(pathString string, path []string, fieldMapping *FieldMapping) string {
fieldName := pathString
if fieldMapping.Name != "" {
parentName := ""
if len(path) > 1 {
parentName = encodePath(path[:len(path)-1]) + pathSeparator
}
fieldName = parentName + fieldMapping.Name
}
return fieldName
}
// UnmarshalJSON offers custom unmarshaling with optional strict validation
func (fm *FieldMapping) UnmarshalJSON(data []byte) error {
var tmp map[string]json.RawMessage
err := json.Unmarshal(data, &tmp)
if err != nil {
return err
}
var invalidKeys []string
for k, v := range tmp {
switch k {
case "name":
err := json.Unmarshal(v, &fm.Name)
if err != nil {
return err
}
case "type":
err := json.Unmarshal(v, &fm.Type)
if err != nil {
return err
}
case "analyzer":
err := json.Unmarshal(v, &fm.Analyzer)
if err != nil {
return err
}
case "store":
err := json.Unmarshal(v, &fm.Store)
if err != nil {
return err
}
case "index":
err := json.Unmarshal(v, &fm.Index)
if err != nil {
return err
}
case "include_term_vectors":
err := json.Unmarshal(v, &fm.IncludeTermVectors)
if err != nil {
return err
}
case "include_in_all":
err := json.Unmarshal(v, &fm.IncludeInAll)
if err != nil {
return err
}
case "date_format":
err := json.Unmarshal(v, &fm.DateFormat)
if err != nil {
return err
}
default:
invalidKeys = append(invalidKeys, k)
}
}
if MappingJSONStrict && len(invalidKeys) > 0 {
return fmt.Errorf("field mapping contains invalid keys: %v", invalidKeys)
}
return nil
}

@ -0,0 +1,430 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package mapping
import (
"encoding/json"
"fmt"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/analysis/analyzer/standard"
"github.com/blevesearch/bleve/analysis/datetime/optional"
"github.com/blevesearch/bleve/document"
"github.com/blevesearch/bleve/registry"
)
var MappingJSONStrict = false
const defaultTypeField = "_type"
const defaultType = "_default"
const defaultField = "_all"
const defaultAnalyzer = standard.Name
const defaultDateTimeParser = optional.Name
// An IndexMappingImpl controls how objects are placed
// into an index.
// First the type of the object is determined.
// Once the type is know, the appropriate
// DocumentMapping is selected by the type.
// If no mapping was determined for that type,
// a DefaultMapping will be used.
type IndexMappingImpl struct {
TypeMapping map[string]*DocumentMapping `json:"types,omitempty"`
DefaultMapping *DocumentMapping `json:"default_mapping"`
TypeField string `json:"type_field"`
DefaultType string `json:"default_type"`
DefaultAnalyzer string `json:"default_analyzer"`
DefaultDateTimeParser string `json:"default_datetime_parser"`
DefaultField string `json:"default_field"`
StoreDynamic bool `json:"store_dynamic"`
IndexDynamic bool `json:"index_dynamic"`
CustomAnalysis *customAnalysis `json:"analysis,omitempty"`
cache *registry.Cache
}
// AddCustomCharFilter defines a custom char filter for use in this mapping
func (im *IndexMappingImpl) AddCustomCharFilter(name string, config map[string]interface{}) error {
_, err := im.cache.DefineCharFilter(name, config)
if err != nil {
return err
}
im.CustomAnalysis.CharFilters[name] = config
return nil
}
// AddCustomTokenizer defines a custom tokenizer for use in this mapping
func (im *IndexMappingImpl) AddCustomTokenizer(name string, config map[string]interface{}) error {
_, err := im.cache.DefineTokenizer(name, config)
if err != nil {
return err
}
im.CustomAnalysis.Tokenizers[name] = config
return nil
}
// AddCustomTokenMap defines a custom token map for use in this mapping
func (im *IndexMappingImpl) AddCustomTokenMap(name string, config map[string]interface{}) error {
_, err := im.cache.DefineTokenMap(name, config)
if err != nil {
return err
}
im.CustomAnalysis.TokenMaps[name] = config
return nil
}
// AddCustomTokenFilter defines a custom token filter for use in this mapping
func (im *IndexMappingImpl) AddCustomTokenFilter(name string, config map[string]interface{}) error {
_, err := im.cache.DefineTokenFilter(name, config)
if err != nil {
return err
}
im.CustomAnalysis.TokenFilters[name] = config
return nil
}
// AddCustomAnalyzer defines a custom analyzer for use in this mapping. The
// config map must have a "type" string entry to resolve the analyzer
// constructor. The constructor is invoked with the remaining entries and
// returned analyzer is registered in the IndexMapping.
