Files
Gists/internal/index/bleve.go
2025-12-26 22:36:28 +08:00

208 lines
5.9 KiB
Go

package index
import (
"errors"
"strconv"
"github.com/blevesearch/bleve/v2"
"github.com/blevesearch/bleve/v2/analysis/analyzer/custom"
"github.com/blevesearch/bleve/v2/analysis/token/camelcase"
"github.com/blevesearch/bleve/v2/analysis/token/lowercase"
"github.com/blevesearch/bleve/v2/analysis/token/unicodenorm"
"github.com/blevesearch/bleve/v2/analysis/tokenizer/unicode"
"github.com/blevesearch/bleve/v2/search/query"
"github.com/rs/zerolog/log"
)
type BleveIndexer struct {
index bleve.Index
path string
}
func NewBleveIndexer(path string) *BleveIndexer {
return &BleveIndexer{path: path}
}
func (i *BleveIndexer) Init() error {
errChan := make(chan error, 1)
go func() {
bleveIndex, err := i.open()
if err != nil {
log.Error().Err(err).Msg("Failed to open Bleve index")
i.Close()
errChan <- err
return
}
i.index = bleveIndex
log.Info().Msg("Bleve indexer initialized")
errChan <- nil
}()
return <-errChan
}
func (i *BleveIndexer) open() (bleve.Index, error) {
bleveIndex, err := bleve.Open(i.path)
if err == nil {
return bleveIndex, nil
}
if !errors.Is(err, bleve.ErrorIndexPathDoesNotExist) {
return nil, err
}
docMapping := bleve.NewDocumentMapping()
docMapping.AddFieldMappingsAt("GistID", bleve.NewNumericFieldMapping())
docMapping.AddFieldMappingsAt("UserID", bleve.NewNumericFieldMapping())
docMapping.AddFieldMappingsAt("Visibility", bleve.NewNumericFieldMapping())
docMapping.AddFieldMappingsAt("Content", bleve.NewTextFieldMapping())
mapping := bleve.NewIndexMapping()
if err = mapping.AddCustomTokenFilter("unicodeNormalize", map[string]any{
"type": unicodenorm.Name,
"form": unicodenorm.NFC,
}); err != nil {
return nil, err
}
if err = mapping.AddCustomAnalyzer("gistAnalyser", map[string]interface{}{
"type": custom.Name,
"char_filters": []string{},
"tokenizer": unicode.Name,
"token_filters": []string{"unicodeNormalize", camelcase.Name, lowercase.Name},
}); err != nil {
return nil, err
}
docMapping.DefaultAnalyzer = "gistAnalyser"
mapping.DefaultMapping = docMapping
return bleve.New(i.path, mapping)
}
func (i *BleveIndexer) Close() {
if i == nil || i.index == nil {
return
}
err := i.index.Close()
if err != nil {
log.Error().Err(err).Msg("Failed to close Bleve index")
}
log.Info().Msg("Bleve indexer closed")
}
func (i *BleveIndexer) Add(gist *Gist) error {
if gist == nil {
return errors.New("failed to add nil gist to index")
}
return (*atomicIndexer.Load()).(*BleveIndexer).index.Index(strconv.Itoa(int(gist.GistID)), gist)
}
func (i *BleveIndexer) Remove(gistID uint) error {
return (*atomicIndexer.Load()).(*BleveIndexer).index.Delete(strconv.Itoa(int(gistID)))
}
func (i *BleveIndexer) Search(queryStr string, queryMetadata SearchGistMetadata, userId uint, page int) ([]uint, uint64, map[string]int, error) {
var err error
var indexerQuery query.Query
if queryStr != "" {
// Use match query with fuzzy matching for more flexible content search
contentQuery := bleve.NewMatchQuery(queryStr)
contentQuery.SetField("Content")
contentQuery.SetFuzziness(2)
indexerQuery = contentQuery
} else {
contentQuery := bleve.NewMatchAllQuery()
indexerQuery = contentQuery
}
// Visibility filtering: show public gists (Visibility=0) OR user's own gists
visibilityZero := float64(0)
truee := true
publicQuery := bleve.NewNumericRangeInclusiveQuery(&visibilityZero, &visibilityZero, &truee, &truee)
publicQuery.SetField("Visibility")
userIdMatch := float64(userId)
userIdQuery := bleve.NewNumericRangeInclusiveQuery(&userIdMatch, &userIdMatch, &truee, &truee)
userIdQuery.SetField("UserID")
accessQuery := bleve.NewDisjunctionQuery(publicQuery, userIdQuery)
indexerQuery = bleve.NewConjunctionQuery(accessQuery, indexerQuery)
// Handle "All" field - search across all metadata fields with OR logic
if queryMetadata.All != "" {
allQueries := make([]query.Query, 0)
// Create match phrase queries for each field
fields := []struct {
field string
value string
}{
{"Username", queryMetadata.All},
{"Title", queryMetadata.All},
{"Extensions", "." + queryMetadata.All},
{"Filenames", queryMetadata.All},
{"Languages", queryMetadata.All},
{"Topics", queryMetadata.All},
}
for _, f := range fields {
q := bleve.NewMatchPhraseQuery(f.value)
q.FieldVal = f.field
allQueries = append(allQueries, q)
}
// Combine all field queries with OR (disjunction)
allDisjunction := bleve.NewDisjunctionQuery(allQueries...)
indexerQuery = bleve.NewConjunctionQuery(indexerQuery, allDisjunction)
} else {
// Original behavior: add each metadata field with AND logic
addQuery := func(field, value string) {
if value != "" && value != "." {
q := bleve.NewMatchPhraseQuery(value)
q.FieldVal = field
indexerQuery = bleve.NewConjunctionQuery(indexerQuery, q)
}
}
addQuery("Username", queryMetadata.Username)
addQuery("Title", queryMetadata.Title)
addQuery("Extensions", "."+queryMetadata.Extension)
addQuery("Filenames", queryMetadata.Filename)
addQuery("Languages", queryMetadata.Language)
addQuery("Topics", queryMetadata.Topic)
}
languageFacet := bleve.NewFacetRequest("Languages", 10)
perPage := 10
offset := (page - 1) * perPage
s := bleve.NewSearchRequestOptions(indexerQuery, perPage+1, offset, false)
s.AddFacet("languageFacet", languageFacet)
s.Fields = []string{"GistID"}
s.IncludeLocations = false
results, err := (*atomicIndexer.Load()).(*BleveIndexer).index.Search(s)
if err != nil {
return nil, 0, nil, err
}
gistIds := make([]uint, 0, len(results.Hits))
for _, hit := range results.Hits {
gistIds = append(gistIds, uint(hit.Fields["GistID"].(float64)))
}
languageCounts := make(map[string]int)
if facets, found := results.Facets["languageFacet"]; found {
for _, term := range facets.Terms.Terms() {
languageCounts[term.Term] = term.Count
}
}
return gistIds, results.Total, languageCounts, nil
}