forked from dgraph-io/tokenizer
-
Notifications
You must be signed in to change notification settings - Fork 0
/
tokenizer_test.go
146 lines (129 loc) · 3.07 KB
/
tokenizer_test.go
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
package tokenizer
import (
"os"
"sort"
"strings"
"testing"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/analysis/analyzer/custom"
"github.com/blevesearch/bleve/analysis/token/lowercase"
"github.com/blevesearch/bleve/analysis/token/unicodenorm"
"github.com/blevesearch/bleve/analysis/tokenizer/unicode"
"github.com/blevesearch/bleve/registry"
"github.com/go-nlp/bpe"
"github.com/go-nlp/corpus"
"github.com/stretchr/testify/require"
)
var tokz *Tokenizer
func init() {
var err error
tokz, err = testingTokenizer()
if err != nil {
panic(err)
}
}
func testingTokenizer() (*Tokenizer, error) {
f, err := os.Open("testdata/corpus_zh.txt")
if err != nil {
return nil, err
}
normalizer := NewNormalizer()
enNorm := func(a string) string {
a = normalizer.MustNorm(a)
return strings.ToLower(a)
}
norm := func(a string) string {
//return strings.Replace(enNorm(a), " ", "", -1)
return a
}
c, err := corpus.FromTextCorpus(f, nil, norm)
if err != nil {
return nil, err
}
g, err := os.Open("testdata/corpus_en.txt")
if err != nil {
return nil, err
}
c2, err := corpus.FromTextCorpus(g, nil, enNorm)
if err != nil {
return nil, err
}
c.Merge(c2)
enc, err := bpe.Learn(c, 6000, 2, false)
if err != nil {
return nil, err
}
tok := NewTokenizer(enc)
return tok, nil
}
func TestTokenizer(t *testing.T) {
tok := tokz
input := "我现在在吃rabbit soup"
tokens, err := tok.Tokenize(input)
require.NoError(t, err)
t.Logf("input %q. tokens %q", input, tokens)
}
// uniqueTerms takes a token stream and returns a string slice of unique terms.
func uniqueTerms(tokens analysis.TokenStream) []string {
var terms []string
for i := range tokens {
terms = append(terms, string(tokens[i].Term))
}
terms = RemoveDuplicates(terms)
return terms
}
// RemoveDuplicates sorts the slice of strings and removes duplicates. changes the input slice.
// This function should be called like: someSlice = RemoveDuplicates(someSlice)
func RemoveDuplicates(s []string) (out []string) {
sort.Strings(s)
out = s[:0]
for i := range s {
if i > 0 && s[i] == s[i-1] {
continue
}
out = append(out, s[i])
}
return
}
func BenchmarkBleve(b *testing.B) {
b.StopTimer()
// set up
unicodenormName := "unicodenorm_nfkc"
bleveCache := registry.NewCache()
_, err := bleveCache.DefineTokenFilter(unicodenormName,
map[string]interface{}{
"type": unicodenorm.Name,
"form": unicodenorm.NFKC,
})
termAnalyzer, err := bleveCache.DefineAnalyzer("term",
map[string]interface{}{
"type": custom.Name,
"tokenizer": unicode.Name,
"token_filters": []string{
lowercase.Name,
unicodenormName,
},
})
if err != nil {
b.Fatal(err)
}
b.ResetTimer()
b.StartTimer()
var tokens []string
for i := 0; i < b.N; i++ {
tks := termAnalyzer.Analyze([]byte("我现在在吃rabbit stew"))
tokens = uniqueTerms(tks)
}
_ = tokens
}
func BenchmarkTokenizer(b *testing.B) {
b.StopTimer()
// set up
b.ResetTimer()
b.StartTimer()
var tokens []string
for i := 0; i < b.N; i++ {
tokens, _ = tokz.Tokenize("我现在在吃rabbit stew")
}
_ = tokens
}