//
// bleve comes with predefined analyzers, like
// github.com/blevesearch/bleve/analysis/analyzers/custom_analyzer. They are
// available only if their package is imported by client code. To achieve this,
// use their metadata to fill configuration entries:
//
// import (
// "github.com/blevesearch/bleve/analysis/analyzers/custom_analyzer"
// "github.com/blevesearch/bleve/analysis/char_filters/html_char_filter"
// "github.com/blevesearch/bleve/analysis/token_filters/lower_case_filter"
// "github.com/blevesearch/bleve/analysis/tokenizers/unicode"
// )
//
// m := bleve.NewIndexMapping()
// err := m.AddCustomAnalyzer("html", map[string]interface{}{
// "type": custom_analyzer.Name,
// "char_filters": []string{
// html_char_filter.Name,
// },
// "tokenizer": unicode.Name,
// "token_filters": []string{
// lower_case_filter.Name,
// ...
// },
// })
func (im *IndexMappingImpl) AddCustomAnalyzer(name string, config map[string]interface{}) error {
_, err := im.cache.DefineAnalyzer(name, config)
if err != nil {
return err
}
im.CustomAnalysis.Analyzers[name] = config
return nil
}
// AddCustomDateTimeParser defines a custom date time parser for use in this mapping
func (im *IndexMappingImpl) AddCustomDateTimeParser(name string, config map[string]interface{}) error {
_, err := im.cache.DefineDateTimeParser(name, config)
if err != nil {
return err
}
im.CustomAnalysis.DateTimeParsers[name] = config
return nil
}
// NewIndexMapping creates a new IndexMapping that will use all the default indexing rules
func NewIndexMapping() *IndexMappingImpl {
return &IndexMappingImpl{
TypeMapping: make(map[string]*DocumentMapping),
DefaultMapping: NewDocumentMapping(),
TypeField: defaultTypeField,
DefaultType: defaultType,
DefaultAnalyzer: defaultAnalyzer,
DefaultDateTimeParser: defaultDateTimeParser,
DefaultField: defaultField,
IndexDynamic: IndexDynamic,
StoreDynamic: StoreDynamic,
CustomAnalysis: newCustomAnalysis(),
cache: registry.NewCache(),
}
}
// Validate will walk the entire structure ensuring the following
// explicitly named and default analyzers can be built
func (im *IndexMappingImpl) Validate() error {
_, err := im.cache.AnalyzerNamed(im.DefaultAnalyzer)
if err != nil {
return err
}
_, err = im.cache.DateTimeParserNamed(im.DefaultDateTimeParser)
if err != nil {
return err
}
err = im.DefaultMapping.Validate(im.cache)
if err != nil {
return err
}
for _, docMapping := range im.TypeMapping {
err = docMapping.Validate(im.cache)
if err != nil {
return err
}
}
return nil
}
// AddDocumentMapping sets a custom document mapping for the specified type
func (im *IndexMappingImpl) AddDocumentMapping(doctype string, dm *DocumentMapping) {
im.TypeMapping[doctype] = dm
}
func (im *IndexMappingImpl) mappingForType(docType string) *DocumentMapping {
docMapping := im.TypeMapping[docType]
if docMapping == nil {
docMapping = im.DefaultMapping
}
return docMapping
}
// UnmarshalJSON offers custom unmarshaling with optional strict validation
func (im *IndexMappingImpl) UnmarshalJSON(data []byte) error {
var tmp map[string]json.RawMessage
err := json.Unmarshal(data, &tmp)
if err != nil {
return err
}
// set defaults for fields which might have been omitted
im.cache = registry.NewCache()
