tiktoken_ruby 0.0.8 → 0.0.10
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/.vscode/settings.json +5 -0
- data/Cargo.lock +117 -133
- data/Gemfile.lock +4 -4
- data/Rakefile +6 -0
- data/ext/tiktoken_ruby/Cargo.toml +2 -2
- data/ext/tiktoken_ruby/src/lib.rs +6 -0
- data/lib/tiktoken_ruby/version.rb +1 -1
- data/lib/tiktoken_ruby.rb +4 -1
- data/script/release +33 -0
- metadata +12 -11
- data/tiktoken_ruby.gemspec +0 -33
checksums.yaml
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
---
|
|
2
2
|
SHA256:
|
|
3
|
-
metadata.gz:
|
|
4
|
-
data.tar.gz:
|
|
3
|
+
metadata.gz: df7ca54ce72c32b1a0a0b278553f35365bf283af271f98436b4d330f6fb05735
|
|
4
|
+
data.tar.gz: 5a225b39a85ce71779bb7e58c6f6c47eb8a884b813377783d833b27af53cbb9e
|
|
5
5
|
SHA512:
|
|
6
|
-
metadata.gz:
|
|
7
|
-
data.tar.gz:
|
|
6
|
+
metadata.gz: d15ec5fee05da4104e7111e237e5bf5e80291d226c51afd90fa580ebed87571140db4b06a17e39a82892489a6a932cb02d594a9cbf047315c5094bb4f25bd99b
|
|
7
|
+
data.tar.gz: 444bd7e59d054395689a3f45beb3c288262e1e9e1903dc565821780ad9929de6aabbdbd9f543506b3cc662a54835892e9b34b9ece259d6f3f27d5642a8f1b75a
|
data/Cargo.lock
CHANGED
|
@@ -1,41 +1,41 @@
|
|
|
1
1
|
# This file is automatically @generated by Cargo.
|
|
2
2
|
# It is not intended for manual editing.
|
|
3
|
-
version =
|
|
3
|
+
version = 4
|
|
4
4
|
|
|
5
5
|
[[package]]
|
|
6
6
|
name = "aho-corasick"
|
|
7
|
-
version = "
|
|
7
|
+
version = "1.1.3"
|
|
8
8
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
9
|
-
checksum = "
|
|
9
|
+
checksum = "8e60d3430d3a69478ad0993f19238d2df97c507009a52b3c10addcd7f6bcb916"
|
|
10
10
|
dependencies = [
|
|
11
11
|
"memchr",
|
|
12
12
|
]
|
|
13
13
|
|
|
14
14
|
[[package]]
|
|
15
15
|
name = "anyhow"
|
|
16
|
-
version = "1.0.
|
|
16
|
+
version = "1.0.95"
|
|
17
17
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
18
|
-
checksum = "
|
|
18
|
+
checksum = "34ac096ce696dc2fcabef30516bb13c0a68a11d30131d3df6f04711467681b04"
|
|
19
19
|
|
|
20
20
|
[[package]]
|
|
21
21
|
name = "autocfg"
|
|
22
|
-
version = "1.
|
|
22
|
+
version = "1.4.0"
|
|
23
23
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
24
|
-
checksum = "
|
|
24
|
+
checksum = "ace50bade8e6234aa140d9a2f552bbee1db4d353f69b8217bc503490fc1a9f26"
|
|
25
25
|
|
|
26
26
|
[[package]]
|
|
27
27
|
name = "base64"
|
|
28
|
-
version = "0.21.
|
|
28
|
+
version = "0.21.7"
|
|
29
29
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
30
|
-
checksum = "
|
|
30
|
+
checksum = "9d297deb1925b89f2ccc13d7635fa0714f12c87adce1c75356b39ca9b7178567"
|
|
31
31
|
|
|
32
32
|
[[package]]
|
|
33
33
|
name = "bindgen"
|
|
34
|
-
version = "0.69.
|
|
34
|
+
version = "0.69.5"
|
|
35
35
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
36
|
-
checksum = "
|
|
36
|
+
checksum = "271383c67ccabffb7381723dea0672a673f292304fcb45c01cc648c7a8d58088"
|
|
37
37
|
dependencies = [
|
|
38
|
-
"bitflags
|
|
38
|
+
"bitflags",
|
|
39
39
|
"cexpr",
|
|
40
40
|
"clang-sys",
|
|
41
41
|
"itertools",
|
|
@@ -66,24 +66,17 @@ checksum = "349f9b6a179ed607305526ca489b34ad0a41aed5f7980fa90eb03160b69598fb"
|
|
|
66
66
|
|
|
67
67
|
[[package]]
|
|
68
68
|
name = "bitflags"
|
|
69
|
-
version = "
|
|
70
|
-
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
71
|
-
checksum = "bef38d45163c2f1dde094a7dfd33ccf595c92905c8f8f4fdc18d06fb1037718a"
|
|
72
|
-
|
|
73
|
-
[[package]]
|
|
74
|
-
name = "bitflags"
|
|
75
|
-
version = "2.4.0"
|
|
69
|
+
version = "2.6.0"
|
|
76
70
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
77
|
-
checksum = "
|
|
71
|
+
checksum = "b048fb63fd8b5923fc5aa7b340d8e156aec7ec02f0c78fa8a6ddc2613f6f71de"
|
|
78
72
|
|
|
79
73
|
[[package]]
|
|
80
74
|
name = "bstr"
|
|
81
|
-
version = "1.
