kabosu 0.6.10.dev.20260226.98055fb → 0.6.10.dev.20260429.b757f55
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +4 -4
- data/Cargo.lock +494 -0
- data/Cargo.toml +6 -0
- data/ext/kabosu/src/dictionary.rs +3 -1
- data/ext/kabosu/src/lib.rs +8 -2
- data/ext/kabosu/src/morpheme.rs +10 -2
- data/ext/kabosu/src/parsing.rs +3 -1
- data/ext/kabosu/src/splitter.rs +2 -1
- data/ext/kabosu/src/tokenizer.rs +3 -1
- data/lib/kabosu/version.rb +1 -1
- metadata +4 -2
checksums.yaml
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
---
|
|
2
2
|
SHA256:
|
|
3
|
-
metadata.gz:
|
|
4
|
-
data.tar.gz:
|
|
3
|
+
metadata.gz: 96eb4ac24dfbacafad9ff3d4da177c047629a50b824b6d71f5959d220b38f11f
|
|
4
|
+
data.tar.gz: 3d70ff51e9634d8d2e7cafb2624aa693c601ea4f46fad4cf534cb7048b4311e4
|
|
5
5
|
SHA512:
|
|
6
|
-
metadata.gz:
|
|
7
|
-
data.tar.gz:
|
|
6
|
+
metadata.gz: 1111993c83cbd77db17dc80c9360b4b37a9f605f87726485fe909ef9e3ad5c788baeda9c0d9ef69527e370ec541822c369e6ba0ba7f57b09e38b6e01f145acbe
|
|
7
|
+
data.tar.gz: 632524affa3a4af3d683fe4e8ed714f035b6a480189c6caba5c9ab487cacb0d83aee641ce58761e260695b0b9d1d46e33b29ba7de72d617ceefb63a54621506b
|
data/Cargo.lock
ADDED
|
@@ -0,0 +1,494 @@
|
|
|
1
|
+
# This file is automatically @generated by Cargo.
|
|
2
|
+
# It is not intended for manual editing.
|
|
3
|
+
version = 4
|
|
4
|
+
|
|
5
|
+
[[package]]
|
|
6
|
+
name = "aho-corasick"
|
|
7
|
+
version = "1.1.4"
|
|
8
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
9
|
+
checksum = "ddd31a130427c27518df266943a5308ed92d4b226cc639f5a8f1002816174301"
|
|
10
|
+
dependencies = [
|
|
11
|
+
"memchr",
|
|
12
|
+
]
|
|
13
|
+
|
|
14
|
+
[[package]]
|
|
15
|
+
name = "bindgen"
|
|
16
|
+
version = "0.72.1"
|
|
17
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
18
|
+
checksum = "993776b509cfb49c750f11b8f07a46fa23e0a1386ffc01fb1e7d343efc387895"
|
|
19
|
+
dependencies = [
|
|
20
|
+
"bitflags",
|
|
21
|
+
"cexpr",
|
|
22
|
+
"clang-sys",
|
|
23
|
+
"itertools",
|
|
24
|
+
"proc-macro2",
|
|
25
|
+
"quote",
|
|
26
|
+
"regex",
|
|
27
|
+
"rustc-hash",
|
|
28
|
+
"shlex",
|
|
29
|
+
"syn",
|
|
30
|
+
]
|
|
31
|
+
|
|
32
|
+
[[package]]
|
|
33
|
+
name = "bit-set"
|
|
34
|
+
version = "0.5.3"
|
|
35
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
36
|
+
checksum = "0700ddab506f33b20a03b13996eccd309a48e5ff77d0d95926aa0210fb4e95f1"
|
|
37
|
+
dependencies = [
|
|
38
|
+
"bit-vec",
|
|
39
|
+
]
|
|
40
|
+
|
|
41
|
+
[[package]]
|
|
42
|
+
name = "bit-vec"
|
|
43
|
+
version = "0.6.3"
|
|
44
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
45
|
+
checksum = "349f9b6a179ed607305526ca489b34ad0a41aed5f7980fa90eb03160b69598fb"
|
|
46
|
+
|
|
47
|
+
[[package]]
|
|
48
|
+
name = "bitflags"
|
|
49
|
+
version = "2.11.1"
|
|
50
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
51
|
+
checksum = "c4512299f36f043ab09a583e57bceb5a5aab7a73db1805848e8fef3c9e8c78b3"
|
|
52
|
+
|
|
53
|
+
[[package]]
|
|
54
|
+
name = "cexpr"
|
|
55
|
+
version = "0.6.0"
|
|
56