im.CustomAnalysis = newCustomAnalysis()
im.TypeField = defaultTypeField
im.DefaultType = defaultType
im.DefaultAnalyzer = defaultAnalyzer
im.DefaultDateTimeParser = defaultDateTimeParser
im.DefaultField = defaultField
im.DefaultMapping = NewDocumentMapping()
im.TypeMapping = make(map[string]*DocumentMapping)
im.StoreDynamic = StoreDynamic
im.IndexDynamic = IndexDynamic
var invalidKeys []string
for k, v := range tmp {
switch k {
case "analysis":
err := json.Unmarshal(v, &im.CustomAnalysis)
if err != nil {
return err
}
case "type_field":
err := json.Unmarshal(v, &im.TypeField)
if err != nil {
return err
}
case "default_type":
err := json.Unmarshal(v, &im.DefaultType)
if err != nil {
return err
}
case "default_analyzer":
err := json.Unmarshal(v, &im.DefaultAnalyzer)
if err != nil {
return err
}
case "default_datetime_parser":
err := json.Unmarshal(v, &im.DefaultDateTimeParser)
if err != nil {
return err
}
case "default_field":
err := json.Unmarshal(v, &im.DefaultField)
if err != nil {
return err
}
case "default_mapping":
err := json.Unmarshal(v, &im.DefaultMapping)
if err != nil {
return err
}
case "types":
err := json.Unmarshal(v, &im.TypeMapping)
if err != nil {
return err
}
case "store_dynamic":
err := json.Unmarshal(v, &im.StoreDynamic)
if err != nil {
return err
}
case "index_dynamic":
err := json.Unmarshal(v, &im.IndexDynamic)
if err != nil {
return err
}
default:
invalidKeys = append(invalidKeys, k)
}
}
if MappingJSONStrict && len(invalidKeys) > 0 {
return fmt.Errorf("index mapping contains invalid keys: %v", invalidKeys)
}
err = im.CustomAnalysis.registerAll(im)
if err != nil {
return err
}
return nil
}
func (im *IndexMappingImpl) determineType(data interface{}) string {
// first see if the object implements Classifier
classifier, ok := data.(Classifier)
if ok {
return classifier.Type()
}
// now see if we can find a type using the mapping
typ, ok := mustString(lookupPropertyPath(data, im.TypeField))
if ok {
return typ
}
return im.DefaultType
}
func (im *IndexMappingImpl) MapDocument(doc *document.Document, data interface{}) error {
docType := im.determineType(data)
docMapping := im.mappingForType(docType)
walkContext := im.newWalkContext(doc, docMapping)
if docMapping.Enabled {
docMapping.walkDocument(data, []string{}, []uint64{}, walkContext)
// see if the _all field was disabled
allMapping := docMapping.documentMappingForPath("_all")
if allMapping == nil || (allMapping.Enabled != false) {
field := document.NewCompositeFieldWithIndexingOptions("_all", true, []string{}, walkContext.excludedFromAll, document.IndexField|document.IncludeTermVectors)
doc.AddField(field)
}
}
return nil
}
type walkContext struct {
doc *document.Document
im *IndexMappingImpl
dm *DocumentMapping
excludedFromAll []string
}
func (im *IndexMappingImpl) newWalkContext(doc *document.Document, dm *DocumentMapping) *walkContext {
return &walkContext{
doc: doc,
im: im,
dm: dm,
excludedFromAll: []string{},
}
}
// AnalyzerNameForPath attempts to find the best analyzer to use with only a
// field name will walk all the document types, look for field mappings at the
// provided path, if one exists and it has an explicit analyzer that is
// returned.