|
|
75
|
+
version = "1.11.1"
|
|
82
76
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
83
|
-
checksum = "
|
|
77
|
+
checksum = "786a307d683a5bf92e6fd5fd69a7eb613751668d1d8d67d802846dfe367c62c8"
|
|
84
78
|
dependencies = [
|
|
85
79
|
"memchr",
|
|
86
|
-
"once_cell",
|
|
87
80
|
"regex-automata",
|
|
88
81
|
"serde",
|
|
89
82
|
]
|
|
@@ -105,9 +98,9 @@ checksum = "baf1de4339761588bc0619e3cbc0120ee582ebb74b53b4efbf79117bd2da40fd"
|
|
|
105
98
|
|
|
106
99
|
[[package]]
|
|
107
100
|
name = "clang-sys"
|
|
108
|
-
version = "1.
|
|
101
|
+
version = "1.8.1"
|
|
109
102
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
110
|
-
checksum = "
|
|
103
|
+
checksum = "0b023947811758c97c59bf9d1c188fd619ad4718dcaa767947df1cadb14f39f4"
|
|
111
104
|
dependencies = [
|
|
112
105
|
"glob",
|
|
113
106
|
"libc",
|
|
@@ -116,15 +109,15 @@ dependencies = [
|
|
|
116
109
|
|
|
117
110
|
[[package]]
|
|
118
111
|
name = "either"
|
|
119
|
-
version = "1.
|
|
112
|
+
version = "1.13.0"
|
|
120
113
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
121
|
-
checksum = "
|
|
114
|
+
checksum = "60b1af1c220855b6ceac025d3f6ecdd2b7c4894bfe9cd9bda4fbb4bc7c0d4cf0"
|
|
122
115
|
|
|
123
116
|
[[package]]
|
|
124
117
|
name = "fancy-regex"
|
|
125
|
-
version = "0.
|
|
118
|
+
version = "0.12.0"
|
|
126
119
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
127
|
-
checksum = "
|
|
120
|
+
checksum = "7493d4c459da9f84325ad297371a6b2b8a162800873a22e3b6b6512e61d18c05"
|
|
128
121
|
dependencies = [
|
|
129
122
|
"bit-set",
|
|
130
123
|
"regex",
|
|
@@ -132,9 +125,9 @@ dependencies = [
|
|
|
132
125
|
|
|
133
126
|
[[package]]
|
|
134
127
|
name = "glob"
|
|
135
|
-
version = "0.3.
|
|
128
|
+
version = "0.3.2"
|
|
136
129
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
137
|
-
checksum = "
|
|
130
|
+
checksum = "a8d1add55171497b4705a648c6b583acafb01d58050a51727785f0b2c8e0a2b2"
|
|
138
131
|
|
|
139
132
|
[[package]]
|
|
140
133
|
name = "itertools"
|
|
@@ -147,9 +140,9 @@ dependencies = [
|
|
|
147
140
|
|
|
148
141
|
[[package]]
|
|
149
142
|
name = "lazy_static"
|
|
150
|
-
version = "1.
|
|
143
|
+
version = "1.5.0"
|
|
151
144
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
152
|
-
checksum = "
|
|
145
|
+
checksum = "bbd2bcb4c963f2ddae06a2efc7e9f3591312473c50c6685e1f298068316e66fe"
|
|
153
146
|
|
|
154
147
|
[[package]]
|
|
155
148
|
name = "lazycell"
|
|
@@ -159,25 +152,25 @@ checksum = "830d08ce1d1d941e6b30645f1a0eb5643013d835ce3779a5fc208261dbe10f55"
|
|
|
159
152
|
|
|
160
153
|
[[package]]
|
|
161
154
|
name = "libc"
|
|
162
|
-
version = "0.2.
|
|
155
|
+
version = "0.2.169"
|
|
163
156
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
164
|
-
checksum = "
|
|
157
|
+
checksum = "b5aba8db14291edd000dfcc4d620c7ebfb122c613afb886ca8803fa4e128a20a"
|
|
165
158
|
|
|
166
159
|
[[package]]
|
|
167
160
|
name = "libloading"
|
|
168
|
-
version = "0.
|
|
161
|
+
version = "0.8.6"
|
|
169
162
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
170
|
-
checksum = "
|
|
163
|
+
checksum = "fc2f4eb4bc735547cfed7c0a4922cbd04a4655978c09b54f1f7b228750664c34"
|
|
171
164
|
dependencies = [
|
|
172
165
|
"cfg-if",
|
|
173
|
-
"
|
|
166
|
+
"windows-targets",
|
|
174
167
|
]
|
|
175
168
|
|
|
176
169
|
[[package]]
|
|
177
170
|
name = "lock_api"
|
|
178
|
-
version = "0.4.
|
|
171
|
+
version = "0.4.12"
|
|
179
172
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
180
|
-
checksum = "
|
|
173
|
+
checksum = "07af8b9cdd281b7915f413fa73f29ebd5d55d0d3f0155584dade1ff18cea1b17"
|
|
181
174
|
dependencies = [
|
|
182
175
|
"autocfg",
|
|
183
176
|
"scopeguard",
|
|
@@ -185,9 +178,9 @@ dependencies = [
|
|
|
185
178
|
|
|
186
179
|
[[package]]
|
|
187
180
|
name = "magnus"
|
|
188
|
-
version = "0.6.