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
57
|
+
checksum = "6fac387a98bb7c37292057cffc56d62ecb629900026402633ae9160df93a8766"
|
|
58
|
+
dependencies = [
|
|
59
|
+
"nom",
|
|
60
|
+
]
|
|
61
|
+
|
|
62
|
+
[[package]]
|
|
63
|
+
name = "cfg-if"
|
|
64
|
+
version = "1.0.4"
|
|
65
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
66
|
+
checksum = "9330f8b2ff13f34540b44e946ef35111825727b38d33286ef986142615121801"
|
|
67
|
+
|
|
68
|
+
[[package]]
|
|
69
|
+
name = "clang-sys"
|
|
70
|
+
version = "1.8.1"
|
|
71
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
72
|
+
checksum = "0b023947811758c97c59bf9d1c188fd619ad4718dcaa767947df1cadb14f39f4"
|
|
73
|
+
dependencies = [
|
|
74
|
+
"glob",
|
|
75
|
+
"libc",
|
|
76
|
+
"libloading",
|
|
77
|
+
]
|
|
78
|
+
|
|
79
|
+
[[package]]
|
|
80
|
+
name = "csv"
|
|
81
|
+
version = "1.4.0"
|
|
82
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
83
|
+
checksum = "52cd9d68cf7efc6ddfaaee42e7288d3a99d613d4b50f76ce9827ae0c6e14f938"
|
|
84
|
+
dependencies = [
|
|
85
|
+
"csv-core",
|
|
86
|
+
"itoa",
|
|
87
|
+
"ryu",
|
|
88
|
+
"serde_core",
|
|
89
|
+
]
|
|
90
|
+
|
|
91
|
+
[[package]]
|
|
92
|
+
name = "csv-core"
|
|
93
|
+
version = "0.1.13"
|
|
94
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
95
|
+
checksum = "704a3c26996a80471189265814dbc2c257598b96b8a7feae2d31ace646bb9782"
|
|
96
|
+
dependencies = [
|
|
97
|
+
"memchr",
|
|
98
|
+
]
|
|
99
|
+
|
|
100
|
+
[[package]]
|
|
101
|
+
name = "either"
|
|
102
|
+
version = "1.15.0"
|
|
103
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
104
|
+
checksum = "48c757948c5ede0e46177b7add2e67155f70e33c07fea8284df6576da70b3719"
|
|
105
|
+
|
|
106
|
+
[[package]]
|
|
107
|
+
name = "equivalent"
|
|
108
|
+
version = "1.0.2"
|
|
109
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
110
|
+
checksum = "877a4ace8713b0bcf2a4e7eec82529c029f1d0619886d18145fea96c3ffe5c0f"
|
|
111
|
+
|
|
112
|
+
[[package]]
|
|
113
|
+
name = "fancy-regex"
|
|
114
|
+
version = "0.13.0"
|
|
115
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
116
|
+
checksum = "531e46835a22af56d1e3b66f04844bed63158bc094a628bec1d321d9b4c44bf2"
|
|
117
|
+
dependencies = [
|
|
118
|
+
"bit-set",
|
|
119
|
+
"regex-automata",
|
|
120
|
+
"regex-syntax",
|
|
121
|
+
]
|
|
122
|
+
|
|
123
|
+
[[package]]
|
|
124
|
+
name = "glob"
|
|
125
|
+
version = "0.3.3"
|
|
126
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
127
|
+
checksum = "0cc23270f6e1808e30a928bdc84dea0b9b4136a8bc82338574f23baf47bbd280"
|
|
128
|
+
|
|
129
|
+
[[package]]
|
|
130
|
+
name = "hashbrown"
|
|
131
|
+
version = "0.17.0"
|
|
132
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
133
|
+
checksum = "4f467dd6dccf739c208452f8014c75c18bb8301b050ad1cfb27153803edb0f51"
|
|
134
|
+
|
|
135
|
+
[[package]]
|
|
136
|
+
name = "indexmap"
|
|
137
|
+
version = "2.14.0"
|
|
138
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
139
|
+
checksum = "d466e9454f08e4a911e14806c24e16fba1b4c121d1ea474396f396069cf949d9"