func (im *IndexMappingImpl) AnalyzerNameForPath(path string) string {
// first we look for explicit mapping on the field
for _, docMapping := range im.TypeMapping {
analyzerName := docMapping.analyzerNameForPath(path)
if analyzerName != "" {
return analyzerName
}
}
// now try the default mapping
pathMapping := im.DefaultMapping.documentMappingForPath(path)
if pathMapping != nil {
if len(pathMapping.Fields) > 0 {
if pathMapping.Fields[0].Analyzer != "" {
return pathMapping.Fields[0].Analyzer
}
}
}
// next we will try default analyzers for the path
pathDecoded := decodePath(path)
for _, docMapping := range im.TypeMapping {
rv := docMapping.defaultAnalyzerName(pathDecoded)
if rv != "" {
return rv
}
}
return im.DefaultAnalyzer
}
func (im *IndexMappingImpl) AnalyzerNamed(name string) *analysis.Analyzer {
analyzer, err := im.cache.AnalyzerNamed(name)
if err != nil {
logger.Printf("error using analyzer named: %s", name)
return nil
}
return analyzer
}
func (im *IndexMappingImpl) DateTimeParserNamed(name string) analysis.DateTimeParser {
if name == "" {
name = im.DefaultDateTimeParser
}
dateTimeParser, err := im.cache.DateTimeParserNamed(name)
if err != nil {
logger.Printf("error using datetime parser named: %s", name)
return nil
}
return dateTimeParser
}
func (im *IndexMappingImpl) datetimeParserNameForPath(path string) string {
// first we look for explicit mapping on the field
for _, docMapping := range im.TypeMapping {
pathMapping := docMapping.documentMappingForPath(path)
if pathMapping != nil {
if len(pathMapping.Fields) > 0 {
if pathMapping.Fields[0].Analyzer != "" {
return pathMapping.Fields[0].Analyzer
}
}
}
}
return im.DefaultDateTimeParser
}
func (im *IndexMappingImpl) AnalyzeText(analyzerName string, text []byte) (analysis.TokenStream, error) {
analyzer, err := im.cache.AnalyzerNamed(analyzerName)
if err != nil {
return nil, err
}
return analyzer.Analyze(text), nil
}
// FieldAnalyzer returns the name of the analyzer used on a field.
func (im *IndexMappingImpl) FieldAnalyzer(field string) string {
return im.AnalyzerNameForPath(field)
}
// wrapper to satisfy new interface
func (im *IndexMappingImpl) DefaultSearchField() string {
return im.DefaultField
}

@ -0,0 +1,49 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package mapping
import (
"io/ioutil"
"log"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/document"
)
// A Classifier is an interface describing any object
// which knows how to identify its own type.
type Classifier interface {
Type() string
}
var logger = log.New(ioutil.Discard, "bleve mapping ", log.LstdFlags)
// SetLog sets the logger used for logging
// by default log messages are sent to ioutil.Discard
func SetLog(l *log.Logger) {
logger = l
}
type IndexMapping interface {
MapDocument(doc *document.Document, data interface{}) error
Validate() error
DateTimeParserNamed(name string) analysis.DateTimeParser
DefaultSearchField() string
AnalyzerNameForPath(path string) string
AnalyzerNamed(name string) *analysis.Analyzer
}

@ -0,0 +1,89 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package mapping
import (
"reflect"
"strings"
)
func lookupPropertyPath(data interface{}, path string) interface{} {
pathParts := decodePath(path)
current := data
for _, part := range pathParts {
current = lookupPropertyPathPart(current, part)
if current == nil {
break
}
}
return current
}
func lookupPropertyPathPart(data interface{}, part string) interface{} {
val := reflect.ValueOf(data)
typ := val.Type()
switch typ.Kind() {
case reflect.Map:
// FIXME can add support for other map keys in the future
if typ.Key().Kind() == reflect.String {
key := reflect.ValueOf(part)
entry := val.MapIndex(key)
if entry.IsValid() {
return entry.Interface()
}
}
case reflect.Struct:
field := val.FieldByName(part)
if field.IsValid() && field.CanInterface() {
return field.Interface()
}
case reflect.Ptr:
ptrElem := val.Elem()
if ptrElem.IsValid() && ptrElem.CanInterface() {
return lookupPropertyPathPart(ptrElem.Interface(), part)
}
}
return nil
}
const pathSeparator = "."