|
|
181
|
+
version = "0.6.4"
|
|
189
182
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
190
|
-
checksum = "
|
|
183
|
+
checksum = "b1597ef40aa8c36be098249e82c9a20cf7199278ac1c1a1a995eeead6a184479"
|
|
191
184
|
dependencies = [
|
|
192
185
|
"magnus-macros",
|
|
193
186
|
"rb-sys",
|
|
@@ -208,9 +201,9 @@ dependencies = [
|
|
|
208
201
|
|
|
209
202
|
[[package]]
|
|
210
203
|
name = "memchr"
|
|
211
|
-
version = "2.
|
|
204
|
+
version = "2.7.4"
|
|
212
205
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
213
|
-
checksum = "
|
|
206
|
+
checksum = "78ca9ab1a0babb1e7d5695e3530886289c18cf2f87ec19a575a0abdce112e3a3"
|
|
214
207
|
|
|
215
208
|
[[package]]
|
|
216
209
|
name = "minimal-lexical"
|
|
@@ -228,17 +221,11 @@ dependencies = [
|
|
|
228
221
|
"minimal-lexical",
|
|
229
222
|
]
|
|
230
223
|
|
|
231
|
-
[[package]]
|
|
232
|
-
name = "once_cell"
|
|
233
|
-
version = "1.17.1"
|
|
234
|
-
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
235
|
-
checksum = "b7e5500299e16ebb147ae15a00a942af264cf3688f47923b8fc2cd5858f23ad3"
|
|
236
|
-
|
|
237
224
|
[[package]]
|
|
238
225
|
name = "parking_lot"
|
|
239
|
-
version = "0.12.
|
|
226
|
+
version = "0.12.3"
|
|
240
227
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
241
|
-
checksum = "
|
|
228
|
+
checksum = "f1bf18183cf54e8d6059647fc3063646a1801cf30896933ec2311622cc4b9a27"
|
|
242
229
|
dependencies = [
|
|
243
230
|
"lock_api",
|
|
244
231
|
"parking_lot_core",
|
|
@@ -246,49 +233,49 @@ dependencies = [
|
|
|
246
233
|
|
|
247
234
|
[[package]]
|
|
248
235
|
name = "parking_lot_core"
|
|
249
|
-
version = "0.9.
|
|
236
|
+
version = "0.9.10"
|
|
250
237
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
251
|
-
checksum = "
|
|
238
|
+
checksum = "1e401f977ab385c9e4e3ab30627d6f26d00e2c73eef317493c4ec6d468726cf8"
|
|
252
239
|
dependencies = [
|
|
253
240
|
"cfg-if",
|
|
254
241
|
"libc",
|
|
255
242
|
"redox_syscall",
|
|
256
243
|
"smallvec",
|
|
257
|
-
"windows-
|
|
244
|
+
"windows-targets",
|
|
258
245
|
]
|
|
259
246
|
|
|
260
247
|
[[package]]
|
|
261
248
|
name = "proc-macro2"
|
|
262
|
-
version = "1.0.
|
|
249
|
+
version = "1.0.92"
|
|
263
250
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
264
|
-
checksum = "
|
|
251
|
+
checksum = "37d3544b3f2748c54e147655edb5025752e2303145b5aefb3c3ea2c78b973bb0"
|
|
265
252
|
dependencies = [
|
|
266
253
|
"unicode-ident",
|
|
267
254
|
]
|
|
268
255
|
|
|
269
256
|
[[package]]
|
|
270
257
|
name = "quote"
|
|
271
|
-
version = "1.0.
|
|
258
|
+
version = "1.0.38"
|
|
272
259
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
273
|
-
checksum = "
|
|
260
|
+
checksum = "0e4dccaaaf89514f546c693ddc140f729f958c247918a13380cccc6078391acc"
|
|
274
261
|
dependencies = [
|
|
275
262
|
"proc-macro2",
|
|
276
263
|
]
|
|
277
264
|
|
|
278
265
|
[[package]]
|
|
279
266
|
name = "rb-sys"
|
|
280
|
-
version = "0.9.
|
|
267
|
+
version = "0.9.105"
|
|
281
268
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
282
|
-
checksum = "
|
|
269
|
+
checksum = "4b3a1f3ce8e7c36d777d52fe7a99039fe4fea7c8ec355a4c4f3a17f92a14029f"
|
|
283
270
|
dependencies = [
|
|
284
271
|
"rb-sys-build",
|
|
285
272
|
]
|
|
286
273
|
|
|
287
274
|
[[package]]
|
|
288
275
|
name = "rb-sys-build"
|
|
289
|
-
version = "0.9.
|
|
276
|
+
version = "0.9.105"
|
|
290
277
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
291
|
-
checksum = "
|
|
278
|
+
checksum = "3e6b246c29c0809e1cbe60a1ba9e093da72a4676d02adc68469297d1e589bbf0"
|
|
292
279
|
dependencies = [
|
|
293
280
|
"bindgen",
|
|
294
281
|
"lazy_static",
|
|
@@ -307,35 +294,41 @@ checksum = "a35802679f07360454b418a5d1735c89716bde01d35b1560fc953c1415a0b3bb"
|
|
|
307
294
|
|
|
308
295
|
[[package]]
|
|
309
296
|
name = "redox_syscall"
|
|
310
|
-
version = "0.
|
|
297
|
+
version = "0.5.8"
|
|
311
298
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
312
|
-
checksum = "
|
|
299
|
+
checksum = "03a862b389f93e68874fbf580b9de08dd02facb9a788ebadaf4a3fd33cf58834"
|
|
313
300
|
dependencies = [
|
|
314
|
-
"bitflags
|
|
301
|
+
"bitflags",
|
|
315
302
|
]
|
|
316
303
|
|
|
317
304
|
[[package]]
|
|
318
305
|
name = "regex"
|
|
319
|
-
version = "1.