|
|
140
|
+
dependencies = [
|
|
141
|
+
"equivalent",
|
|
142
|
+
"hashbrown",
|
|
143
|
+
]
|
|
144
|
+
|
|
145
|
+
[[package]]
|
|
146
|
+
name = "itertools"
|
|
147
|
+
version = "0.13.0"
|
|
148
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
149
|
+
checksum = "413ee7dfc52ee1a4949ceeb7dbc8a33f2d6c088194d9f922fb8318faf1f01186"
|
|
150
|
+
dependencies = [
|
|
151
|
+
"either",
|
|
152
|
+
]
|
|
153
|
+
|
|
154
|
+
[[package]]
|
|
155
|
+
name = "itoa"
|
|
156
|
+
version = "1.0.18"
|
|
157
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
158
|
+
checksum = "8f42a60cbdf9a97f5d2305f08a87dc4e09308d1276d28c869c684d7777685682"
|
|
159
|
+
|
|
160
|
+
[[package]]
|
|
161
|
+
name = "kabosu"
|
|
162
|
+
version = "0.6.10"
|
|
163
|
+
dependencies = [
|
|
164
|
+
"magnus",
|
|
165
|
+
"sudachi",
|
|
166
|
+
]
|
|
167
|
+
|
|
168
|
+
[[package]]
|
|
169
|
+
name = "lazy_static"
|
|
170
|
+
version = "1.5.0"
|
|
171
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
172
|
+
checksum = "bbd2bcb4c963f2ddae06a2efc7e9f3591312473c50c6685e1f298068316e66fe"
|
|
173
|
+
|
|
174
|
+
[[package]]
|
|
175
|
+
name = "libc"
|
|
176
|
+
version = "0.2.186"
|
|
177
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
178
|
+
checksum = "68ab91017fe16c622486840e4c83c9a37afeff978bd239b5293d61ece587de66"
|
|
179
|
+
|
|
180
|
+
[[package]]
|
|
181
|
+
name = "libloading"
|
|
182
|
+
version = "0.8.9"
|
|
183
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
184
|
+
checksum = "d7c4b02199fee7c5d21a5ae7d8cfa79a6ef5bb2fc834d6e9058e89c825efdc55"
|
|
185
|
+
dependencies = [
|
|
186
|
+
"cfg-if",
|
|
187
|
+
"windows-link",
|
|
188
|
+
]
|
|
189
|
+
|
|
190
|
+
[[package]]
|
|
191
|
+
name = "magnus"
|
|
192
|
+
version = "0.8.2"
|
|
193
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
194
|
+
checksum = "3b36a5b126bbe97eb0d02d07acfeb327036c6319fd816139a49824a83b7f9012"
|
|
195
|
+
dependencies = [
|
|
196
|
+
"magnus-macros",
|
|
197
|
+
"rb-sys",
|
|
198
|
+
"rb-sys-env",
|
|
199
|
+
"seq-macro",
|
|
200
|
+
]
|
|
201
|
+
|
|
202
|
+
[[package]]
|
|
203
|
+
name = "magnus-macros"
|
|
204
|
+
version = "0.8.0"
|
|
205
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
206
|
+
checksum = "47607461fd8e1513cb4f2076c197d8092d921a1ea75bd08af97398f593751892"
|
|
207
|
+
dependencies = [
|
|
208
|
+
"proc-macro2",
|
|
209
|
+
"quote",
|
|
210
|
+
"syn",
|
|
211
|
+
]
|
|
212
|
+
|
|
213
|
+
[[package]]
|
|
214
|
+
name = "memchr"
|
|
215
|
+
version = "2.8.0"
|
|
216
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
217
|
+
checksum = "f8ca58f447f06ed17d5fc4043ce1b10dd205e060fb3ce5b979b8ed8e59ff3f79"
|
|
218
|
+
|
|
219
|
+
[[package]]
|
|
220
|
+
name = "memmap2"
|
|
221
|
+
version = "0.9.10"
|
|
222
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
223
|
+
checksum = "714098028fe011992e1c3962653c96b2d578c4b4bce9036e15ff220319b1e0e3"
|
|
224
|
+
dependencies = [
|
|
225
|
+
"libc",
|
|
226
|
+
]
|
|
227
|
+
|
|
228
|
+
[[package]]
|
|
229
|
+
