-
Notifications
You must be signed in to change notification settings - Fork 0
/
pattern.go
280 lines (224 loc) · 6.27 KB
/
pattern.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
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
package hre
import (
"bytes"
"fmt"
"regexp"
"github.com/pkg/errors"
"github.com/hangulize/stringset"
)
// Pattern represents an HRE (Hangulize-specific Regular Expression) pattern.
//
// The transcription logic includes several rewriting rules. A rule has a
// Pattern and an RPattern. A sub-word which is matched with the Pattern, will
// be rewritten by the RPattern.
//
// rewrite:
// "'" -> ""
// "^gli$" -> "li"
// "^glia$" -> "g.lia"
// "^glioma$" -> "g.lioma"
// "^gli{@}" -> "li"
// "{@}gli" -> "li"
// "gn{@}" -> "nJ"
// "gn" -> "n"
//
// Some expressions in Pattern have special meaning:
//
// "^" // start of chunk
// "^^" // start of string
// "$" // end of chunk
// "$$" // end of string
// "{...}" // zero-width match
// "{~...}" // zero-width negative match
// "<var>" // one of var values (defined in spec)
//
type Pattern struct {
expr string
re *regexp.Regexp // positive regexp
negA *regexp.Regexp // negative ahead regexp
negB *regexp.Regexp // negative behind regexp
negAWidth int // max width of nagative lookahead
negBWidth int // max width of nagative lookbehind
// Letters used in the positive/negative regexps.
letters stringset.StringSet
// References to expanded vars.
usedVars [][]string
}
func (p *Pattern) String() string {
return fmt.Sprintf(`"%s"`, p.expr)
}
// -----------------------------------------------------------------------------
// NewPattern compiles an HRE pattern from an expression.
func NewPattern(
expr string,
macros map[string]string,
vars map[string][]string,
) (*Pattern, error) {
if len(expr) == 0 {
return nil, errors.New("empty pattern not allowed")
}
reExpr := expr
reExpr = expandMacros(reExpr, macros)
reExpr, usedVars := expandVars(reExpr, vars)
reExpr, negAExpr, negBExpr, negAWidth, negBWidth, err :=
expandLookaround(reExpr)
if err != nil {
return nil, errors.Wrapf(err, "failed to compile pattern: %#v", expr)
}
reExpr = expandEdges(reExpr)
// Collect letters in the regexps.
combinedExpr := reExpr + negAExpr + negBExpr
letters := stringset.NewStringSet(splitLetters(regexpLetters(combinedExpr))...)
// Compile regexp.
re, err := regexp.Compile(reExpr)
if err != nil {
return nil, errors.Wrapf(err, "failed to compile pattern: %#v", expr)
}
// Compile negative lookahead/behind regexps.
var negA *regexp.Regexp
var negB *regexp.Regexp
if negAExpr != `` {
negA, err = regexp.Compile(negAExpr)
if err != nil {
err = errors.Wrapf(err, "failed to compile pattern: %#v", expr)
return nil, err
}
}
if negBExpr != `` {
negB, err = regexp.Compile(negBExpr)
if err != nil {
err = errors.Wrapf(err, "failed to compile pattern: %#v", expr)
return nil, err
}
}
p := &Pattern{expr, re, negA, negB, negAWidth, negBWidth, letters, usedVars}
return p, nil
}
// -----------------------------------------------------------------------------
// Letters returns the set of natural letters used in the expression in
// ascending order.
func (p *Pattern) Letters() []string {
return p.letters.Array()
}
// Explain shows the HRE expression with
// the underlying standard regexp patterns.
func (p *Pattern) Explain() string {
if p == nil {
return fmt.Sprintf("%#v", nil)
}
return fmt.Sprintf(
"expr:/%s/, re:/%s/, negA:/%s/, negB:/%s/",
p.expr, p.re, p.negA, p.negB,
)
}
// NegativeLookaroundWidths returns the potential widths of negative lookahead
// and negative lookbehind.
//
// -1 means unlimited. An unlimited negative lookround width leads to a
// polynominal time to match. Otherwise, the match consumes only a linear time.
func (p *Pattern) NegativeLookaroundWidths() (negAWidth int, negBWidth int) {
negAWidth, negBWidth = p.negAWidth, p.negBWidth
return
}
// -----------------------------------------------------------------------------
func pickStartStop(m []int) (int, int) {
start := m[5]
if start == -1 {
start = m[0]
}
stop := m[len(m)-4]
if stop == -1 {
stop = m[1]
}
return start, stop
}
// Find searches up to n matches in the word. If n is -1, it will search all
// matches. The result is an array of submatch locations.
func (p *Pattern) Find(word string, n int) [][]int {
var matches [][]int
offset := 0
length := len(word)
for offset < length && (n < 0 || len(matches) < n) {
// Find submatches on a shifted word.
m := p.re.FindStringSubmatchIndex(word[offset:])
for i := range m {
if m[i] != -1 {
m[i] += offset
}
}
// Submatches look like:
//
// 0 ┌4 ┌5 -2┐ -1┐
// └(edge)(look)abc(look)(edge)┐
// └2 └3 -4┘ -3┘ 1
//
lenM := len(m)
if lenM == 0 {
// No more match.
break
}
if lenM < 10 {
// Not expected number of groups.
panic(fmt.Errorf("unexpected submatches from %v: %v", p, m))
}
// Pick the actual start and stop.
start, stop := pickStartStop(m)
// The match MUST NOT be zero-width.
if stop-start == 0 {
panic(fmt.Errorf("zero-width match from %v", p))
}
// Shift the cursor for the next iteration.
offset = stop
// Test negative lookahead.
if p.negA != nil {
var negAStart, negAStop int
negAStart = m[lenM-4]
if p.negAWidth == -1 {
negAStop = len(word)
} else {
negAStop = m[lenM-4] + p.negAWidth
}
if p.negA.MatchString(substr(word, negAStart, negAStop)) {
continue
}
}
// Test negative lookbehind.
if p.negB != nil {
var negBStart, negBStop int
negBStop = m[5]
if p.negBWidth == -1 {
negBStart = 0
} else {
negBStart = m[5] - p.negBWidth
}
if p.negB.MatchString(substr(word, negBStart, negBStop)) {
continue
}
}
// Successfully matched.
match := []int{start, stop}
// Keep submatches in the core match.
match = append(match, m[6:lenM-4]...)
// Export the match.
matches = append(matches, match)
}
return matches
}
// Replace finds matches and replaces by the given RPattern.
func (p *Pattern) Replace(word string, rpat *RPattern, n int) string {
var buf bytes.Buffer
cur := 0
for _, m := range p.Find(word, n) {
start, stop := m[0], m[1]
buf.WriteString(word[cur:start])
repl, err := rpat.Interpolate(p, word, m)
if err == nil {
buf.WriteString(repl)
} else {
buf.WriteString(word[start:stop])
}
cur = stop
}
buf.WriteString(word[cur:])
return buf.String()
}