func decodePath(path string) []string {
return strings.Split(path, pathSeparator)
}
func encodePath(pathElements []string) string {
return strings.Join(pathElements, pathSeparator)
}
func mustString(data interface{}) (string, bool) {
if data != nil {
str, ok := data.(string)
if ok {
return str, true
}
}
return "", false
}
// parseTagName extracts the field name from a struct tag
func parseTagName(tag string) string {
if idx := strings.Index(tag, ","); idx != -1 {
return tag[:idx]
}
return tag
}

@ -0,0 +1,34 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package numeric
import (
"math"
)
func Float64ToInt64(f float64) int64 {
fasint := int64(math.Float64bits(f))
if fasint < 0 {
fasint = fasint ^ 0x7fffffffffffffff
}
return fasint
}
func Int64ToFloat64(i int64) float64 {
if i < 0 {
i ^= 0x7fffffffffffffff
}
return math.Float64frombits(uint64(i))
}

@ -0,0 +1,92 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package numeric
import "fmt"
const ShiftStartInt64 byte = 0x20
// PrefixCoded is a byte array encoding of
// 64-bit numeric values shifted by 0-63 bits
type PrefixCoded []byte
func NewPrefixCodedInt64(in int64, shift uint) (PrefixCoded, error) {
if shift > 63 {
return nil, fmt.Errorf("cannot shift %d, must be between 0 and 63", shift)
}
nChars := ((63 - shift) / 7) + 1
rv := make(PrefixCoded, nChars+1)
rv[0] = ShiftStartInt64 + byte(shift)
sortableBits := int64(uint64(in) ^ 0x8000000000000000)
sortableBits = int64(uint64(sortableBits) >> shift)
for nChars > 0 {
// Store 7 bits per byte for compatibility
// with UTF-8 encoding of terms
rv[nChars] = byte(sortableBits & 0x7f)
nChars--
sortableBits = int64(uint64(sortableBits) >> 7)
}
return rv, nil
}
func MustNewPrefixCodedInt64(in int64, shift uint) PrefixCoded {
rv, err := NewPrefixCodedInt64(in, shift)
if err != nil {
panic(err)
}
return rv
}
// Shift returns the number of bits shifted
// returns 0 if in uninitialized state
func (p PrefixCoded) Shift() (uint, error) {
if len(p) > 0 {
shift := p[0] - ShiftStartInt64
if shift < 0 || shift < 63 {
return uint(shift), nil
}
}
return 0, fmt.Errorf("invalid prefix coded value")
}
func (p PrefixCoded) Int64() (int64, error) {
shift, err := p.Shift()
if err != nil {
return 0, err
}
var sortableBits int64
for _, inbyte := range p[1:] {
sortableBits <<= 7
sortableBits |= int64(inbyte)
}
return int64(uint64((sortableBits << shift)) ^ 0x8000000000000000), nil
}
func ValidPrefixCodedTerm(p string) (bool, int) {
if len(p) > 0 {
if p[0] < ShiftStartInt64 || p[0] > ShiftStartInt64+63 {
return false, 0
}
shift := p[0] - ShiftStartInt64
nChars := ((63 - int(shift)) / 7) + 1
if len(p) != nChars+1 {
return false, 0
}
return true, int(shift)
}
return false, 0
}

@ -0,0 +1,186 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package bleve
import (
"time"
"github.com/blevesearch/bleve/search/query"
)
// NewBoolFieldQuery creates a new Query for boolean fields
func NewBoolFieldQuery(val bool) *query.BoolFieldQuery {
return query.NewBoolFieldQuery(val)
}
// NewBooleanQuery creates a compound Query composed
// of several other Query objects.
// These other query objects are added using the
// AddMust() AddShould() and AddMustNot() methods.
// Result documents must satisfy ALL of the
// must Queries.
// Result documents must satisfy NONE of the must not
// Queries.