name = "minimal-lexical"
|
|
230
|
+
version = "0.2.1"
|
|
231
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
232
|
+
checksum = "68354c5c6bd36d73ff3feceb05efa59b6acb7626617f4962be322a825e61f79a"
|
|
233
|
+
|
|
234
|
+
[[package]]
|
|
235
|
+
name = "nom"
|
|
236
|
+
version = "7.1.3"
|
|
237
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
238
|
+
checksum = "d273983c5a657a70a3e8f2a01329822f3b8c8172b73826411a55751e404a0a4a"
|
|
239
|
+
dependencies = [
|
|
240
|
+
"memchr",
|
|
241
|
+
"minimal-lexical",
|
|
242
|
+
]
|
|
243
|
+
|
|
244
|
+
[[package]]
|
|
245
|
+
name = "proc-macro2"
|
|
246
|
+
version = "1.0.106"
|
|
247
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
248
|
+
checksum = "8fd00f0bb2e90d81d1044c2b32617f68fcb9fa3bb7640c23e9c748e53fb30934"
|
|
249
|
+
dependencies = [
|
|
250
|
+
"unicode-ident",
|
|
251
|
+
]
|
|
252
|
+
|
|
253
|
+
[[package]]
|
|
254
|
+
name = "quote"
|
|
255
|
+
version = "1.0.45"
|
|
256
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
257
|
+
checksum = "41f2619966050689382d2b44f664f4bc593e129785a36d6ee376ddf37259b924"
|
|
258
|
+
dependencies = [
|
|
259
|
+
"proc-macro2",
|
|
260
|
+
]
|
|
261
|
+
|
|
262
|
+
[[package]]
|
|
263
|
+
name = "rb-sys"
|
|
264
|
+
version = "0.9.127"
|
|
265
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
266
|
+
checksum = "d7d7c9560fe42dcffa576941394075f18a17dce89fcf718a2fa90b7dc2134d12"
|
|
267
|
+
dependencies = [
|
|
268
|
+
"rb-sys-build",
|
|
269
|
+
]
|
|
270
|
+
|
|
271
|
+
[[package]]
|
|
272
|
+
name = "rb-sys-build"
|
|
273
|
+
version = "0.9.127"
|
|
274
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
275
|
+
checksum = "f1688e8f32967ba48c89e4dfa283b57f901075f542fc7ee9c3d7c5f9091ca1d9"
|
|
276
|
+
dependencies = [
|
|
277
|
+
"bindgen",
|
|
278
|
+
"lazy_static",
|
|
279
|
+
"proc-macro2",
|
|
280
|
+
"quote",
|
|
281
|
+
"regex",
|
|
282
|
+
"shell-words",
|
|
283
|
+
"syn",
|
|
284
|
+
]
|
|
285
|
+
|
|
286
|
+
[[package]]
|
|
287
|
+
name = "rb-sys-env"
|
|
288
|
+
version = "0.2.3"
|
|
289
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
290
|
+
checksum = "cca7ad6a7e21e72151d56fe2495a259b5670e204c3adac41ee7ef676ea08117a"
|
|
291
|
+
|
|
292
|
+
[[package]]
|
|
293
|
+
name = "regex"
|
|
294
|
+
version = "1.12.3"
|
|
295
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
296
|
+
checksum = "e10754a14b9137dd7b1e3e5b0493cc9171fdd105e0ab477f51b72e7f3ac0e276"
|
|
297
|
+
dependencies = [
|
|
298
|
+
"aho-corasick",
|
|
299
|
+
"memchr",
|
|
300
|
+
"regex-automata",
|
|
301
|
+
"regex-syntax",
|
|
302
|
+
]
|
|
303
|
+
|
|
304
|
+
[[package]]
|
|
305
|
+
name = "regex-automata"
|
|
306
|
+
version = "0.4.14"
|
|
307
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
308
|
+
checksum = "6e1dd4122fc1595e8162618945476892eefca7b88c52820e74af6262213cae8f"
|
|
309
|
+
dependencies = [
|
|
310
|
+
"aho-corasick",
|
|
311
|
+
"memchr",
|
|
312
|
+
"regex-syntax",
|
|
313
|
+
]
|
|
314
|
+
|
|
315
|
+
[[package]]
|
|
316
|
+
