2014-08-14 03:14:47 +02:00
|
|
|
|
package el
|
|
|
|
|
|
|
|
|
|
import (
|
2014-08-28 21:38:57 +02:00
|
|
|
|
"github.com/blevesearch/bleve/analysis"
|
|
|
|
|
"github.com/blevesearch/bleve/registry"
|
2014-08-14 03:14:47 +02:00
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
const StopName = "stop_el"
|
2014-07-31 01:29:52 +02:00
|
|
|
|
|
|
|
|
|
// this content was obtained from:
|
|
|
|
|
// lucene-4.7.2/analysis/common/src/resources/org/apache/lucene/analysis/
|
|
|
|
|
// ` was changed to ' to allow for literal string
|
|
|
|
|
|
|
|
|
|
var GreekStopWords = []byte(`# Lucene Greek Stopwords list
|
|
|
|
|
# Note: by default this file is used after GreekLowerCaseFilter,
|
|
|
|
|
# so when modifying this file use 'σ' instead of 'ς'
|
|
|
|
|
ο
|
|
|
|
|
η
|
|
|
|
|
το
|
|
|
|
|
οι
|
|
|
|
|
τα
|
|
|
|
|
του
|
|
|
|
|
τησ
|
|
|
|
|
των
|
|
|
|
|
τον
|
|
|
|
|
την
|
|
|
|
|
και
|
|
|
|
|
κι
|
|
|
|
|
κ
|
|
|
|
|
ειμαι
|
|
|
|
|
εισαι
|
|
|
|
|
ειναι
|
|
|
|
|
ειμαστε
|
|
|
|
|
ειστε
|
|
|
|
|
στο
|
|
|
|
|
στον
|
|
|
|
|
στη
|
|
|
|
|
στην
|
|
|
|
|
μα
|
|
|
|
|
αλλα
|
|
|
|
|
απο
|
|
|
|
|
για
|
|
|
|
|
προσ
|
|
|
|
|
με
|
|
|
|
|
σε
|
|
|
|
|
ωσ
|
|
|
|
|
παρα
|
|
|
|
|
αντι
|
|
|
|
|
κατα
|
|
|
|
|
μετα
|
|
|
|
|
θα
|
|
|
|
|
να
|
|
|
|
|
δε
|
|
|
|
|
δεν
|
|
|
|
|
μη
|
|
|
|
|
μην
|
|
|
|
|
επι
|
|
|
|
|
ενω
|
|
|
|
|
εαν
|
|
|
|
|
αν
|
|
|
|
|
τοτε
|
|
|
|
|
που
|
|
|
|
|
πωσ
|
|
|
|
|
ποιοσ
|
|
|
|
|
ποια
|
|
|
|
|
ποιο
|
|
|
|
|
ποιοι
|
|
|
|
|
ποιεσ
|
|
|
|
|
ποιων
|
|
|
|
|
ποιουσ
|
|
|
|
|
αυτοσ
|
|
|
|
|
αυτη
|
|
|
|
|
αυτο
|
|
|
|
|
αυτοι
|
|
|
|
|
αυτων
|
|
|
|
|
αυτουσ
|
|
|
|
|
αυτεσ
|
|
|
|
|
αυτα
|
|
|
|
|
εκεινοσ
|
|
|
|
|
εκεινη
|
|
|
|
|
εκεινο
|
|
|
|
|
εκεινοι
|
|
|
|
|
εκεινεσ
|
|
|
|
|
εκεινα
|
|
|
|
|
εκεινων
|
|
|
|
|
εκεινουσ
|
|
|
|
|
οπωσ
|
|
|
|
|
ομωσ
|
|
|
|
|
ισωσ
|
|
|
|
|
οσο
|
|
|
|
|
οτι
|
|
|
|
|
`)
|
2014-08-14 03:14:47 +02:00
|
|
|
|
|
|
|
|
|
func TokenMapConstructor(config map[string]interface{}, cache *registry.Cache) (analysis.TokenMap, error) {
|
|
|
|
|
rv := analysis.NewTokenMap()
|
|
|
|
|
err := rv.LoadBytes(GreekStopWords)
|
|
|
|
|
return rv, err
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func init() {
|
|
|
|
|
registry.RegisterTokenMap(StopName, TokenMapConstructor)
|
|
|
|
|
}
|