// Result documents that ALSO satisfy any of the should
// Queries will score higher.
func NewBooleanQuery() *query.BooleanQuery {
return query.NewBooleanQuery(nil, nil, nil)
}
// NewConjunctionQuery creates a new compound Query.
// Result documents must satisfy all of the queries.
func NewConjunctionQuery(conjuncts ...query.Query) *query.ConjunctionQuery {
return query.NewConjunctionQuery(conjuncts)
}
// NewDateRangeQuery creates a new Query for ranges
// of date values.
// Date strings are parsed using the DateTimeParser configured in the
// top-level config.QueryDateTimeParser
// Either, but not both endpoints can be nil.
func NewDateRangeQuery(start, end time.Time) *query.DateRangeQuery {
return query.NewDateRangeQuery(start, end)
}
// NewDateRangeInclusiveQuery creates a new Query for ranges
// of date values.
// Date strings are parsed using the DateTimeParser configured in the
// top-level config.QueryDateTimeParser
// Either, but not both endpoints can be nil.
// startInclusive and endInclusive control inclusion of the endpoints.
func NewDateRangeInclusiveQuery(start, end time.Time, startInclusive, endInclusive *bool) *query.DateRangeQuery {
return query.NewDateRangeInclusiveQuery(start, end, startInclusive, endInclusive)
}
// NewDisjunctionQuery creates a new compound Query.
// Result documents satisfy at least one Query.
func NewDisjunctionQuery(disjuncts ...query.Query) *query.DisjunctionQuery {
return query.NewDisjunctionQuery(disjuncts)
}
// NewDocIDQuery creates a new Query object returning indexed documents among
// the specified set. Combine it with ConjunctionQuery to restrict the scope of
// other queries output.
func NewDocIDQuery(ids []string) *query.DocIDQuery {
return query.NewDocIDQuery(ids)
}
// NewFuzzyQuery creates a new Query which finds
// documents containing terms within a specific
// fuzziness of the specified term.
// The default fuzziness is 1.
//
// The current implementation uses Levenshtein edit
// distance as the fuzziness metric.
func NewFuzzyQuery(term string) *query.FuzzyQuery {
return query.NewFuzzyQuery(term)
}
// NewMatchAllQuery creates a Query which will
// match all documents in the index.
func NewMatchAllQuery() *query.MatchAllQuery {
return query.NewMatchAllQuery()
}
// NewMatchNoneQuery creates a Query which will not
// match any documents in the index.
func NewMatchNoneQuery() *query.MatchNoneQuery {
return query.NewMatchNoneQuery()
}
// NewMatchPhraseQuery creates a new Query object
// for matching phrases in the index.
// An Analyzer is chosen based on the field.
// Input text is analyzed using this analyzer.
// Token terms resulting from this analysis are
// used to build a search phrase. Result documents
// must match this phrase. Queried field must have been indexed with
// IncludeTermVectors set to true.
func NewMatchPhraseQuery(matchPhrase string) *query.MatchPhraseQuery {
return query.NewMatchPhraseQuery(matchPhrase)
}
// NewMatchQuery creates a Query for matching text.
// An Analyzer is chosen based on the field.
// Input text is analyzed using this analyzer.
// Token terms resulting from this analysis are
// used to perform term searches. Result documents
// must satisfy at least one of these term searches.
func NewMatchQuery(match string) *query.MatchQuery {
return query.NewMatchQuery(match)
}
// NewNumericRangeQuery creates a new Query for ranges
// of numeric values.
// Either, but not both endpoints can be nil.
// The minimum value is inclusive.
// The maximum value is exclusive.
func NewNumericRangeQuery(min, max *float64) *query.NumericRangeQuery {
return query.NewNumericRangeQuery(min, max)
}
// NewNumericRangeInclusiveQuery creates a new Query for ranges
// of numeric values.
// Either, but not both endpoints can be nil.
// Control endpoint inclusion with inclusiveMin, inclusiveMax.
func NewNumericRangeInclusiveQuery(min, max *float64, minInclusive, maxInclusive *bool) *query.NumericRangeQuery {
return query.NewNumericRangeInclusiveQuery(min, max, minInclusive, maxInclusive)
}
// NewPhraseQuery creates a new Query for finding
// exact term phrases in the index.
