python3Packages.tokenizers: 0.21.1 -> 0.21.2 (#419584)

authored by Gaétan Lepage and committed by GitHub 82eeda2a 5d6b848b

+1361 -16
+1334
pkgs/development/python-modules/tokenizers/Cargo.lock
··· 1 + # This file is automatically @generated by Cargo. 2 + # It is not intended for manual editing. 3 + version = 4 4 + 5 + [[package]] 6 + name = "ahash" 7 + version = "0.8.12" 8 + source = "registry+https://github.com/rust-lang/crates.io-index" 9 + checksum = "5a15f179cd60c4584b8a8c596927aadc462e27f2ca70c04e0071964a73ba7a75" 10 + dependencies = [ 11 + "cfg-if", 12 + "getrandom", 13 + "once_cell", 14 + "serde", 15 + "version_check", 16 + "zerocopy", 17 + ] 18 + 19 + [[package]] 20 + name = "aho-corasick" 21 + version = "1.1.3" 22 + source = "registry+https://github.com/rust-lang/crates.io-index" 23 + checksum = "8e60d3430d3a69478ad0993f19238d2df97c507009a52b3c10addcd7f6bcb916" 24 + dependencies = [ 25 + "memchr", 26 + ] 27 + 28 + [[package]] 29 + name = "anstream" 30 + version = "0.6.19" 31 + source = "registry+https://github.com/rust-lang/crates.io-index" 32 + checksum = "301af1932e46185686725e0fad2f8f2aa7da69dd70bf6ecc44d6b703844a3933" 33 + dependencies = [ 34 + "anstyle", 35 + "anstyle-parse", 36 + "anstyle-query", 37 + "anstyle-wincon", 38 + "colorchoice", 39 + "is_terminal_polyfill", 40 + "utf8parse", 41 + ] 42 + 43 + [[package]] 44 + name = "anstyle" 45 + version = "1.0.11" 46 + source = "registry+https://github.com/rust-lang/crates.io-index" 47 + checksum = "862ed96ca487e809f1c8e5a8447f6ee2cf102f846893800b20cebdf541fc6bbd" 48 + 49 + [[package]] 50 + name = "anstyle-parse" 51 + version = "0.2.7" 52 + source = "registry+https://github.com/rust-lang/crates.io-index" 53 + checksum = "4e7644824f0aa2c7b9384579234ef10eb7efb6a0deb83f9630a49594dd9c15c2" 54 + dependencies = [ 55 + "utf8parse", 56 + ] 57 + 58 + [[package]] 59 + name = "anstyle-query" 60 + version = "1.1.3" 61 + source = "registry+https://github.com/rust-lang/crates.io-index" 62 + checksum = "6c8bdeb6047d8983be085bab0ba1472e6dc604e7041dbf6fcd5e71523014fae9" 63 + dependencies = [ 64 + "windows-sys 0.59.0", 65 + ] 66 + 67 + [[package]] 68 + name = "anstyle-wincon" 69 + version = "3.0.9" 70 + source = "registry+https://github.com/rust-lang/crates.io-index" 71 + checksum = "403f75924867bb1033c59fbf0797484329750cfbe3c4325cd33127941fabc882" 72 + dependencies = [ 73 + "anstyle", 74 + "once_cell_polyfill", 75 + "windows-sys 0.59.0", 76 + ] 77 + 78 + [[package]] 79 + name = "autocfg" 80 + version = "1.5.0" 81 + source = "registry+https://github.com/rust-lang/crates.io-index" 82 + checksum = "c08606f8c3cbf4ce6ec8e28fb0014a2c086708fe954eaa885384a6165172e7e8" 83 + 84 + [[package]] 85 + name = "base64" 86 + version = "0.13.1" 87 + source = "registry+https://github.com/rust-lang/crates.io-index" 88 + checksum = "9e1b586273c5702936fe7b7d6896644d8be71e6314cfe09d3167c95f712589e8" 89 + 90 + [[package]] 91 + name = "bitflags" 92 + version = "2.9.1" 93 + source = "registry+https://github.com/rust-lang/crates.io-index" 94 + checksum = "1b8e56985ec62d17e9c1001dc89c88ecd7dc08e47eba5ec7c29c7b5eeecde967" 95 + 96 + [[package]] 97 + name = "bumpalo" 98 + version = "3.18.1" 99 + source = "registry+https://github.com/rust-lang/crates.io-index" 100 + checksum = "793db76d6187cd04dff33004d8e6c9cc4e05cd330500379d2394209271b4aeee" 101 + 102 + [[package]] 103 + name = "castaway" 104 + version = "0.2.3" 105 + source = "registry+https://github.com/rust-lang/crates.io-index" 106 + checksum = "0abae9be0aaf9ea96a3b1b8b1b55c602ca751eba1b1500220cea4ecbafe7c0d5" 107 + dependencies = [ 108 + "rustversion", 109 + ] 110 + 111 + [[package]] 112 + name = "cc" 113 + version = "1.2.27" 114 + source = "registry+https://github.com/rust-lang/crates.io-index" 115 + checksum = "d487aa071b5f64da6f19a3e848e3578944b726ee5a4854b82172f02aa876bfdc" 116 + dependencies = [ 117 + "shlex", 118 + ] 119 + 120 + [[package]] 121 + name = "cfg-if" 122 + version = "1.0.1" 123 + source = "registry+https://github.com/rust-lang/crates.io-index" 124 + checksum = "9555578bc9e57714c812a1f84e4fc5b4d21fcb063490c624de019f7464c91268" 125 + 126 + [[package]] 127 + name = "colorchoice" 128 + version = "1.0.4" 129 + source = "registry+https://github.com/rust-lang/crates.io-index" 130 + checksum = "b05b61dc5112cbb17e4b6cd61790d9845d13888356391624cbe7e41efeac1e75" 131 + 132 + [[package]] 133 + name = "compact_str" 134 + version = "0.9.0" 135 + source = "registry+https://github.com/rust-lang/crates.io-index" 136 + checksum = "3fdb1325a1cece981e8a296ab8f0f9b63ae357bd0784a9faaf548cc7b480707a" 137 + dependencies = [ 138 + "castaway", 139 + "cfg-if", 140 + "itoa", 141 + "rustversion", 142 + "ryu", 143 + "serde", 144 + "static_assertions", 145 + ] 146 + 147 + [[package]] 148 + name = "console" 149 + version = "0.15.11" 150 + source = "registry+https://github.com/rust-lang/crates.io-index" 151 + checksum = "054ccb5b10f9f2cbf51eb355ca1d05c2d279ce1804688d0db74b4733a5aeafd8" 152 + dependencies = [ 153 + "encode_unicode", 154 + "libc", 155 + "once_cell", 156 + "unicode-width", 157 + "windows-sys 0.59.0", 158 + ] 159 + 160 + [[package]] 161 + name = "crossbeam-deque" 162 + version = "0.8.6" 163 + source = "registry+https://github.com/rust-lang/crates.io-index" 164 + checksum = "9dd111b7b7f7d55b72c0a6ae361660ee5853c9af73f70c3c2ef6858b950e2e51" 165 + dependencies = [ 166 + "crossbeam-epoch", 167 + "crossbeam-utils", 168 + ] 169 + 170 + [[package]] 171 + name = "crossbeam-epoch" 172 + version = "0.9.18" 173 + source = "registry+https://github.com/rust-lang/crates.io-index" 174 + checksum = "5b82ac4a3c2ca9c3460964f020e1402edd5753411d7737aa39c3714ad1b5420e" 175 + dependencies = [ 176 + "crossbeam-utils", 177 + ] 178 + 179 + [[package]] 180 + name = "crossbeam-utils" 181 + version = "0.8.21" 182 + source = "registry+https://github.com/rust-lang/crates.io-index" 183 + checksum = "d0a5c400df2834b80a4c3327b3aad3a4c4cd4de0629063962b03235697506a28" 184 + 185 + [[package]] 186 + name = "darling" 187 + version = "0.20.11" 188 + source = "registry+https://github.com/rust-lang/crates.io-index" 189 + checksum = "fc7f46116c46ff9ab3eb1597a45688b6715c6e628b5c133e288e709a29bcb4ee" 190 + dependencies = [ 191 + "darling_core", 192 + "darling_macro", 193 + ] 194 + 195 + [[package]] 196 + name = "darling_core" 197 + version = "0.20.11" 198 + source = "registry+https://github.com/rust-lang/crates.io-index" 199 + checksum = "0d00b9596d185e565c2207a0b01f8bd1a135483d02d9b7b0a54b11da8d53412e" 200 + dependencies = [ 201 + "fnv", 202 + "ident_case", 203 + "proc-macro2", 204 + "quote", 205 + "strsim", 206 + "syn", 207 + ] 208 + 209 + [[package]] 210 + name = "darling_macro" 211 + version = "0.20.11" 212 + source = "registry+https://github.com/rust-lang/crates.io-index" 213 + checksum = "fc34b93ccb385b40dc71c6fceac4b2ad23662c7eeb248cf10d529b7e055b6ead" 214 + dependencies = [ 215 + "darling_core", 216 + "quote", 217 + "syn", 218 + ] 219 + 220 + [[package]] 221 + name = "dary_heap" 222 + version = "0.3.7" 223 + source = "registry+https://github.com/rust-lang/crates.io-index" 224 + checksum = "04d2cd9c18b9f454ed67da600630b021a8a80bf33f8c95896ab33aaf1c26b728" 225 + dependencies = [ 226 + "serde", 227 + ] 228 + 229 + [[package]] 230 + name = "derive_builder" 231 + version = "0.20.2" 232 + source = "registry+https://github.com/rust-lang/crates.io-index" 233 + checksum = "507dfb09ea8b7fa618fcf76e953f4f5e192547945816d5358edffe39f6f94947" 234 + dependencies = [ 235 + "derive_builder_macro", 236 + ] 237 + 238 + [[package]] 239 + name = "derive_builder_core" 240 + version = "0.20.2" 241 + source = "registry+https://github.com/rust-lang/crates.io-index" 242 + checksum = "2d5bcf7b024d6835cfb3d473887cd966994907effbe9227e8c8219824d06c4e8" 243 + dependencies = [ 244 + "darling", 245 + "proc-macro2", 246 + "quote", 247 + "syn", 248 + ] 249 + 250 + [[package]] 251 + name = "derive_builder_macro" 252 + version = "0.20.2" 253 + source = "registry+https://github.com/rust-lang/crates.io-index" 254 + checksum = "ab63b0e2bf4d5928aff72e83a7dace85d7bba5fe12dcc3c5a572d78caffd3f3c" 255 + dependencies = [ 256 + "derive_builder_core", 257 + "syn", 258 + ] 259 + 260 + [[package]] 261 + name = "either" 262 + version = "1.15.0" 263 + source = "registry+https://github.com/rust-lang/crates.io-index" 264 + checksum = "48c757948c5ede0e46177b7add2e67155f70e33c07fea8284df6576da70b3719" 265 + 266 + [[package]] 267 + name = "encode_unicode" 268 + version = "1.0.0" 269 + source = "registry+https://github.com/rust-lang/crates.io-index" 270 + checksum = "34aa73646ffb006b8f5147f3dc182bd4bcb190227ce861fc4a4844bf8e3cb2c0" 271 + 272 + [[package]] 273 + name = "env_filter" 274 + version = "0.1.3" 275 + source = "registry+https://github.com/rust-lang/crates.io-index" 276 + checksum = "186e05a59d4c50738528153b83b0b0194d3a29507dfec16eccd4b342903397d0" 277 + dependencies = [ 278 + "log", 279 + "regex", 280 + ] 281 + 282 + [[package]] 283 + name = "env_logger" 284 + version = "0.11.8" 285 + source = "registry+https://github.com/rust-lang/crates.io-index" 286 + checksum = "13c863f0904021b108aa8b2f55046443e6b1ebde8fd4a15c399893aae4fa069f" 287 + dependencies = [ 288 + "anstream", 289 + "anstyle", 290 + "env_filter", 291 + "jiff", 292 + "log", 293 + ] 294 + 295 + [[package]] 296 + name = "errno" 297 + version = "0.3.13" 298 + source = "registry+https://github.com/rust-lang/crates.io-index" 299 + checksum = "778e2ac28f6c47af28e4907f13ffd1e1ddbd400980a9abd7c8df189bf578a5ad" 300 + dependencies = [ 301 + "libc", 302 + "windows-sys 0.60.2", 303 + ] 304 + 305 + [[package]] 306 + name = "esaxx-rs" 307 + version = "0.1.10" 308 + source = "registry+https://github.com/rust-lang/crates.io-index" 309 + checksum = "d817e038c30374a4bcb22f94d0a8a0e216958d4c3dcde369b1439fec4bdda6e6" 310 + dependencies = [ 311 + "cc", 312 + ] 313 + 314 + [[package]] 315 + name = "fastrand" 316 + version = "2.3.0" 317 + source = "registry+https://github.com/rust-lang/crates.io-index" 318 + checksum = "37909eebbb50d72f9059c3b6d82c0463f2ff062c9e95845c43a6c9c0355411be" 319 + 320 + [[package]] 321 + name = "fnv" 322 + version = "1.0.7" 323 + source = "registry+https://github.com/rust-lang/crates.io-index" 324 + checksum = "3f9eec918d3f24069decb9af1554cad7c880e2da24a9afd88aca000531ab82c1" 325 + 326 + [[package]] 327 + name = "getrandom" 328 + version = "0.3.3" 329 + source = "registry+https://github.com/rust-lang/crates.io-index" 330 + checksum = "26145e563e54f2cadc477553f1ec5ee650b00862f0a58bcd12cbdc5f0ea2d2f4" 331 + dependencies = [ 332 + "cfg-if", 333 + "libc", 334 + "r-efi", 335 + "wasi", 336 + ] 337 + 338 + [[package]] 339 + name = "heck" 340 + version = "0.5.0" 341 + source = "registry+https://github.com/rust-lang/crates.io-index" 342 + checksum = "2304e00983f87ffb38b55b444b5e3b60a884b5d30c0fca7d82fe33449bbe55ea" 343 + 344 + [[package]] 345 + name = "ident_case" 346 + version = "1.0.1" 347 + source = "registry+https://github.com/rust-lang/crates.io-index" 348 + checksum = "b9e0384b61958566e926dc50660321d12159025e767c18e043daf26b70104c39" 349 + 350 + [[package]] 351 + name = "indicatif" 352 + version = "0.17.11" 353 + source = "registry+https://github.com/rust-lang/crates.io-index" 354 + checksum = "183b3088984b400f4cfac3620d5e076c84da5364016b4f49473de574b2586235" 355 + dependencies = [ 356 + "console", 357 + "number_prefix", 358 + "portable-atomic", 359 + "unicode-width", 360 + "web-time", 361 + ] 362 + 363 + [[package]] 364 + name = "indoc" 365 + version = "2.0.6" 366 + source = "registry+https://github.com/rust-lang/crates.io-index" 367 + checksum = "f4c7245a08504955605670dbf141fceab975f15ca21570696aebe9d2e71576bd" 368 + 369 + [[package]] 370 + name = "is_terminal_polyfill" 371 + version = "1.70.1" 372 + source = "registry+https://github.com/rust-lang/crates.io-index" 373 + checksum = "7943c866cc5cd64cbc25b2e01621d07fa8eb2a1a23160ee81ce38704e97b8ecf" 374 + 375 + [[package]] 376 + name = "itertools" 377 + version = "0.14.0" 378 + source = "registry+https://github.com/rust-lang/crates.io-index" 379 + checksum = "2b192c782037fadd9cfa75548310488aabdbf3d2da73885b31bd0abd03351285" 380 + dependencies = [ 381 + "either", 382 + ] 383 + 384 + [[package]] 385 + name = "itoa" 386 + version = "1.0.15" 387 + source = "registry+https://github.com/rust-lang/crates.io-index" 388 + checksum = "4a5f13b858c8d314ee3e8f639011f7ccefe71f97f96e50151fb991f267928e2c" 389 + 390 + [[package]] 391 + name = "jiff" 392 + version = "0.2.15" 393 + source = "registry+https://github.com/rust-lang/crates.io-index" 394 + checksum = "be1f93b8b1eb69c77f24bbb0afdf66f54b632ee39af40ca21c4365a1d7347e49" 395 + dependencies = [ 396 + "jiff-static", 397 + "log", 398 + "portable-atomic", 399 + "portable-atomic-util", 400 + "serde", 401 + ] 402 + 403 + [[package]] 404 + name = "jiff-static" 405 + version = "0.2.15" 406 + source = "registry+https://github.com/rust-lang/crates.io-index" 407 + checksum = "03343451ff899767262ec32146f6d559dd759fdadf42ff0e227c7c48f72594b4" 408 + dependencies = [ 409 + "proc-macro2", 410 + "quote", 411 + "syn", 412 + ] 413 + 414 + [[package]] 415 + name = "js-sys" 416 + version = "0.3.77" 417 + source = "registry+https://github.com/rust-lang/crates.io-index" 418 + checksum = "1cfaf33c695fc6e08064efbc1f72ec937429614f25eef83af942d0e227c3a28f" 419 + dependencies = [ 420 + "once_cell", 421 + "wasm-bindgen", 422 + ] 423 + 424 + [[package]] 425 + name = "libc" 426 + version = "0.2.174" 427 + source = "registry+https://github.com/rust-lang/crates.io-index" 428 + checksum = "1171693293099992e19cddea4e8b849964e9846f4acee11b3948bcc337be8776" 429 + 430 + [[package]] 431 + name = "linux-raw-sys" 432 + version = "0.9.4" 433 + source = "registry+https://github.com/rust-lang/crates.io-index" 434 + checksum = "cd945864f07fe9f5371a27ad7b52a172b4b499999f1d97574c9fa68373937e12" 435 + 436 + [[package]] 437 + name = "log" 438 + version = "0.4.27" 439 + source = "registry+https://github.com/rust-lang/crates.io-index" 440 + checksum = "13dc2df351e3202783a1fe0d44375f7295ffb4049267b0f3018346dc122a1d94" 441 + 442 + [[package]] 443 + name = "macro_rules_attribute" 444 + version = "0.2.2" 445 + source = "registry+https://github.com/rust-lang/crates.io-index" 446 + checksum = "65049d7923698040cd0b1ddcced9b0eb14dd22c5f86ae59c3740eab64a676520" 447 + dependencies = [ 448 + "macro_rules_attribute-proc_macro", 449 + "paste", 450 + ] 451 + 452 + [[package]] 453 + name = "macro_rules_attribute-proc_macro" 454 + version = "0.2.2" 455 + source = "registry+https://github.com/rust-lang/crates.io-index" 456 + checksum = "670fdfda89751bc4a84ac13eaa63e205cf0fd22b4c9a5fbfa085b63c1f1d3a30" 457 + 458 + [[package]] 459 + name = "matrixmultiply" 460 + version = "0.3.10" 461 + source = "registry+https://github.com/rust-lang/crates.io-index" 462 + checksum = "a06de3016e9fae57a36fd14dba131fccf49f74b40b7fbdb472f96e361ec71a08" 463 + dependencies = [ 464 + "autocfg", 465 + "rawpointer", 466 + ] 467 + 468 + [[package]] 469 + name = "memchr" 470 + version = "2.7.5" 471 + source = "registry+https://github.com/rust-lang/crates.io-index" 472 + checksum = "32a282da65faaf38286cf3be983213fcf1d2e2a58700e808f83f4ea9a4804bc0" 473 + 474 + [[package]] 475 + name = "memoffset" 476 + version = "0.9.1" 477 + source = "registry+https://github.com/rust-lang/crates.io-index" 478 + checksum = "488016bfae457b036d996092f6cb448677611ce4449e970ceaf42695203f218a" 479 + dependencies = [ 480 + "autocfg", 481 + ] 482 + 483 + [[package]] 484 + name = "minimal-lexical" 485 + version = "0.2.1" 486 + source = "registry+https://github.com/rust-lang/crates.io-index" 487 + checksum = "68354c5c6bd36d73ff3feceb05efa59b6acb7626617f4962be322a825e61f79a" 488 + 489 + [[package]] 490 + name = "monostate" 491 + version = "0.1.14" 492 + source = "registry+https://github.com/rust-lang/crates.io-index" 493 + checksum = "aafe1be9d0c75642e3e50fedc7ecadf1ef1cbce6eb66462153fc44245343fbee" 494 + dependencies = [ 495 + "monostate-impl", 496 + "serde", 497 + ] 498 + 499 + [[package]] 500 + name = "monostate-impl" 501 + version = "0.1.14" 