better indexing and classifier

This commit is contained in:
ston1th 2019-05-03 00:05:51 +02:00
commit 98b53d3a33
9 changed files with 168 additions and 32 deletions

View file

@ -7,16 +7,25 @@ import (
"git.giftfish.de/ston1th/docstore/pkg/core"
"git.giftfish.de/ston1th/godrop/v2"
"github.com/blevesearch/bleve"
"github.com/blevesearch/bleve/mapping"
"html/template"
"os"
"strings"
"time"
)
const maxSearchResult = 101
type indexDocument struct {
Text string `json:"text"`
Date string `json:"date"`
Type string `json:"type"`
const docType = "document"
type document struct {
Text string `json:"text"`
Keyword []string `json:"keyword"`
Created time.Time `json:"created"`
}
func (d *document) Type() string {
return docType
}
type Index struct {
@ -30,8 +39,10 @@ func NewIndex(path string) (i *Index, err error) {
return
}
if _, err = os.Stat(path); os.IsNotExist(err) {
mapping := bleve.NewIndexMapping()
bi, err = bleve.New(path, mapping)
var m *mapping.IndexMappingImpl
m, err = createMapping()
//m := bleve.NewIndexMapping()
bi, err = bleve.New(path, m)
if err != nil {
err = errors.New("index: " + err.Error())
return
@ -51,7 +62,7 @@ func (i *Index) Close() error {
return i.i.Close()
}
func (i *Index) Add(text, docType string) (id string, err error) {
func (i *Index) Add(text string) (id string, err error) {
id = newID()
for {
doc, _ := i.i.Document(id)
@ -60,7 +71,8 @@ func (i *Index) Add(text, docType string) (id string, err error) {
}
id = newID()
}
err = i.i.Index(id, indexDocument{text, core.Now(), docType})
k := classify(text)
err = i.i.Index(id, &document{text, k, time.Now()})
return
}
@ -86,12 +98,20 @@ func (i *Index) Search(search string) (results []core.Result, err error) {
continue
}
for fragField, frags := range hit.Fragments {
if fragField == "text" || fragField == "type" || fragField == "date" {
if fragField == "text" || fragField == "created" {
for _, f := range frags {
r.HTML += template.HTML(f)
}
break
}
if fragField == "keyword" {
var k string
for _, f := range frags {
k += f
}
r.Keyword = strings.ReplaceAll(strings.ReplaceAll(k, "<mark>", ""), "</mark>", "")
break
}
}
results = append(results, r)
}