// The provided terms must exist in the correct
// order, at the correct index offsets, in the
// specified field. Queried field must have been indexed with
// IncludeTermVectors set to true.
func NewPhraseQuery(terms []string, field string) *query.PhraseQuery {
return query.NewPhraseQuery(terms, field)
}
// NewPrefixQuery creates a new Query which finds
// documents containing terms that start with the
// specified prefix.
func NewPrefixQuery(prefix string) *query.PrefixQuery {
return query.NewPrefixQuery(prefix)
}
// NewRegexpQuery creates a new Query which finds
// documents containing terms that match the
// specified regular expression.
func NewRegexpQuery(regexp string) *query.RegexpQuery {
return query.NewRegexpQuery(regexp)
}
// NewQueryStringQuery creates a new Query used for
// finding documents that satisfy a query string. The
// query string is a small query language for humans.
func NewQueryStringQuery(q string) *query.QueryStringQuery {
return query.NewQueryStringQuery(q)
}
// NewTermQuery creates a new Query for finding an
// exact term match in the index.
func NewTermQuery(term string) *query.TermQuery {
return query.NewTermQuery(term)
}
// NewWildcardQuery creates a new Query which finds
// documents containing terms that match the
// specified wildcard. In the wildcard pattern '*'
// will match any sequence of 0 or more characters,
// and '?' will match any single character.
func NewWildcardQuery(wildcard string) *query.WildcardQuery {
return query.NewWildcardQuery(wildcard)
}

@ -0,0 +1,89 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package registry
import (
"fmt"
"github.com/blevesearch/bleve/analysis"
)
func RegisterAnalyzer(name string, constructor AnalyzerConstructor) {
_, exists := analyzers[name]
if exists {
panic(fmt.Errorf("attempted to register duplicate analyzer named '%s'", name))
}
analyzers[name] = constructor
}
type AnalyzerConstructor func(config map[string]interface{}, cache *Cache) (*analysis.Analyzer, error)
type AnalyzerRegistry map[string]AnalyzerConstructor
type AnalyzerCache struct {
*ConcurrentCache
}
func NewAnalyzerCache() *AnalyzerCache {
return &AnalyzerCache{
NewConcurrentCache(),
}
}
func AnalyzerBuild(name string, config map[string]interface{}, cache *Cache) (interface{}, error) {
cons, registered := analyzers[name]
if !registered {
return nil, fmt.Errorf("no analyzer with name or type '%s' registered", name)
}
analyzer, err := cons(config, cache)
if err != nil {
return nil, fmt.Errorf("error building analyzer: %v", err)
}
return analyzer, nil
}
func (c *AnalyzerCache) AnalyzerNamed(name string, cache *Cache) (*analysis.Analyzer, error) {
item, err := c.ItemNamed(name, cache, AnalyzerBuild)
if err != nil {
return nil, err
}
return item.(*analysis.Analyzer), nil
}
func (c *AnalyzerCache) DefineAnalyzer(name string, typ string, config map[string]interface{}, cache *Cache) (*analysis.Analyzer, error) {
item, err := c.DefineItem(name, typ, config, cache, AnalyzerBuild)
if err != nil {
if err == ErrAlreadyDefined {
return nil, fmt.Errorf("analyzer named '%s' already defined", name)
}
return nil, err
}
return item.(*analysis.Analyzer), nil
}
func AnalyzerTypesAndInstances() ([]string, []string) {
emptyConfig := map[string]interface{}{}
emptyCache := NewCache()
var types []string
var instances []string
for name, cons := range analyzers {
_, err := cons(emptyConfig, emptyCache)
if err == nil {
instances = append(instances, name)
} else {
types = append(types, name)
}
}
return types, instances
}

@ -0,0 +1,87 @@
// Copyright (c) 2016 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package registry
import (
"fmt"
"sync"
)
var ErrAlreadyDefined = fmt.Errorf("item already defined")