name = "regex-syntax"
|
|
317
|
+
version = "0.8.10"
|
|
318
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
319
|
+
checksum = "dc897dd8d9e8bd1ed8cdad82b5966c3e0ecae09fb1907d58efaa013543185d0a"
|
|
320
|
+
|
|
321
|
+
[[package]]
|
|
322
|
+
name = "rustc-hash"
|
|
323
|
+
version = "2.1.2"
|
|
324
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
325
|
+
checksum = "94300abf3f1ae2e2b8ffb7b58043de3d399c73fa6f4b73826402a5c457614dbe"
|
|
326
|
+
|
|
327
|
+
[[package]]
|
|
328
|
+
name = "ryu"
|
|
329
|
+
version = "1.0.23"
|
|
330
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
331
|
+
checksum = "9774ba4a74de5f7b1c1451ed6cd5285a32eddb5cccb8cc655a4e50009e06477f"
|
|
332
|
+
|
|
333
|
+
[[package]]
|
|
334
|
+
name = "seq-macro"
|
|
335
|
+
version = "0.3.6"
|
|
336
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
337
|
+
checksum = "1bc711410fbe7399f390ca1c3b60ad0f53f80e95c5eb935e52268a0e2cd49acc"
|
|
338
|
+
|
|
339
|
+
[[package]]
|
|
340
|
+
name = "serde"
|
|
341
|
+
version = "1.0.228"
|
|
342
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
343
|
+
checksum = "9a8e94ea7f378bd32cbbd37198a4a91436180c5bb472411e48b5ec2e2124ae9e"
|
|
344
|
+
dependencies = [
|
|
345
|
+
"serde_core",
|
|
346
|
+
"serde_derive",
|
|
347
|
+
]
|
|
348
|
+
|
|
349
|
+
[[package]]
|
|
350
|
+
name = "serde_core"
|
|
351
|
+
version = "1.0.228"
|
|
352
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
353
|
+
checksum = "41d385c7d4ca58e59fc732af25c3983b67ac852c1a25000afe1175de458b67ad"
|
|
354
|
+
dependencies = [
|
|
355
|
+
"serde_derive",
|
|
356
|
+
]
|
|
357
|
+
|
|
358
|
+
[[package]]
|
|
359
|
+
name = "serde_derive"
|
|
360
|
+
version = "1.0.228"
|
|
361
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
362
|
+
checksum = "d540f220d3187173da220f885ab66608367b6574e925011a9353e4badda91d79"
|
|
363
|
+
dependencies = [
|
|
364
|
+
"proc-macro2",
|
|
365
|
+
"quote",
|
|
366
|
+
"syn",
|
|
367
|
+
]
|
|
368
|
+
|
|
369
|
+
[[package]]
|
|
370
|
+
name = "serde_json"
|
|
371
|
+
version = "1.0.149"
|
|
372
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
373
|
+
checksum = "83fc039473c5595ace860d8c4fafa220ff474b3fc6bfdb4293327f1a37e94d86"
|
|
374
|
+
dependencies = [
|
|
375
|
+
"itoa",
|
|
376
|
+
"memchr",
|
|
377
|
+
"serde",
|
|
378
|
+
"serde_core",
|
|
379
|
+
"zmij",
|
|
380
|
+
]
|
|
381
|
+
|
|
382
|
+
[[package]]
|
|
383
|
+
name = "shell-words"
|
|
384
|
+
version = "1.1.1"
|
|
385
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
386
|
+
checksum = "dc6fe69c597f9c37bfeeeeeb33da3530379845f10be461a66d16d03eca2ded77"
|
|
387
|
+
|
|
388
|
+
[[package]]
|
|
389
|
+
name = "shlex"
|
|
390
|
+
version = "1.3.0"
|
|
391
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
392
|
+
checksum = "0fda2ff0d084019ba4d7c6f371c95d8fd75ce3524c3cb8fb653a3023f6323e64"
|
|
393
|
+
|
|
394
|
+
[[package]]
|
|
395
|
+
name = "sudachi"
|
|
396
|
+
version = "0.6.10"
|
|
397
|
+
source = "git+https://github.com/WorksApplications/sudachi.rs?tag=v0.6.10#7e2f287bbfffc036421cf960802e41a696727747"