|
|
306
|
+
version = "1.11.1"
|
|
320
307
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
321
|
-
checksum = "
|
|
308
|
+
checksum = "b544ef1b4eac5dc2db33ea63606ae9ffcfac26c1416a2806ae0bf5f56b201191"
|
|
322
309
|
dependencies = [
|
|
323
310
|
"aho-corasick",
|
|
324
311
|
"memchr",
|
|
312
|
+
"regex-automata",
|
|
325
313
|
"regex-syntax",
|
|
326
314
|
]
|
|
327
315
|
|
|
328
316
|
[[package]]
|
|
329
317
|
name = "regex-automata"
|
|
330
|
-
version = "0.
|
|
318
|
+
version = "0.4.9"
|
|
331
319
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
332
|
-
checksum = "
|
|
320
|
+
checksum = "809e8dc61f6de73b46c85f4c96486310fe304c434cfa43669d7b40f711150908"
|
|
321
|
+
dependencies = [
|
|
322
|
+
"aho-corasick",
|
|
323
|
+
"memchr",
|
|
324
|
+
"regex-syntax",
|
|
325
|
+
]
|
|
333
326
|
|
|
334
327
|
[[package]]
|
|
335
328
|
name = "regex-syntax"
|
|
336
|
-
version = "0.
|
|
329
|
+
version = "0.8.5"
|
|
337
330
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
338
|
-
checksum = "
|
|
331
|
+
checksum = "2b15c43186be67a4fd63bee50d0303afffcef381492ebe2c5d87f324e1b8815c"
|
|
339
332
|
|
|
340
333
|
[[package]]
|
|
341
334
|
name = "rustc-hash"
|
|
@@ -345,9 +338,9 @@ checksum = "08d43f7aa6b08d49f382cde6a7982047c3426db949b1424bc4b7ec9ae12c6ce2"
|
|
|
345
338
|
|
|
346
339
|
[[package]]
|
|
347
340
|
name = "scopeguard"
|
|
348
|
-
version = "1.
|
|
341
|
+
version = "1.2.0"
|
|
349
342
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
350
|
-
checksum = "
|
|
343
|
+
checksum = "94143f37725109f92c262ed2cf5e59bce7498c01bcc1502d7b9afe439a4e9f49"
|
|
351
344
|
|
|
352
345
|
[[package]]
|
|
353
346
|
name = "seq-macro"
|
|
@@ -357,9 +350,23 @@ checksum = "a3f0bf26fd526d2a95683cd0f87bf103b8539e2ca1ef48ce002d67aad59aa0b4"
|
|
|
357
350
|
|
|
358
351
|
[[package]]
|
|
359
352
|
name = "serde"
|
|
360
|
-
version = "1.0.
|
|
353
|
+
version = "1.0.217"
|
|
354
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
355
|
+
checksum = "02fc4265df13d6fa1d00ecff087228cc0a2b5f3c0e87e258d8b94a156e984c70"
|
|
356
|
+
dependencies = [
|
|
357
|
+
"serde_derive",
|
|
358
|
+
]
|
|
359
|
+
|
|
360
|
+
[[package]]
|
|
361
|
+
name = "serde_derive"
|
|
362
|
+
version = "1.0.217"
|
|
361
363
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
362
|
-
checksum = "
|
|
364
|
+
checksum = "5a9bf7cf98d04a2b28aead066b7496853d4779c9cc183c440dbac457641e19a0"
|
|
365
|
+
dependencies = [
|
|
366
|
+
"proc-macro2",
|
|
367
|
+
"quote",
|
|
368
|
+
"syn",
|
|
369
|
+
]
|
|
363
370
|
|
|
364
371
|
[[package]]
|
|
365
372
|
name = "shell-words"
|
|
@@ -369,21 +376,21 @@ checksum = "24188a676b6ae68c3b2cb3a01be17fbf7240ce009799bb56d5b1409051e78fde"
|
|
|
369
376
|
|
|
370
377
|
[[package]]
|
|
371
378
|
name = "shlex"
|
|
372
|
-
version = "1.
|
|
379
|
+
version = "1.3.0"
|
|
373
380
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
374
|
-
checksum = "
|
|
381
|
+
checksum = "0fda2ff0d084019ba4d7c6f371c95d8fd75ce3524c3cb8fb653a3023f6323e64"
|
|
375
382
|
|
|
376
383
|
[[package]]
|
|
377
384
|
name = "smallvec"
|
|
378
|
-
version = "1.
|
|
385
|
+
version = "1.13.2"
|
|
379
386
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
380
|
-
checksum = "
|
|
387
|
+
checksum = "3c5e1a9a646d36c3599cd173a41282daf47c44583ad367b8e6837255952e5c67"
|
|
381
388
|
|
|
382
389
|
[[package]]
|
|
383
390
|
name = "syn"
|
|
384
|
-
version = "2.0.
|
|
391
|
+
version = "2.0.93"
|
|
385
392
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
386
|
-
checksum = "
|
|
393
|
+
checksum = "9c786062daee0d6db1132800e623df74274a0a87322d8e183338e01b3d98d058"
|
|
387
394
|
dependencies = [
|
|
388
395
|
"proc-macro2",
|
|
389
396
|
"quote",
|
|
@@ -392,8 +399,9 @@ dependencies = [
|
|
|
392
399
|
|
|
393
400
|
[[package]]
|
|
394
401
|
name = "tiktoken-rs"
|
|
395
|
-
version = "0.