type CacheBuild func(name string, config map[string]interface{}, cache *Cache) (interface{}, error)
type ConcurrentCache struct {
mutex sync.RWMutex
data map[string]interface{}
}
func NewConcurrentCache() *ConcurrentCache {
return &ConcurrentCache{
data: make(map[string]interface{}),
}
}
func (c *ConcurrentCache) ItemNamed(name string, cache *Cache, build CacheBuild) (interface{}, error) {
c.mutex.RLock()
item, cached := c.data[name]
if cached {
c.mutex.RUnlock()
return item, nil
}
// give up read lock
c.mutex.RUnlock()
// try to build it
newItem, err := build(name, nil, cache)
if err != nil {
return nil, err
}
// acquire write lock
c.mutex.Lock()
defer c.mutex.Unlock()
// check again because it could have been created while trading locks
item, cached = c.data[name]
if cached {
return item, nil
}
c.data[name] = newItem
return newItem, nil
}
func (c *ConcurrentCache) DefineItem(name string, typ string, config map[string]interface{}, cache *Cache, build CacheBuild) (interface{}, error) {
c.mutex.RLock()
_, cached := c.data[name]
if cached {
c.mutex.RUnlock()
return nil, ErrAlreadyDefined
}
// give up read lock so others lookups can proceed
c.mutex.RUnlock()
// really not there, try to build it
newItem, err := build(typ, config, cache)
if err != nil {
return nil, err
}
// now we've built it, acquire lock
c.mutex.Lock()
defer c.mutex.Unlock()
// check again because it could have been created while trading locks
_, cached = c.data[name]
if cached {
return nil, ErrAlreadyDefined
}
c.data[name] = newItem
return newItem, nil
}

@ -0,0 +1,89 @@
// Copyright (c) 2014 Couchbase, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package registry
import (
"fmt"
"github.com/blevesearch/bleve/analysis"
)
func RegisterCharFilter(name string, constructor CharFilterConstructor) {
_, exists := charFilters[name]
if exists {
panic(fmt.Errorf("attempted to register duplicate char filter named '%s'", name))
}
charFilters[name] = constructor
}
type CharFilterConstructor func(config map[string]interface{}, cache *Cache) (analysis.CharFilter, error)
type CharFilterRegistry map[string]CharFilterConstructor
type CharFilterCache struct {
*ConcurrentCache
}
func NewCharFilterCache() *CharFilterCache {
return &CharFilterCache{
NewConcurrentCache(),
}
}
func CharFilterBuild(name string, config map[string]interface{}, cache *Cache) (interface{}, error) {
cons, registered := charFilters[name]
if !registered {
return nil, fmt.Errorf("no char filter with name or type '%s' registered", name)
}
charFilter, err := cons(config, cache)
if err != nil {
return nil, fmt.Errorf("error building char filter: %v", err)
}
return charFilter, nil
}
func (c *CharFilterCache) CharFilterNamed(name string, cache *Cache) (analysis.CharFilter, error) {
item, err := c.ItemNamed(name, cache, CharFilterBuild)
if err != nil {
return nil, err
}
return item.(analysis.CharFilter), nil
}
func (c *CharFilterCache) DefineCharFilter(name string, typ string, config map[string]interface{}, cache *Cache) (analysis.CharFilter, error) {
item, err := c.DefineItem(name, typ, config, cache, CharFilterBuild)
if err != nil {
if err == ErrAlreadyDefined {
return nil, fmt.Errorf("char filter named '%s' already defined", name)
}
return nil, err
}
return item.(analysis.CharFilter), nil
}
func CharFilterTypesAndInstances() ([]string, []string) {
emptyConfig := map[string]interface{}{}
emptyCache := NewCache()
var types []string
var instances []string
for name, cons := range charFilters {
_, err := cons(emptyConfig, emptyCache)
if err == nil {
instances = append(instances, name)
} else {
types = append(types, name)
}
}
return types, instances
}

Some files were not shown because too many files have changed in this diff Show More

Loading…
Cancel
Save