|
|
398
|
+
dependencies = [
|
|
399
|
+
"aho-corasick",
|
|
400
|
+
"bitflags",
|
|
401
|
+
"csv",
|
|
402
|
+
"fancy-regex",
|
|
403
|
+
"indexmap",
|
|
404
|
+
"itertools",
|
|
405
|
+
"lazy_static",
|
|
406
|
+
"libloading",
|
|
407
|
+
"memmap2",
|
|
408
|
+
"nom",
|
|
409
|
+
"regex",
|
|
410
|
+
"serde",
|
|
411
|
+
"serde_json",
|
|
412
|
+
"thiserror",
|
|
413
|
+
"unicode-normalization",
|
|
414
|
+
"yada",
|
|
415
|
+
]
|
|
416
|
+
|
|
417
|
+
[[package]]
|
|
418
|
+
name = "syn"
|
|
419
|
+
version = "2.0.117"
|
|
420
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
421
|
+
checksum = "e665b8803e7b1d2a727f4023456bbbbe74da67099c585258af0ad9c5013b9b99"
|
|
422
|
+
dependencies = [
|
|
423
|
+
"proc-macro2",
|
|
424
|
+
"quote",
|
|
425
|
+
"unicode-ident",
|
|
426
|
+
]
|
|
427
|
+
|
|
428
|
+
[[package]]
|
|
429
|
+
name = "thiserror"
|
|
430
|
+
version = "1.0.69"
|
|
431
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
432
|
+
checksum = "b6aaf5339b578ea85b50e080feb250a3e8ae8cfcdff9a461c9ec2904bc923f52"
|
|
433
|
+
dependencies = [
|
|
434
|
+
"thiserror-impl",
|
|
435
|
+
]
|
|
436
|
+
|
|
437
|
+
[[package]]
|
|
438
|
+
name = "thiserror-impl"
|
|
439
|
+
version = "1.0.69"
|
|
440
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
441
|
+
checksum = "4fee6c4efc90059e10f81e6d42c60a18f76588c3d74cb83a0b242a2b6c7504c1"
|
|
442
|
+
dependencies = [
|
|
443
|
+
"proc-macro2",
|
|
444
|
+
"quote",
|
|
445
|
+
"syn",
|
|
446
|
+
]
|
|
447
|
+
|
|
448
|
+
[[package]]
|
|
449
|
+
name = "tinyvec"
|
|
450
|
+
version = "1.11.0"
|
|
451
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
452
|
+
checksum = "3e61e67053d25a4e82c844e8424039d9745781b3fc4f32b8d55ed50f5f667ef3"
|
|
453
|
+
dependencies = [
|
|
454
|
+
"tinyvec_macros",
|
|
455
|
+
]
|
|
456
|
+
|
|
457
|
+
[[package]]
|
|
458
|
+
name = "tinyvec_macros"
|
|
459
|
+
version = "0.1.1"
|
|
460
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
461
|
+
checksum = "1f3ccbac311fea05f86f61904b462b55fb3df8837a366dfc601a0161d0532f20"
|
|
462
|
+
|
|
463
|
+
[[package]]
|
|
464
|
+
name = "unicode-ident"
|
|
465
|
+
version = "1.0.24"
|
|
466
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
467
|
+
checksum = "e6e4313cd5fcd3dad5cafa179702e2b244f760991f45397d14d4ebf38247da75"
|
|
468
|
+
|
|
469
|
+
[[package]]
|
|
470
|
+
name = "unicode-normalization"
|
|
471
|
+
version = "0.1.25"
|
|
472
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
473
|
+
checksum = "5fd4f6878c9cb28d874b009da9e8d183b5abc80117c40bbd187a1fde336be6e8"
|
|
474
|
+
dependencies = [
|
|
475
|
+
"tinyvec",
|
|
476
|
+
]
|
|
477
|
+
|
|
478
|
+
[[package]]
|
|
479
|
+
name = "windows-link"
|
|
480
|
+
version = "0.2.1"
|
|
481
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
482
|
+
checksum = "f0805222e57f7521d6a62e36fa9163bc891acd422f971defe97d64e70d0a4fe5"
|
|
483
|
+
|
|
484
|
+
[[package]]
|
|
485
|
+
name = "yada"
|
|
486
|
+
version = "0.5.1"