|
|
396
|
-
source = "
|
|
402
|
+
version = "0.5.9"
|
|
403
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
404
|
+
checksum = "c314e7ce51440f9e8f5a497394682a57b7c323d0f4d0a6b1b13c429056e0e234"
|
|
397
405
|
dependencies = [
|
|
398
406
|
"anyhow",
|
|
399
407
|
"base64",
|
|
@@ -415,50 +423,20 @@ dependencies = [
|
|
|
415
423
|
|
|
416
424
|
[[package]]
|
|
417
425
|
name = "unicode-ident"
|
|
418
|
-
version = "1.0.
|
|
426
|
+
version = "1.0.14"
|
|
419
427
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
420
|
-
checksum = "
|
|
421
|
-
|
|
422
|
-
[[package]]
|
|
423
|
-
name = "winapi"
|
|
424
|
-
version = "0.3.9"
|
|
425
|
-
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
426
|
-
checksum = "5c839a674fcd7a98952e593242ea400abe93992746761e38641405d28b00f419"
|
|
427
|
-
dependencies = [
|
|
428
|
-
"winapi-i686-pc-windows-gnu",
|
|
429
|
-
"winapi-x86_64-pc-windows-gnu",
|
|
430
|
-
]
|
|
431
|
-
|
|
432
|
-
[[package]]
|
|
433
|
-
name = "winapi-i686-pc-windows-gnu"
|
|
434
|
-
version = "0.4.0"
|
|
435
|
-
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
436
|
-
checksum = "ac3b87c63620426dd9b991e5ce0329eff545bccbbb34f3be09ff6fb6ab51b7b6"
|
|
437
|
-
|
|
438
|
-
[[package]]
|
|
439
|
-
name = "winapi-x86_64-pc-windows-gnu"
|
|
440
|
-
version = "0.4.0"
|
|
441
|
-
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
442
|
-
checksum = "712e227841d057c1ee1cd2fb22fa7e5a5461ae8e48fa2ca79ec42cfc1931183f"
|
|
443
|
-
|
|
444
|
-
[[package]]
|
|
445
|
-
name = "windows-sys"
|
|
446
|
-
version = "0.45.0"
|
|
447
|
-
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
448
|
-
checksum = "75283be5efb2831d37ea142365f009c02ec203cd29a3ebecbc093d52315b66d0"
|
|
449
|
-
dependencies = [
|
|
450
|
-
"windows-targets",
|
|
451
|
-
]
|
|
428
|
+
checksum = "adb9e6ca4f869e1180728b7950e35922a7fc6397f7b641499e8f3ef06e50dc83"
|
|
452
429
|
|
|
453
430
|
[[package]]
|
|
454
431
|
name = "windows-targets"
|
|
455
|
-
version = "0.
|
|
432
|
+
version = "0.52.6"
|
|
456
433
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
457
|
-
checksum = "
|
|
434
|
+
checksum = "9b724f72796e036ab90c1021d4780d4d3d648aca59e491e6b98e725b84e99973"
|
|
458
435
|
dependencies = [
|
|
459
436
|
"windows_aarch64_gnullvm",
|
|
460
437
|
"windows_aarch64_msvc",
|
|
461
438
|
"windows_i686_gnu",
|
|
439
|
+
"windows_i686_gnullvm",
|
|
462
440
|
"windows_i686_msvc",
|
|
463
441
|
"windows_x86_64_gnu",
|
|
464
442
|
"windows_x86_64_gnullvm",
|
|
@@ -467,42 +445,48 @@ dependencies = [
|
|
|
467
445
|
|
|
468
446
|
[[package]]
|
|
469
447
|
name = "windows_aarch64_gnullvm"
|
|
470
|
-
version = "0.
|
|
448
|
+
version = "0.52.6"
|
|
471
449
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
472
|
-
checksum = "
|
|
450
|
+
checksum = "32a4622180e7a0ec044bb555404c800bc9fd9ec262ec147edd5989ccd0c02cd3"
|
|
473
451
|
|
|
474
452
|
[[package]]
|
|
475
453
|
name = "windows_aarch64_msvc"
|
|
476
|
-
version = "0.
|
|
454
|
+
version = "0.52.6"
|
|
477
455
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
478
|
-
checksum = "
|
|
456
|
+
checksum = "09ec2a7bb152e2252b53fa7803150007879548bc709c039df7627cabbd05d469"
|
|
479
457
|
|
|
480
458
|
[[package]]
|
|
481
459
|
name = "windows_i686_gnu"
|
|
482
|
-
version = "0.
|
|
460
|
+
version = "0.52.6"
|
|
461
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
462
|
+
checksum = "8e9b5ad5ab802e97eb8e295ac6720e509ee4c243f69d781394014ebfe8bbfa0b"
|
|
463
|
+
|
|
464
|
+
[[package]]
|
|
465
|
+
name = "windows_i686_gnullvm"
|
|
466
|
+
version = "0.52.6"
|
|
483
467
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
484
|
-
checksum = "
|
|
468
|
+
checksum = "0eee52d38c090b3caa76c563b86c3a4bd71ef1a819287c19d586d7334ae8ed66"
|
|
485
469
|
|
|
486
470
|
[[package]]
|
|
487
471
|
name = "windows_i686_msvc"
|
|
488
|
-
version = "0.
