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
147
148
149
150
151
152
153
154
|
package internal
import (
"regexp"
"sort"
"strings"
"sync"
"github.com/jdkato/prose/chunk"
"github.com/jdkato/prose/tag"
"github.com/jdkato/prose/tokenize"
)
var (
licenseMarkReadmeRe = regexp.MustCompile(`(?i)(copy(right|ing))|\(c\)|©|(licen[cs][ei])|released under`)
garbageReadmeRe = regexp.MustCompile(`([Cc]opy(right|ing))|\(c\)|©`)
licenseReadmeRe = regexp.MustCompile(`\s*[Ll]icen[cs]e\s*`)
licenseNamePartRe = regexp.MustCompile(`([a-z]+)|([0-9]+)`)
digitsRe = regexp.MustCompile(`[0-9]+`)
disabledNamePartsRe = regexp.MustCompile(`clause|or|only|deprecated|later`)
tagger = tag.NewPerceptronTagger()
chunkLock sync.Mutex
)
// investigateReadmeFile uses NER to match license name mentions.
// It takes two arguments: licenseNameParts and licenseNameSizes.
// The idea is to map substrings to real licenses, and the confidence is
// <the number of matches> / <overall number of substrings>.
func investigateReadmeFile(
text string, licenseNameParts map[string][]substring,
licenseNameSizes map[string]int) map[string]float32 {
matches := licenseMarkReadmeRe.FindAllStringIndex(text, -1)
if len(matches) == 0 {
return map[string]float32{}
}
// shoot in the dark. Is it a license text?
beginIndex := matches[0][0]
for ; beginIndex >= 1 && text[beginIndex-1:beginIndex+1] != "\n\n"; beginIndex-- {
}
endIndex := matches[len(matches)-1][1]
for ; endIndex < len(text)-1 && text[endIndex:endIndex+2] != "\n\n"; endIndex++ {
}
candidates := globalLicenseDatabase().QueryLicenseText(text[beginIndex:endIndex])
beginIndex = matches[0][0]
endIndex = beginIndex + 50
if len(matches) > 1 {
endIndex = matches[len(matches)-1][1]
} else {
beginIndex -= 50
if beginIndex < 0 {
beginIndex = 0
} else {
for ; text[beginIndex] != ' ' && text[beginIndex] != '\t' &&
text[beginIndex] != '\n' && beginIndex < matches[0][0]; beginIndex++ {
}
}
for ; endIndex < len(text) && text[endIndex] != ' ' && text[endIndex] != '\t' &&
text[endIndex] != '\n'; endIndex++ {
}
}
if endIndex > len(text) {
endIndex = len(text)
}
suspectedText := text[beginIndex:endIndex]
suspectedWords := tokenize.TextToWords(suspectedText)
chunks := readmeChunks(tagger.Tag(suspectedWords))
for _, entity := range chunks {
if garbageReadmeRe.MatchString(entity) {
continue
}
scores := map[string]map[string]int{}
entity = licenseReadmeRe.ReplaceAllString(entity, " ")
substrs := splitLicenseName(entity)
for _, substr := range substrs {
for _, match := range licenseNameParts[substr.value] {
common := match.count
if substr.count < common {
common = substr.count
}
matchSubstrs := scores[match.value]
if matchSubstrs == nil {
matchSubstrs = map[string]int{}
scores[match.value] = matchSubstrs
}
matchSubstrs[substr.value] = common
}
}
// if the only reason a license matched is a single digit, drop it
toRemove := []string{}
for key, matchSubstrs := range scores {
if len(matchSubstrs) == 1 {
for substr := range matchSubstrs {
if digitsRe.MatchString(substr) {
toRemove = append(toRemove, key)
}
}
}
}
for _, key := range toRemove {
delete(scores, key)
}
for key, val := range scores {
matchSize := 0
for _, n := range val {
matchSize += n
}
confidence := float32(matchSize) / float32(licenseNameSizes[key])
if candidates[key] < confidence && confidence >= 0.3 {
candidates[key] = confidence
}
}
}
return candidates
}
func readmeChunks(tokens []tag.Token) []string {
chunkLock.Lock()
defer chunkLock.Unlock()
return chunk.Chunk(tokens, chunk.TreebankNamedEntities)
}
func splitLicenseName(name string) []substring {
counts := map[string]int{}
parts := licenseNamePartRe.FindAllString(strings.ToLower(name), -1)
for i, part := range parts {
if part[len(part)-1] == 'v' && i < len(parts)-1 && digitsRe.MatchString(parts[i+1]) {
part = part[:len(part)-1]
if len(part) == 0 {
continue
}
}
if disabledNamePartsRe.MatchString(part) {
continue
}
// BSD hack
if part == "simplified" {
part = "2"
}
counts[part]++
}
result := make([]substring, len(counts))
i := 0
for key, val := range counts {
result[i] = substring{value: key, count: val}
i++
}
sort.Slice(result, func(i int, j int) bool {
return result[i].value > result[j].value
})
return result
}
|
