Improve issue search (#2387)

* Improve issue indexer

* Fix new issue sqlite bug

* Different test indexer paths for each db

* Add integration indexer paths to make clean
This commit is contained in:
Ethan Koenig 2017-09-16 13:16:21 -07:00 committed by Lauris BH
parent 52e11b24bf
commit b0f7457d9e
122 changed files with 15280 additions and 1458 deletions

View file

@ -6,112 +6,21 @@ package models
import (
"fmt"
"os"
"strconv"
"strings"
"code.gitea.io/gitea/modules/indexer"
"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) {
terms := strings.Fields(strings.ToLower(keyword))
indexerQuery := bleve.NewConjunctionQuery(
numericQuery(repoID, "RepoID"),
bleve.NewDisjunctionQuery(
bleve.NewPhraseQuery(terms, "Title"),
bleve.NewPhraseQuery(terms, "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
}
// issueIndexerUpdateQueue queue of issue ids to be updated
var issueIndexerUpdateQueue chan int64
// 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)
indexer.InitIssueIndexer(populateIssueIndexer)
issueIndexerUpdateQueue = make(chan int64, 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
@ -127,57 +36,64 @@ func populateIssueIndexer() error {
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,
})
if err != nil {
return fmt.Errorf("Issues: %v", err)
updates := make([]indexer.IssueIndexerUpdate, len(issues))
for i, issue := range issues {
updates[i] = issue.update()
}
for _, issue := range issues {
err = batch.Index(issue.indexUID(), issue.issueData())
if err != nil {
return fmt.Errorf("batch.Index: %v", err)
}
if err = indexer.BatchUpdateIssues(updates...); err != nil {
return fmt.Errorf("BatchUpdate: %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 {
case issueID := <-issueIndexerUpdateQueue:
issue, err := GetIssueByID(issueID)
if err != nil {
log.Error(4, "issuesIndexer.Index: %v", err)
continue
}
if err = indexer.UpdateIssue(issue.update()); 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,
func (issue *Issue) update() indexer.IssueIndexerUpdate {
comments := make([]string, 0, 5)
for _, comment := range issue.Comments {
if comment.Type == CommentTypeComment {
comments = append(comments, comment.Content)
}
}
return indexer.IssueIndexerUpdate{
IssueID: issue.ID,
Data: &indexer.IssueIndexerData{
RepoID: issue.RepoID,
Title: issue.Title,
Content: issue.Content,
Comments: comments,
},
}
}
// UpdateIssueIndexer add/update an issue to the issue indexer
func UpdateIssueIndexer(issue *Issue) {
go func() {
issueIndexerUpdateQueue <- issue
}()
func UpdateIssueIndexer(issueID int64) {
select {
case issueIndexerUpdateQueue <- issueID:
default:
go func() {
issueIndexerUpdateQueue <- issueID
}()
}
}