|
|
472
|
+
version = "0.52.6"
|
|
489
473
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
490
|
-
checksum = "
|
|
474
|
+
checksum = "240948bc05c5e7c6dabba28bf89d89ffce3e303022809e73deaefe4f6ec56c66"
|
|
491
475
|
|
|
492
476
|
[[package]]
|
|
493
477
|
name = "windows_x86_64_gnu"
|
|
494
|
-
version = "0.
|
|
478
|
+
version = "0.52.6"
|
|
495
479
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
496
|
-
checksum = "
|
|
480
|
+
checksum = "147a5c80aabfbf0c7d901cb5895d1de30ef2907eb21fbbab29ca94c5b08b1a78"
|
|
497
481
|
|
|
498
482
|
[[package]]
|
|
499
483
|
name = "windows_x86_64_gnullvm"
|
|
500
|
-
version = "0.
|
|
484
|
+
version = "0.52.6"
|
|
501
485
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
502
|
-
checksum = "
|
|
486
|
+
checksum = "24d5b23dc417412679681396f2b49f3de8c1473deb516bd34410872eff51ed0d"
|
|
503
487
|
|
|
504
488
|
[[package]]
|
|
505
489
|
name = "windows_x86_64_msvc"
|
|
506
|
-
version = "0.
|
|
490
|
+
version = "0.52.6"
|
|
507
491
|
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
508
|
-
checksum = "
|
|
492
|
+
checksum = "589f6da84c646204747d1270a2a5661ea66ed1cced2631d546fdfb155959f9ec"
|
data/Gemfile.lock
CHANGED
|
@@ -1,8 +1,8 @@
|
|
|
1
1
|
PATH
|
|
2
2
|
remote: .
|
|
3
3
|
specs:
|
|
4
|
-
tiktoken_ruby (0.0.
|
|
5
|
-
rb_sys (>= 0.9.
|
|
4
|
+
tiktoken_ruby (0.0.10)
|
|
5
|
+
rb_sys (>= 0.9.87)
|
|
6
6
|
|
|
7
7
|
GEM
|
|
8
8
|
remote: https://rubygems.org/
|
|
@@ -22,7 +22,7 @@ GEM
|
|
|
22
22
|
rake (13.1.0)
|
|
23
23
|
rake-compiler (1.2.5)
|
|
24
24
|
rake
|
|
25
|
-
rb_sys (0.9.
|
|
25
|
+
rb_sys (0.9.105)
|
|
26
26
|
regexp_parser (2.9.0)
|
|
27
27
|
rexml (3.2.6)
|
|
28
28
|
rspec (3.12.0)
|
|
@@ -89,4 +89,4 @@ DEPENDENCIES
|
|
|
89
89
|
yard-doctest
|
|
90
90
|
|
|
91
91
|
BUNDLED WITH
|
|
92
|
-
2.
|
|
92
|
+
2.5.18
|
data/Rakefile
CHANGED
|
@@ -21,3 +21,9 @@ end
|
|
|
21
21
|
task build: :compile
|
|
22
22
|
|
|
23
23
|
task default: %i[compile spec standard]
|
|
24
|
+
|
|
25
|
+
# Packaging default (non-precompiled) gem
|
|
26
|
+
require "rubygems/package_task"
|
|
27
|
+
gem_path = Gem::PackageTask.new(GEMSPEC).define
|
|
28
|
+
desc "Package the Ruby gem"
|
|
29
|
+
task "package" => [gem_path]
|
|
@@ -11,5 +11,5 @@ crate-type = ["cdylib"]
|
|
|
11
11
|
|
|
12
12
|
[dependencies]
|
|
13
13
|
magnus = { version = "0.6.1" }
|
|
14
|
-
rb-sys = { version = "
|
|
15
|
-
tiktoken-rs = {
|
|
14
|
+
rb-sys = { version = "0.9.87", features = ["stable-api-compiled-fallback"] }
|
|
15
|
+
tiktoken-rs = { version = "0.5.9" }
|
|
@@ -20,6 +20,11 @@ fn cl100k_base() -> CoreBPEWrapper {
|
|
|
20
20
|
CoreBPEWrapper::new(core_bpe)
|
|
21
21
|
}
|
|
22
22
|
|
|
23
|
+
fn o200k_base() -> CoreBPEWrapper {
|
|
24
|
+
let core_bpe = tiktoken_rs::o200k_base().unwrap();
|
|
25
|
+
CoreBPEWrapper::new(core_bpe)
|
|
26
|
+
}
|
|
27
|
+
|
|
23
28
|
fn module() -> Result<RModule, magnus::Error> {
|
|
24
29
|
define_module("Tiktoken")
|
|
25
30
|
}
|
|
@@ -37,6 +42,7 @@ fn init() -> Result<(), Error> {
|
|
|
37
42
|
factory_module.define_singleton_method("p50k_base", function!(p50k_base, 0))?;
|
|
38
43
|
factory_module.define_singleton_method("p50k_edit", function!(p50k_edit, 0))?;
|
|
39
44
|
factory_module.define_singleton_method("cl100k_base", function!(cl100k_base, 0))?;
|
|
45
|
+
factory_module.define_singleton_method("o200k_base", function!(o200k_base, 0))?;
|
|
40
46
|
|
|
41
47
|
let ext_module = module.define_module("Ext")?;
|
|
42
48
|
let bpe_class = ext_module.define_class("CoreBPE", class::object())?;
|
data/lib/tiktoken_ruby.rb
CHANGED
|
@@ -64,13 +64,15 @@ module Tiktoken
|
|
|
64
64
|
:r50k_base,
|
|
65
65
|
:p50k_base,
|
|
66
66
|
:p50k_edit,
|
|
67
|
-
:cl100k_base
|
|
67
|
+
:cl100k_base,
|
|
68
|
+
:o200k_base
|
|
68
69
|
]
|
|
69
70
|
|
|
70
71
|
# taken from the python library here https://github.com/openai/tiktoken/blob/main/tiktoken/model.py
|
|
71
72
|
# that is also MIT licensed but by OpenAI
|
|
72
73
|
MODEL_TO_ENCODING_NAME = {
|
|
73
74
|
# chat
|
|
75
|
+
"gpt-4o": "o200k_base",
|
|
74
76
|
"gpt-4": "cl100k_base",
|
|
75
77
|
"gpt-3.5-turbo": "cl100k_base",
|
|
76
78
|
"gpt-35-turbo": "cl100k_base", # Azure deployment name
|
|
@@ -120,6 +122,7 @@ module Tiktoken
|
|
|
120
122
|
|
|
121
123
|
MODEL_PREFIX_TO_ENCODING = {
|
|
122
124
|
# chat
|
|
125
|
+
"gpt-4o-": "o200k_base", # e.g., gpt-4o-2024-05-13, etc.