502 + source = "registry+https://github.com/rust-lang/crates.io-index" 503 + checksum = "c402a4092d5e204f32c9e155431046831fa712637043c58cb73bc6bc6c9663b5" 504 + dependencies = [ 505 + "proc-macro2", 506 + "quote", 507 + "syn", 508 + ] 509 + 510 + [[package]] 511 + name = "ndarray" 512 + version = "0.16.1" 513 + source = "registry+https://github.com/rust-lang/crates.io-index" 514 + checksum = "882ed72dce9365842bf196bdeedf5055305f11fc8c03dee7bb0194a6cad34841" 515 + dependencies = [ 516 + "matrixmultiply", 517 + "num-complex", 518 + "num-integer", 519 + "num-traits", 520 + "portable-atomic", 521 + "portable-atomic-util", 522 + "rawpointer", 523 + ] 524 + 525 + [[package]] 526 + name = "nom" 527 + version = "7.1.3" 528 + source = "registry+https://github.com/rust-lang/crates.io-index" 529 + checksum = "d273983c5a657a70a3e8f2a01329822f3b8c8172b73826411a55751e404a0a4a" 530 + dependencies = [ 531 + "memchr", 532 + "minimal-lexical", 533 + ] 534 + 535 + [[package]] 536 + name = "num-complex" 537 + version = "0.4.6" 538 + source = "registry+https://github.com/rust-lang/crates.io-index" 539 + checksum = "73f88a1307638156682bada9d7604135552957b7818057dcef22705b4d509495" 540 + dependencies = [ 541 + "num-traits", 542 + ] 543 + 544 + [[package]] 545 + name = "num-integer" 546 + version = "0.1.46" 547 + source = "registry+https://github.com/rust-lang/crates.io-index" 548 + checksum = "7969661fd2958a5cb096e56c8e1ad0444ac2bbcd0061bd28660485a44879858f" 549 + dependencies = [ 550 + "num-traits", 551 + ] 552 + 553 + [[package]] 554 + name = "num-traits" 555 + version = "0.2.19" 556 + source = "registry+https://github.com/rust-lang/crates.io-index" 557 + checksum = "071dfc062690e90b734c0b2273ce72ad0ffa95f0c74596bc250dcfd960262841" 558 + dependencies = [ 559 + "autocfg", 560 + ] 561 + 562 + [[package]] 563 + name = "number_prefix" 564 + version = "0.4.0" 565 + source = "registry+https://github.com/rust-lang/crates.io-index" 566 + checksum = "830b246a0e5f20af87141b25c173cd1b609bd7779a4617d6ec582abaf90870f3" 567 + 568 + [[package]] 569 + name = "numpy" 570 + version = "0.25.0" 571 + source = "registry+https://github.com/rust-lang/crates.io-index" 572 + checksum = "29f1dee9aa8d3f6f8e8b9af3803006101bb3653866ef056d530d53ae68587191" 573 + dependencies = [ 574 + "libc", 575 + "ndarray", 576 + "num-complex", 577 + "num-integer", 578 + "num-traits", 579 + "pyo3", 580 + "pyo3-build-config", 581 + "rustc-hash", 582 + ] 583 + 584 + [[package]] 585 + name = "once_cell" 586 + version = "1.21.3" 587 + source = "registry+https://github.com/rust-lang/crates.io-index" 588 + checksum = "42f5e15c9953c5e4ccceeb2e7382a716482c34515315f7b03532b8b4e8393d2d" 589 + 590 + [[package]] 591 + name = "once_cell_polyfill" 592 + version = "1.70.1" 593 + source = "registry+https://github.com/rust-lang/crates.io-index" 594 + checksum = "a4895175b425cb1f87721b59f0f286c2092bd4af812243672510e1ac53e2e0ad" 595 + 596 + [[package]] 597 + name = "onig" 598 + version = "6.5.1" 599 + source = "registry+https://github.com/rust-lang/crates.io-index" 600 + checksum = "336b9c63443aceef14bea841b899035ae3abe89b7c486aaf4c5bd8aafedac3f0" 601 + dependencies = [ 602 + "bitflags", 603 + "libc", 604 + "once_cell", 605 + "onig_sys", 606 + ] 607 + 608 + [[package]] 609 + name = "onig_sys" 610 + version = "69.9.1" 611 + source = "registry+https://github.com/rust-lang/crates.io-index" 612 + checksum = "c7f86c6eef3d6df15f23bcfb6af487cbd2fed4e5581d58d5bf1f5f8b7f6727dc" 613 + dependencies = [ 614 + "cc", 615 + "pkg-config", 616 + ] 617 + 618 + [[package]] 619 + name = "paste" 620 + version = "1.0.15" 621 + source = "registry+https://github.com/rust-lang/crates.io-index" 622 + checksum = "57c0d7b74b563b49d38dae00a0c37d4d6de9b432382b2892f0574ddcae73fd0a" 623 + 624 + [[package]] 625 + name = "pkg-config" 626 + version = "0.3.32" 627 + source = "registry+https://github.com/rust-lang/crates.io-index" 628 + checksum = "7edddbd0b52d732b21ad9a5fab5c704c14cd949e5e9a1ec5929a24fded1b904c" 629 + 630 + [[package]] 631 + name = "portable-atomic" 632 + version = "1.11.1" 633 + source = "registry+https://github.com/rust-lang/crates.io-index" 634 + checksum = "f84267b20a16ea918e43c6a88433c2d54fa145c92a811b5b047ccbe153674483" 635 + 636 + [[package]] 637 + name = "portable-atomic-util" 638 + version = "0.2.4" 639 + source = "registry+https://github.com/rust-lang/crates.io-index" 640 + checksum = "d8a2f0d8d040d7848a709caf78912debcc3f33ee4b3cac47d73d1e1069e83507" 641 + dependencies = [ 642 + "portable-atomic", 643 + ] 644 + 645 + [[package]] 646 + name = "ppv-lite86" 647 + version = "0.2.21" 648 + source = "registry+https://github.com/rust-lang/crates.io-index" 649 + checksum = "85eae3c4ed2f50dcfe72643da4befc30deadb458a9b590d720cde2f2b1e97da9" 650 + dependencies = [ 651 + "zerocopy", 652 + ] 653 + 654 + [[package]] 655 + name = "proc-macro2" 656 + version = "1.0.95" 657 + source = "registry+https://github.com/rust-lang/crates.io-index" 658 + checksum = "02b3e5e68a3a1a02aad3ec490a98007cbc13c37cbe84a3cd7b8e406d76e7f778" 659 + dependencies = [ 660 + "unicode-ident", 661 + ] 662 + 663 + [[package]] 664 + name = "pyo3" 665 + version = "0.25.1" 666 + source = "registry+https://github.com/rust-lang/crates.io-index" 667 + checksum = "8970a78afe0628a3e3430376fc5fd76b6b45c4d43360ffd6cdd40bdde72b682a" 668 + dependencies = [ 669 + "indoc", 