|
|
487
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
488
|
+
checksum = "aed111bd9e48a802518765906cbdadf0b45afb72b9c81ab049a3b86252adffdd"
|
|
489
|
+
|
|
490
|
+
[[package]]
|
|
491
|
+
name = "zmij"
|
|
492
|
+
version = "1.0.21"
|
|
493
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
494
|
+
checksum = "b8848ee67ecc8aedbaf3e4122217aff892639231befc6a1b58d29fff4c2cabaa"
|
data/Cargo.toml
ADDED
|
@@ -54,7 +54,9 @@ impl RbDictionary {
|
|
|
54
54
|
let lexicon = self.inner.lexicon();
|
|
55
55
|
|
|
56
56
|
for entry in lexicon.lookup(text.as_bytes(), 0) {
|
|
57
|
-
let info = lexicon
|
|
57
|
+
let info = lexicon
|
|
58
|
+
.get_word_info(entry.word_id)
|
|
59
|
+
.map_err(sudachi_error)?;
|
|
58
60
|
let surface_slice = text.get(..entry.end).ok_or_else(|| {
|
|
59
61
|
sudachi_error(format!(
|
|
60
62
|
"lookup returned invalid UTF-8 boundary: end={}",
|
data/ext/kabosu/src/lib.rs
CHANGED
|
@@ -51,7 +51,10 @@ fn init(ruby: &Ruby) -> Result<(), Error> {
|
|
|
51
51
|
let morph_class = module.define_class("Morpheme", ruby.class_object())?;
|
|
52
52
|
morph_class.define_method("surface", method!(RbMorpheme::surface, 0))?;
|
|
53
53
|
morph_class.define_method("part_of_speech", method!(RbMorpheme::part_of_speech, 0))?;
|
|
54
|
-
morph_class.define_method(
|
|
54
|
+
morph_class.define_method(
|
|
55
|
+
"part_of_speech_id",
|
|
56
|
+
method!(RbMorpheme::part_of_speech_id, 0),
|
|
57
|
+
)?;
|
|
55
58
|
morph_class.define_method("dictionary_form", method!(RbMorpheme::dictionary_form, 0))?;
|
|
56
59
|
morph_class.define_method("normalized_form", method!(RbMorpheme::normalized_form, 0))?;
|
|
57
60
|
morph_class.define_method("reading_form", method!(RbMorpheme::reading_form, 0))?;
|
|
@@ -67,7 +70,10 @@ fn init(ruby: &Ruby) -> Result<(), Error> {
|
|
|
67
70
|
morph_class.define_method("end", method!(RbMorpheme::end, 0))?;
|
|
68
71
|
morph_class.define_method("begin_c", method!(RbMorpheme::begin_c, 0))?;
|
|
69
72
|
morph_class.define_method("end_c", method!(RbMorpheme::end_c, 0))?;
|
|
70
|
-
morph_class.define_method(
|
|
73
|
+
morph_class.define_method(
|
|
74
|
+
"synonym_group_ids",
|
|
75
|
+
method!(RbMorpheme::synonym_group_ids, 0),
|
|
76
|
+
)?;
|
|
71
77
|
morph_class.define_method("a_unit_split", method!(RbMorpheme::a_unit_split, 0))?;
|
|
72
78
|
morph_class.define_method("b_unit_split", method!(RbMorpheme::b_unit_split, 0))?;
|
|
73
79
|
morph_class.define_method("word_structure", method!(RbMorpheme::word_structure, 0))?;
|
data/ext/kabosu/src/morpheme.rs
CHANGED
|
@@ -197,7 +197,11 @@ impl RbMorpheme {
|
|
|
197
197
|
let info = match info_result {
|
|
198
198
|
Ok(Ok(info)) => info,
|
|
199
199
|
Ok(Err(e)) => return Err(sudachi_error(e)),
|
|
200
|
-
Err(_) =>
|
|
200
|
+
Err(_) => {
|
|
201
|
+
return Err(sudachi_error(
|
|
202
|
+
"panic while reading child word info for split",
|
|
203
|
+
))
|
|
204
|
+
}
|
|
201
205
|
};
|
|
202
206
|
|
|
203
207
|
// head_word_length is in codepoints; clamp to remaining characters.