|
|
123
126
|
"gpt-4-": "cl100k_base", # e.g., gpt-4-0314, etc., plus gpt-4-32k
|
|
124
127
|
"gpt-3.5-turbo-": "cl100k_base", # e.g, gpt-3.5-turbo-0301, -0401, etc.
|
|
125
128
|
"gpt-35-turbo-": "cl100k_base", # Azure deployment name
|
data/script/release
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
1
|
+
#!/usr/bin/env bash
|
|
2
|
+
|
|
3
|
+
set -e
|
|
4
|
+
|
|
5
|
+
if [ -z "${TIKTOKEN_PUBLISH_KEY}" ]; then
|
|
6
|
+
echo "Error: TIKTOKEN_PUBLISH_KEY is not set. This is the RubyGems API key to push the gem."
|
|
7
|
+
exit 1
|
|
8
|
+
fi
|
|
9
|
+
|
|
10
|
+
version=$(grep VERSION lib/tiktoken_ruby/version.rb | head -n 1 | cut -d'"' -f2)
|
|
11
|
+
echo "Building tiktoken_ruby v$version"
|
|
12
|
+
|
|
13
|
+
targets=(
|
|
14
|
+
"arm64-darwin"
|
|
15
|
+
"x86_64-darwin"
|
|
16
|
+
"aarch64-linux"
|
|
17
|
+
"x86_64-linux"
|
|
18
|
+
"x86_64-linux-musl"
|
|
19
|
+
"arm-linux"
|
|
20
|
+
"x64-mingw-ucrt"
|
|
21
|
+
)
|
|
22
|
+
|
|
23
|
+
# for target in "${targets[@]}"; do
|
|
24
|
+
# bundle exec rb-sys-dock -p "$target" --ruby-versions 3.2 --build
|
|
25
|
+
# done
|
|
26
|
+
|
|
27
|
+
for gem in pkg/tiktoken_ruby-"$version"-*.gem ; do
|
|
28
|
+
GEM_HOST_API_KEY="${TIKTOKEN_PUBLISH_KEY}" gem push "$gem" --host https://rubygems.org
|
|
29
|
+
done
|
|
30
|
+
|
|
31
|
+
# last but not least, the uncompiled gem
|
|
32
|
+
bundle exec rake package
|
|
33
|
+
GEM_HOST_API_KEY="${TIKTOKEN_PUBLISH_KEY}" gem push "pkg/tiktoken_ruby-$version.gem" --host https://rubygems.org
|
metadata
CHANGED
|
@@ -1,14 +1,14 @@
|
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
|
2
2
|
name: tiktoken_ruby
|
|
3
3
|
version: !ruby/object:Gem::Version
|
|
4
|
-
version: 0.0.
|
|
4
|
+
version: 0.0.10
|
|
5
5
|
platform: ruby
|
|
6
6
|
authors:
|
|
7
7
|
- IAPark
|
|
8
|
-
autorequire:
|
|
8
|
+
autorequire:
|
|
9
9
|
bindir: exe
|
|
10
10
|
cert_chain: []
|
|
11
|
-
date: 2024-
|
|
11
|
+
date: 2024-12-29 00:00:00.000000000 Z
|
|
12
12
|
dependencies:
|
|
13
13
|
- !ruby/object:Gem::Dependency
|
|
14
14
|
name: rb_sys
|
|
@@ -16,14 +16,14 @@ dependencies:
|
|
|
16
16
|
requirements:
|
|
17
17
|
- - ">="
|
|
18
18
|
- !ruby/object:Gem::Version
|
|
19
|
-
version: 0.9.
|
|
19
|
+
version: 0.9.87
|
|
20
20
|
type: :runtime
|
|
21
21
|
prerelease: false
|
|
22
22
|
version_requirements: !ruby/object:Gem::Requirement
|
|
23
23
|
requirements:
|
|
24
24
|
- - ">="
|
|
25
25
|
- !ruby/object:Gem::Version
|
|
26
|
-
version: 0.9.