670 + "libc", 671 + "memoffset", 672 + "once_cell", 673 + "portable-atomic", 674 + "pyo3-build-config", 675 + "pyo3-ffi", 676 + "pyo3-macros", 677 + "unindent", 678 + ] 679 + 680 + [[package]] 681 + name = "pyo3-build-config" 682 + version = "0.25.1" 683 + source = "registry+https://github.com/rust-lang/crates.io-index" 684 + checksum = "458eb0c55e7ece017adeba38f2248ff3ac615e53660d7c71a238d7d2a01c7598" 685 + dependencies = [ 686 + "once_cell", 687 + "target-lexicon", 688 + ] 689 + 690 + [[package]] 691 + name = "pyo3-ffi" 692 + version = "0.25.1" 693 + source = "registry+https://github.com/rust-lang/crates.io-index" 694 + checksum = "7114fe5457c61b276ab77c5055f206295b812608083644a5c5b2640c3102565c" 695 + dependencies = [ 696 + "libc", 697 + "pyo3-build-config", 698 + ] 699 + 700 + [[package]] 701 + name = "pyo3-macros" 702 + version = "0.25.1" 703 + source = "registry+https://github.com/rust-lang/crates.io-index" 704 + checksum = "a8725c0a622b374d6cb051d11a0983786448f7785336139c3c94f5aa6bef7e50" 705 + dependencies = [ 706 + "proc-macro2", 707 + "pyo3-macros-backend", 708 + "quote", 709 + "syn", 710 + ] 711 + 712 + [[package]] 713 + name = "pyo3-macros-backend" 714 + version = "0.25.1" 715 + source = "registry+https://github.com/rust-lang/crates.io-index" 716 + checksum = "4109984c22491085343c05b0dbc54ddc405c3cf7b4374fc533f5c3313a572ccc" 717 + dependencies = [ 718 + "heck", 719 + "proc-macro2", 720 + "pyo3-build-config", 721 + "quote", 722 + "syn", 723 + ] 724 + 725 + [[package]] 726 + name = "quote" 727 + version = "1.0.40" 728 + source = "registry+https://github.com/rust-lang/crates.io-index" 729 + checksum = "1885c039570dc00dcb4ff087a89e185fd56bae234ddc7f056a945bf36467248d" 730 + dependencies = [ 731 + "proc-macro2", 732 + ] 733 + 734 + [[package]] 735 + name = "r-efi" 736 + version = "5.3.0" 737 + source = "registry+https://github.com/rust-lang/crates.io-index" 738 + checksum = "69cdb34c158ceb288df11e18b4bd39de994f6657d83847bdffdbd7f346754b0f" 739 + 740 + [[package]] 741 + name = "rand" 742 + version = "0.9.1" 743 + source = "registry+https://github.com/rust-lang/crates.io-index" 744 + checksum = "9fbfd9d094a40bf3ae768db9361049ace4c0e04a4fd6b359518bd7b73a73dd97" 745 + dependencies = [ 746 + "rand_chacha", 747 + "rand_core", 748 + ] 749 + 750 + [[package]] 751 + name = "rand_chacha" 752 + version = "0.9.0" 753 + source = "registry+https://github.com/rust-lang/crates.io-index" 754 + checksum = "d3022b5f1df60f26e1ffddd6c66e8aa15de382ae63b3a0c1bfc0e4d3e3f325cb" 755 + dependencies = [ 756 + "ppv-lite86", 757 + "rand_core", 758 + ] 759 + 760 + [[package]] 761 + name = "rand_core" 762 + version = "0.9.3" 763 + source = "registry+https://github.com/rust-lang/crates.io-index" 764 + checksum = "99d9a13982dcf210057a8a78572b2217b667c3beacbf3a0d8b454f6f82837d38" 765 + dependencies = [ 766 + "getrandom", 767 + ] 768 + 769 + [[package]] 770 + name = "rawpointer" 771 + version = "0.2.1" 772 + source = "registry+https://github.com/rust-lang/crates.io-index" 773 + checksum = "60a357793950651c4ed0f3f52338f53b2f809f32d83a07f72909fa13e4c6c1e3" 774 + 775 + [[package]] 776 + name = "rayon" 777 + version = "1.10.0" 778 + source = "registry+https://github.com/rust-lang/crates.io-index" 779 + checksum = "b418a60154510ca1a002a752ca9714984e21e4241e804d32555251faf8b78ffa" 780 + dependencies = [ 781 + "either", 782 + "rayon-core", 783 + ] 784 + 785 + [[package]] 786 + name = "rayon-cond" 787 + version = "0.4.0" 788 + source = "registry+https://github.com/rust-lang/crates.io-index" 789 + checksum = "2964d0cf57a3e7a06e8183d14a8b527195c706b7983549cd5462d5aa3747438f" 790 + dependencies = [ 791 + "either", 792 + "itertools", 793 + "rayon", 794 + ] 795 + 796 + [[package]] 797 + name = "rayon-core" 798 + version = "1.12.1" 799 + source = "registry+https://github.com/rust-lang/crates.io-index" 800 + checksum = "1465873a3dfdaa8ae7cb14b4383657caab0b3e8a0aa9ae8e04b044854c8dfce2" 801 + dependencies = [ 802 + "crossbeam-deque", 803 + "crossbeam-utils", 804 + ] 805 + 806 + [[package]] 807 + name = "regex" 808 + version = "1.11.1" 809 + source = "registry+https://github.com/rust-lang/crates.io-index" 810 + checksum = "b544ef1b4eac5dc2db33ea63606ae9ffcfac26c1416a2806ae0bf5f56b201191" 811 + dependencies = [ 812 + "aho-corasick", 813 + "memchr", 814 + "regex-automata", 815 + "regex-syntax", 816 + ] 817 + 818 + [[package]] 819 + name = "regex-automata" 820 + version = "0.4.9" 821 + source = "registry+https://github.com/rust-lang/crates.io-index" 822 + checksum = "809e8dc61f6de73b46c85f4c96486310fe304c434cfa43669d7b40f711150908" 823 + dependencies = [ 824 + "aho-corasick", 825 + "memchr", 826 + "regex-syntax", 827 + ] 828 + 829 + [[package]] 830 + name = "regex-syntax" 831 + version = "0.8.5" 832 + source = "registry+https://github.com/rust-lang/crates.io-index" 833 + checksum = "2b15c43186be67a4fd63bee50d0303afffcef381492ebe2c5d87f324e1b8815c" 834 + 835 + [[package]] 836 + name = "rustc-hash" 837 + version = "2.1.1" 838 + source = "registry+https://github.com/rust-lang/crates.io-index" 839 + checksum = "357703d41365b4b27c590e3ed91eabb1b663f07c4c084095e60cbed4362dff0d" 840 + 841 + [[package]] 842 + name = "rustix" 843 + version = "1.0.7" 844 + source = "registry+https://github.com/rust-lang/crates.io-index" 845 + checksum = "c71e83d6afe7ff64890ec6b71d6a69bb8a610ab78ce364b3352876bb4c801266" 