|
|
@@ -354,7 +358,11 @@ impl RbMorpheme {
|
|
|
354
358
|
ary.clear()?;
|
|
355
359
|
ary
|
|
356
360
|
}
|
|
357
|
-
None => ruby.ary_new_capa(if children.is_empty() {
|
|
361
|
+
None => ruby.ary_new_capa(if children.is_empty() {
|
|
362
|
+
1
|
|
363
|
+
} else {
|
|
364
|
+
children.len()
|
|
365
|
+
}),
|
|
358
366
|
};
|
|
359
367
|
|
|
360
368
|
if children.is_empty() {
|
data/ext/kabosu/src/parsing.rs
CHANGED
|
@@ -35,7 +35,9 @@ pub(crate) fn parse_subset(value: Option<Value>) -> Result<InfoSubset, Error> {
|
|
|
35
35
|
} else if let Ok(s) = String::try_convert(raw) {
|
|
36
36
|
s
|
|
37
37
|
} else {
|
|
38
|
-
return Err(arg_error(
|
|
38
|
+
return Err(arg_error(
|
|
39
|
+
"fields must contain only String or Symbol values",
|
|
40
|
+
));
|
|
39
41
|
};
|
|
40
42
|
|
|
41
43
|
let normalized = name.to_ascii_lowercase();
|
data/ext/kabosu/src/splitter.rs
CHANGED
|
@@ -34,7 +34,8 @@ impl SplitSentencesJob {
|
|
|
34
34
|
|
|
35
35
|
match &self.dict_path {
|
|
36
36
|
Some(path) => {
|
|
37
|
-
let cfg = Config::new(None, None, Some(path.clone().into()))
|
|
37
|
+
let cfg = Config::new(None, None, Some(path.clone().into()))
|
|
38
|
+
.map_err(|e| e.to_string())?;
|
|
38
39
|
let dict = JapaneseDictionary::from_cfg(&cfg).map_err(|e| e.to_string())?;
|
|
39
40
|
let base = match self.limit {
|
|
40
41
|
Some(lim) => SentenceSplitter::with_limit(lim),
|
data/ext/kabosu/src/tokenizer.rs
CHANGED
|
@@ -95,7 +95,9 @@ fn analyze_text(
|
|
|
95
95
|
tokenizer.do_tokenize().map_err(|e| e.to_string())?;
|
|
96
96
|
|
|
97
97
|
let mut mlist = MorphemeList::empty(pool.dict.clone());
|
|
98
|
-
mlist
|
|
98
|
+
mlist
|
|
99
|
+
.collect_results(tokenizer)
|
|
100
|
+
.map_err(|e| e.to_string())?;
|
|
99
101
|
|
|
100
102
|
let mut morphemes = Vec::with_capacity(mlist.len());
|
|
101
103
|
for i in 0..mlist.len() {
|
data/lib/kabosu/version.rb
CHANGED
metadata
CHANGED
|
@@ -1,14 +1,14 @@
|
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
|
2
2
|
name: kabosu
|
|
3
3
|
version: !ruby/object:Gem::Version
|
|
4
|
-
version: 0.6.10.dev.
|
|
4
|
+
version: 0.6.10.dev.20260429.b757f55
|
|
5
5
|
platform: ruby
|
|
6
6
|
authors:
|
|
7
7
|
- davafons
|
|
8
8
|
autorequire:
|
|
9
9
|
bindir: bin
|
|
10
10
|
cert_chain: []
|
|
11
|
-
date: 2026-
|
|
11
|
+
date: 2026-04-29 00:00:00.000000000 Z
|
|
12
12
|
dependencies:
|
|
13
13
|
- !ruby/object:Gem::Dependency
|
|
14
14
|
name: rb_sys
|
|
@@ -88,6 +88,8 @@ extensions:
|
|
|
88
88
|
- ext/kabosu/extconf.rb
|
|
89
89
|
extra_rdoc_files: []
|
|
90
90
|
files:
|
|
91
|
+
- Cargo.lock
|
|
92
|
+
- Cargo.toml
|
|
91
93
|
- LICENSE
|
|
92
94
|
- README.md
|
|
93
95
|
- ext/kabosu/Cargo.toml
|