|
|
26
|
+
version: 0.9.87
|
|
27
27
|
description: An unofficial Ruby wrapper for Tiktoken, a BPE tokenizer written by and
|
|
28
28
|
used by OpenAI. It can be used to count the number of tokens in text before sending
|
|
29
29
|
it to OpenAI APIs.
|
|
@@ -36,6 +36,7 @@ extra_rdoc_files: []
|
|
|
36
36
|
files:
|
|
37
37
|
- ".rspec"
|
|
38
38
|
- ".standard.yml"
|
|
39
|
+
- ".vscode/settings.json"
|
|
39
40
|
- Cargo.lock
|
|
40
41
|
- Cargo.toml
|
|
41
42
|
- Gemfile
|
|
@@ -51,8 +52,8 @@ files:
|
|
|
51
52
|
- lib/tiktoken_ruby.rb
|
|
52
53
|
- lib/tiktoken_ruby/encoding.rb
|
|
53
54
|
- lib/tiktoken_ruby/version.rb
|
|
55
|
+
- script/release
|
|
54
56
|
- sig/tiktoken_ruby.rbs
|
|
55
|
-
- tiktoken_ruby.gemspec
|
|
56
57
|
homepage: https://github.com/IAPark/tiktoken_ruby
|
|
57
58
|
licenses:
|
|
58
59
|
- MIT
|
|
@@ -60,7 +61,7 @@ metadata:
|
|
|
60
61
|
homepage_uri: https://github.com/IAPark/tiktoken_ruby
|
|
61
62
|
source_code_uri: https://github.com/IAPark/tiktoken_ruby
|
|
62
63
|
documentation_uri: https://rubydoc.info/github/IAPark/tiktoken_ruby/main
|
|
63
|
-
post_install_message:
|
|
64
|
+
post_install_message:
|
|
64
65
|
rdoc_options: []
|
|
65
66
|
require_paths:
|
|
66
67
|
- lib
|
|
@@ -68,15 +69,15 @@ required_ruby_version: !ruby/object:Gem::Requirement
|
|
|
68
69
|
requirements:
|
|
69
70
|
- - ">="
|
|
70
71
|
- !ruby/object:Gem::Version
|
|
71
|
-
version:
|
|
72
|
+
version: 3.1.0
|
|
72
73
|
required_rubygems_version: !ruby/object:Gem::Requirement
|
|
73
74
|
requirements:
|
|
74
75
|
- - ">="
|
|
75
76
|
- !ruby/object:Gem::Version
|
|
76
|
-
version: 3.
|
|
77
|
+
version: 3.4.0
|
|
77
78
|
requirements: []
|
|
78
|
-
rubygems_version: 3.
|
|
79
|
-
signing_key:
|
|
79
|
+
rubygems_version: 3.5.23
|
|
80
|
+
signing_key:
|
|
80
81
|
specification_version: 4
|
|
81
82
|
summary: Ruby wrapper for Tiktoken
|
|
82
83
|
test_files: []
|
data/tiktoken_ruby.gemspec
DELETED
|
@@ -1,33 +0,0 @@
|
|
|
1
|
-
# frozen_string_literal: true
|
|
2
|
-
|
|
3
|
-
require_relative "lib/tiktoken_ruby/version"
|
|
4
|
-
|
|
5
|
-
Gem::Specification.new do |spec|
|
|
6
|
-
spec.name = "tiktoken_ruby"
|
|
7
|
-
spec.version = Tiktoken::VERSION
|
|
8
|
-
spec.authors = ["IAPark"]
|
|
9
|
-
spec.email = ["isaac.a.park@gmail.com"]
|
|
10
|
-
spec.summary = "Ruby wrapper for Tiktoken"
|
|
11
|
-
spec.description = "An unofficial Ruby wrapper for Tiktoken, " \
|
|
12
|
-
"a BPE tokenizer written by and used by OpenAI. It can be used to " \
|
|
13
|
-
"count the number of tokens in text before sending it to OpenAI APIs."
|
|
14
|
-
spec.homepage = "https://github.com/IAPark/tiktoken_ruby"
|
|
15
|
-
spec.license = "MIT"
|
|
16
|
-
spec.required_ruby_version = ">= 2.7.0"
|
|
17
|
-
spec.required_rubygems_version = ">= 3.1.0"
|
|
18
|
-
spec.platform = Gem::Platform::RUBY
|
|
19
|
-
|
|
20
|
-
spec.metadata["homepage_uri"] = spec.homepage
|
|
21
|
-
spec.metadata["source_code_uri"] = "https://github.com/IAPark/tiktoken_ruby"
|
|
22
|
-
spec.metadata["documentation_uri"] = "https://rubydoc.info/github/IAPark/tiktoken_ruby/main"
|
|
23
|
-
spec.files = Dir.chdir(__dir__) do
|
|
24
|
-
`git ls-files -z`.split("\x0").reject do |f|
|
|
25
|
-
(f == __FILE__) || f.match(%r{\A(?:(?:bin|test|spec|features)/|\.(?:git|circleci)|appveyor)})
|
|
26
|
-
end
|
|
27
|
-
end
|
|
28
|
-
spec.bindir = "exe"
|
|
29
|
-
spec.executables = spec.files.grep(%r{\Aexe/}) { |f| File.basename(f) }
|
|
30
|
-
spec.require_paths = ["lib"]
|
|
31
|
-
spec.extensions = ["ext/tiktoken_ruby/extconf.rb"]
|
|
32
|
-
spec.add_dependency "rb_sys", ">= 0.9.86"
|
|
33
|
-
end
|