846 + dependencies = [ 847 + "bitflags", 848 + "errno", 849 + "libc", 850 + "linux-raw-sys", 851 + "windows-sys 0.59.0", 852 + ] 853 + 854 + [[package]] 855 + name = "rustversion" 856 + version = "1.0.21" 857 + source = "registry+https://github.com/rust-lang/crates.io-index" 858 + checksum = "8a0d197bd2c9dc6e53b84da9556a69ba4cdfab8619eb41a8bd1cc2027a0f6b1d" 859 + 860 + [[package]] 861 + name = "ryu" 862 + version = "1.0.20" 863 + source = "registry+https://github.com/rust-lang/crates.io-index" 864 + checksum = "28d3b2b1366ec20994f1fd18c3c594f05c5dd4bc44d8bb0c1c632c8d6829481f" 865 + 866 + [[package]] 867 + name = "serde" 868 + version = "1.0.219" 869 + source = "registry+https://github.com/rust-lang/crates.io-index" 870 + checksum = "5f0e2c6ed6606019b4e29e69dbaba95b11854410e5347d525002456dbbb786b6" 871 + dependencies = [ 872 + "serde_derive", 873 + ] 874 + 875 + [[package]] 876 + name = "serde_derive" 877 + version = "1.0.219" 878 + source = "registry+https://github.com/rust-lang/crates.io-index" 879 + checksum = "5b0276cf7f2c73365f7157c8123c21cd9a50fbbd844757af28ca1f5925fc2a00" 880 + dependencies = [ 881 + "proc-macro2", 882 + "quote", 883 + "syn", 884 + ] 885 + 886 + [[package]] 887 + name = "serde_json" 888 + version = "1.0.140" 889 + source = "registry+https://github.com/rust-lang/crates.io-index" 890 + checksum = "20068b6e96dc6c9bd23e01df8827e6c7e1f2fddd43c21810382803c136b99373" 891 + dependencies = [ 892 + "itoa", 893 + "memchr", 894 + "ryu", 895 + "serde", 896 + ] 897 + 898 + [[package]] 899 + name = "shlex" 900 + version = "1.3.0" 901 + source = "registry+https://github.com/rust-lang/crates.io-index" 902 + checksum = "0fda2ff0d084019ba4d7c6f371c95d8fd75ce3524c3cb8fb653a3023f6323e64" 903 + 904 + [[package]] 905 + name = "smallvec" 906 + version = "1.15.1" 907 + source = "registry+https://github.com/rust-lang/crates.io-index" 908 + checksum = "67b1b7a3b5fe4f1376887184045fcf45c69e92af734b7aaddc05fb777b6fbd03" 909 + 910 + [[package]] 911 + name = "spm_precompiled" 912 + version = "0.1.4" 913 + source = "registry+https://github.com/rust-lang/crates.io-index" 914 + checksum = "5851699c4033c63636f7ea4cf7b7c1f1bf06d0cc03cfb42e711de5a5c46cf326" 915 + dependencies = [ 916 + "base64", 917 + "nom", 918 + "serde", 919 + "unicode-segmentation", 920 + ] 921 + 922 + [[package]] 923 + name = "static_assertions" 924 + version = "1.1.0" 925 + source = "registry+https://github.com/rust-lang/crates.io-index" 926 + checksum = "a2eb9349b6444b326872e140eb1cf5e7c522154d69e7a0ffb0fb81c06b37543f" 927 + 928 + [[package]] 929 + name = "strsim" 930 + version = "0.11.1" 931 + source = "registry+https://github.com/rust-lang/crates.io-index" 932 + checksum = "7da8b5736845d9f2fcb837ea5d9e2628564b3b043a70948a3f0b778838c5fb4f" 933 + 934 + [[package]] 935 + name = "syn" 936 + version = "2.0.104" 937 + source = "registry+https://github.com/rust-lang/crates.io-index" 938 + checksum = "17b6f705963418cdb9927482fa304bc562ece2fdd4f616084c50b7023b435a40" 939 + dependencies = [ 940 + "proc-macro2", 941 + "quote", 942 + "unicode-ident", 943 + ] 944 + 945 + [[package]] 946 + name = "target-lexicon" 947 + version = "0.13.2" 948 + source = "registry+https://github.com/rust-lang/crates.io-index" 949 + checksum = "e502f78cdbb8ba4718f566c418c52bc729126ffd16baee5baa718cf25dd5a69a" 950 + 951 + [[package]] 952 + name = "tempfile" 953 + version = "3.20.0" 954 + source = "registry+https://github.com/rust-lang/crates.io-index" 955 + checksum = "e8a64e3985349f2441a1a9ef0b853f869006c3855f2cda6862a94d26ebb9d6a1" 956 + dependencies = [ 957 + "fastrand", 958 + "getrandom", 959 + "once_cell", 960 + "rustix", 961 + "windows-sys 0.59.0", 962 + ] 963 + 964 + [[package]] 965 + name = "thiserror" 966 + version = "2.0.12" 967 + source = "registry+https://github.com/rust-lang/crates.io-index" 968 + checksum = "567b8a2dae586314f7be2a752ec7474332959c6460e02bde30d702a66d488708" 969 + dependencies = [ 970 + "thiserror-impl", 971 + ] 972 + 973 + [[package]] 974 + name = "thiserror-impl" 975 + version = "2.0.12" 976 + source = "registry+https://github.com/rust-lang/crates.io-index" 977 + checksum = "7f7cf42b4507d8ea322120659672cf1b9dbb93f8f2d4ecfd6e51350ff5b17a1d" 978 + dependencies = [ 979 + "proc-macro2", 980 + "quote", 981 + "syn", 982 + ] 983 + 984 + [[package]] 985 + name = "tokenizers" 986 + version = "0.21.2" 987 + dependencies = [ 988 + "ahash", 989 + "aho-corasick", 990 + "compact_str", 991 + "dary_heap", 992 + "derive_builder", 993 + "esaxx-rs", 994 + "getrandom", 995 + "indicatif", 996 + "itertools", 997 + "log", 998 + "macro_rules_attribute", 999 + "monostate", 1000 + "onig", 1001 + "paste", 1002 + "rand", 1003 + "rayon", 1004 + "rayon-cond", 1005 + "regex", 1006 + "regex-syntax", 1007 + "serde", 1008 + "serde_json", 1009 + "spm_precompiled", 1010 + "thiserror", 1011 + "unicode-normalization-alignments", 1012 + "unicode-segmentation", 1013 + "unicode_categories", 1014 + ] 1015 + 1016 + [[package]] 1017 + name = "tokenizers-python" 1018 + version = "0.21.2" 1019 + dependencies = [ 1020 + "ahash", 1021 + "env_logger", 1022 + "itertools", 1023 + "libc", 1024 + "ndarray", 1025 + "numpy", 1026 + "pyo3", 1027 + "rayon", 1028 + "serde", 1029 + "serde_json", 1030 + "tempfile", 1031 + "tokenizers", 1032 + ] 1033 + 1034 + [[package]] 1035 + name = "unicode-ident" 1036 + version = "1.0.18" 1037 + source = "registry+https://github.com/rust-lang/crates.io-index" 1038 + checksum = "5a5f39404a5da50712a4c1eecf25e90dd62b613502b7e925fd4e4d19b5c96512" 1039 + 1040 + [[package]] 1041 + name = "unicode-normalization-alignments" 1042 + version = "0.1.12" 1043 + source = "registry+https://github.com/rust-lang/crates.io-index" 1044 + checksum = "43f613e4fa046e69818dd287fdc4bc78175ff20331479dab6e1b0f98d57062de" 1045 + dependencies = [ 1046 + "smallvec", 1047 + ] 1048 + 1049 + [[package]] 1050 + name = "unicode-segmentation" 1051 + version = "1.12.0" 1052 + source = "registry+https://github.com/rust-lang/crates.io-index" 1053 + checksum = "f6ccf251212114b54433ec949fd6a7841275f9ada20dddd2f29e9ceea4501493" 1054 + 1055 + [[package]] 1056 + name = "unicode-width" 1057 + version = "0.2.1" 1058 + source = "registry+https://github.com/rust-lang/crates.io-index" 1059 + checksum = "4a1a07cc7db3810833284e8d372ccdc6da29741639ecc70c9ec107df0fa6154c" 1060 + 1061 + [[package]] 1062 + name = "unicode_categories" 1063 + version = "0.1.1" 1064 + source = "registry+https://github.com/rust-lang/crates.io-index" 1065 + checksum = "39ec24b3121d976906ece63c9daad25b85969647682eee313cb5779fdd69e14e" 1066 + 1067 + [[package]] 1068 + name = "unindent" 1069 + version = "0.2.4" 1070 + source = "registry+https://github.com/rust-lang/crates.io-index" 1071 + checksum = "7264e107f553ccae879d21fbea1d6724ac785e8c3bfc762137959b5802826ef3" 1072 + 1073 + [[package]] 1074 + name = "utf8parse" 1075 + version = "0.2.2" 1076 + source = "registry+https://github.com/rust-lang/crates.io-index" 1077 + checksum = "06abde3611657adf66d383f00b093d7faecc7fa57071cce2578660c9f1010821" 1078 + 1079 + [[package]] 1080 + name = "version_check" 1081 + version = "0.9.5" 1082 + source = "registry+https://github.com/rust-lang/crates.io-index" 1083 + checksum = "0b928f33d975fc6ad9f86c8f283853ad26bdd5b10b7f1542aa2fa15e2289105a" 1084 + 1085 + [[package]] 1086 + name = "wasi" 1087 + version = "0.14.2+wasi-0.2.4" 1088 + source = "registry+https://github.com/rust-lang/crates.io-index" 1089 + checksum = "9683f9a5a998d873c0d21fcbe3c083009670149a8fab228644b8bd36b2c48cb3" 1090 + dependencies = [ 1091 + "wit-bindgen-rt", 1092 + ] 1093 + 1094 + [[package]] 1095 + name = "wasm-bindgen" 1096 + version = "0.2.100" 1097 + source = "registry+https://github.com/rust-lang/crates.io-index" 1098 + checksum = "1edc8929d7499fc4e8f0be2262a241556cfc54a0bea223790e71446f2aab1ef5" 1099 + dependencies = [ 1100 + "cfg-if", 1101 + "once_cell", 1102 + "wasm-bindgen-macro", 1103 + ] 1104 + 1105 + [[package]] 1106 + name = "wasm-bindgen-backend" 1107 + version = "0.2.100" 1108 + source = "registry+https://github.com/rust-lang/crates.io-index" 1109 + checksum = "2f0a0651a5c2bc21487bde11ee802ccaf4c51935d0d3d42a6101f98161700bc6" 1110 + dependencies = [ 1111 + "bumpalo", 1112 + "log", 1113 + "proc-macro2", 1114 + "quote", 1115 + "syn", 1116 + "wasm-bindgen-shared", 1117 + ] 1118 + 1119 + [[package]] 1120 + name = "wasm-bindgen-macro" 1121 + version = "0.2.100" 1122 + source = "registry+https://github.com/rust-lang/crates.io-index" 1123 + checksum = "7fe63fc6d09ed3792bd0897b314f53de8e16568c2b3f7982f468c0bf9bd0b407" 1124 + dependencies = [ 1125 + "quote", 1126 + "wasm-bindgen-macro-support", 1127 + ] 1128 + 1129 + [[package]] 1130 + name = "wasm-bindgen-macro-support" 1131 + version = "0.2.100" 1132 + source = "registry+https://github.com/rust-lang/crates.io-index" 1133 + checksum = "8ae87ea40c9f689fc23f209965b6fb8a99ad69aeeb0231408be24920604395de" 1134 + dependencies = [ 1135 + "proc-macro2", 1136 + "quote", 1137 + "syn", 1138 + "wasm-bindgen-backend", 1139 + "wasm-bindgen-shared", 1140 + ] 1141 + 1142 + [[package]] 1143 + name = "wasm-bindgen-shared" 1144 + version = "0.2.100" 1145 + source = "registry+https://github.com/rust-lang/crates.io-index" 1146 + checksum = "1a05d73b933a847d6cccdda8f838a22ff101ad9bf93e33684f39c1f5f0eece3d" 1147 + dependencies = [ 1148 + "unicode-ident", 1149 + ] 1150 + 1151 + [[package]] 1152 + name = "web-time" 1153 + version = "1.1.0" 1154 + source = "registry+https://github.com/rust-lang/crates.io-index" 1155 + checksum = "5a6580f308b1fad9207618087a65c04e7a10bc77e02c8e84e9b00dd4b12fa0bb" 1156 + dependencies = [ 1157 + "js-sys", 1158 + "wasm-bindgen", 1159 + ] 1160 + 1161 + [[package]] 1162 + name = "windows-sys" 1163 + version = "0.59.0" 1164 + source = "registry+https://github.com/rust-lang/crates.io-index" 1165 + checksum = "1e38bc4d79ed67fd075bcc251a1c39b32a1776bbe92e5bef1f0bf1f8c531853b" 1166 + dependencies = [ 1167 + "windows-targets 0.52.6", 1168 + ] 1169 + 1170 + [[package]] 1171 + name = "windows-sys" 1172 + version = "0.60.2" 1173 + source = "registry+https://github.com/rust-lang/crates.io-index" 1174 + checksum = "f2f500e4d28234f72040990ec9d39e3a6b950f9f22d3dba18416c35882612bcb" 1175 + dependencies = [ 1176 + "windows-targets 0.53.2", 1177 + ] 1178 + 1179 + [[package]] 1180 + name = "windows-targets" 1181 + version = "0.52.6" 1182 + source = "registry+https://github.com/rust-lang/crates.io-index" 1183 + checksum = "9b724f72796e036ab90c1021d4780d4d3d648aca59e491e6b98e725b84e99973" 1184 + dependencies = [ 1185 + "windows_aarch64_gnullvm 0.52.6", 1186 + "windows_aarch64_msvc 0.52.6", 1187 + "windows_i686_gnu 0.52.6", 1188 + "windows_i686_gnullvm 0.52.6", 1189 + "windows_i686_msvc 0.52.6", 1190 + "windows_x86_64_gnu 0.52.6", 1191 + "windows_x86_64_gnullvm 0.52.6", 1192 + "windows_x86_64_msvc 0.52.6", 1193 + ] 1194 + 1195 + [[package]] 1196 + name = "windows-targets" 1197 + version = "0.53.2" 1198 + source = "registry+https://github.com/rust-lang/crates.io-index" 1199 + checksum = "c66f69fcc9ce11da9966ddb31a40968cad001c5bedeb5c2b82ede4253ab48aef" 1200 + dependencies = [ 1201 + "windows_aarch64_gnullvm 0.53.0", 1202 + "windows_aarch64_msvc 0.53.0", 1203 + "windows_i686_gnu 0.53.0", 1204 + "windows_i686_gnullvm 0.53.0", 1205 + "windows_i686_msvc 0.53.0", 1206 + "windows_x86_64_gnu 0.53.0", 1207 + "windows_x86_64_gnullvm 0.53.0", 1208 + "windows_x86_64_msvc 0.53.0", 1209 + ] 1210 + 1211 + [[package]] 1212 + name = "windows_aarch64_gnullvm" 1213 + version = "0.52.6" 1214 + source = "registry+https://github.com/rust-lang/crates.io-index" 1215 + checksum = "32a4622180e7a0ec044bb555404c800bc9fd9ec262ec147edd5989ccd0c02cd3" 1216 + 1217 + [[package]] 1218 + name = "windows_aarch64_gnullvm" 1219 + version = "0.53.0" 1220 + source = "registry+https://github.com/rust-lang/crates.io-index" 1221 + checksum = "86b8d5f90ddd19cb4a147a5fa63ca848db3df085e25fee3cc10b39b6eebae764" 1222 + 1223 + [[package]] 1224 + name = "windows_aarch64_msvc" 1225 + version = "0.52.6" 1226 + source = "registry+https://github.com/rust-lang/crates.io-index" 1227 + checksum = "09ec2a7bb152e2252b53fa7803150007879548bc709c039df7627cabbd05d469" 1228 + 1229 + [[package]] 1230 + name = "windows_aarch64_msvc" 1231 + version = "0.53.0" 1232 + source = "registry+https://github.com/rust-lang/crates.io-index" 1233 + checksum = "c7651a1f62a11b8cbd5e0d42526e55f2c99886c77e007179efff86c2b137e66c" 1234 + 1235 + [[package]] 1236 + name = "windows_i686_gnu" 1237 + version = "0.52.6" 1238 + source = "registry+https://github.com/rust-lang/crates.io-index" 1239 + checksum = "8e9b5ad5ab802e97eb8e295ac6720e509ee4c243f69d781394014ebfe8bbfa0b" 1240 + 1241 + [[package]] 1242 + name = "windows_i686_gnu" 1243 + version = "0.53.0" 1244 + source = "registry+https://github.com/rust-lang/crates.io-index" 1245 + checksum = "c1dc67659d35f387f5f6c479dc4e28f1d4bb90ddd1a5d3da2e5d97b42d6272c3" 1246 + 1247 + [[package]] 1248 + name = "windows_i686_gnullvm" 1249 + version = "0.52.6" 1250 + source = "registry+https://github.com/rust-lang/crates.io-index" 1251 + checksum = "0eee52d38c090b3caa76c563b86c3a4bd71ef1a819287c19d586d7334ae8ed66" 1252 + 1253 + [[package]] 1254 + name = "windows_i686_gnullvm" 1255 + version = "0.53.0" 1256 + source = "registry+https://github.com/rust-lang/crates.io-index" 1257 + checksum = "9ce6ccbdedbf6d6354471319e781c0dfef054c81fbc7cf83f338a4296c0cae11" 1258 + 1259 + [[package]] 1260 + name = "windows_i686_msvc" 1261 + version = "0.52.6" 1262 + source = "registry+https://github.com/rust-lang/crates.io-index" 1263 + checksum = "240948bc05c5e7c6dabba28bf89d89ffce3e303022809e73deaefe4f6ec56c66" 1264 + 1265 + [[package]] 1266 + name = "windows_i686_msvc" 1267 + version = "0.53.0" 1268 + source = "registry+https://github.com/rust-lang/crates.io-index" 1269 + checksum = "581fee95406bb13382d2f65cd4a908ca7b1e4c2f1917f143ba16efe98a589b5d" 1270 + 1271 + [[package]] 1272 + name = "windows_x86_64_gnu" 1273 + version = "0.52.6" 1274 + source = "registry+https://github.com/rust-lang/crates.io-index" 1275 + checksum = "147a5c80aabfbf0c7d901cb5895d1de30ef2907eb21fbbab29ca94c5b08b1a78" 1276 + 1277 + [[package]] 1278 + name = "windows_x86_64_gnu" 1279 + version = "0.53.0" 1280 + source = "registry+https://github.com/rust-lang/crates.io-index" 1281 + checksum = "2e55b5ac9ea33f2fc1716d1742db15574fd6fc8dadc51caab1c16a3d3b4190ba" 1282 + 1283 + [[package]] 1284 + name = "windows_x86_64_gnullvm" 1285 + version = "0.52.6" 1286 + source = "registry+https://github.com/rust-lang/crates.io-index" 1287 + checksum = "24d5b23dc417412679681396f2b49f3de8c1473deb516bd34410872eff51ed0d" 1288 + 1289 + [[package]] 1290 + name = "windows_x86_64_gnullvm" 1291 + version = "0.53.0" 1292 + source = "registry+https://github.com/rust-lang/crates.io-index" 1293 + checksum = "0a6e035dd0599267ce1ee132e51c27dd29437f63325753051e71dd9e42406c57" 1294 + 1295 + [[package]] 1296 + name = "windows_x86_64_msvc" 1297 + version = "0.52.6" 1298 + source = "registry+https://github.com/rust-lang/crates.io-index" 1299 + checksum = "589f6da84c646204747d1270a2a5661ea66ed1cced2631d546fdfb155959f9ec" 1300 + 1301 + [[package]] 1302 + name = "windows_x86_64_msvc" 1303 + version = "0.53.0" 1304 + source = "registry+https://github.com/rust-lang/crates.io-index" 1305 + checksum = "271414315aff87387382ec3d271b52d7ae78726f5d44ac98b4f4030c91880486" 1306 + 1307 + [[package]] 1308 + name = "wit-bindgen-rt" 1309 + version = "0.39.0" 1310 + source = "registry+https://github.com/rust-lang/crates.io-index" 1311 + checksum = "6f42320e61fe2cfd34354ecb597f86f413484a798ba44a8ca1165c58d42da6c1" 1312 + dependencies = [ 1313 + "bitflags", 1314 + ] 1315 + 1316 + [[package]] 1317 + name = "zerocopy" 1318 + version = "0.8.26" 1319 + source = "registry+https://github.com/rust-lang/crates.io-index" 1320 + checksum = "1039dd0d3c310cf05de012d8a39ff557cb0d23087fd44cad61df08fc31907a2f" 1321 + dependencies = [ 1322 + "zerocopy-derive", 1323 + ] 1324 + 1325 + [[package]] 1326 + name = "zerocopy-derive" 1327 + version = "0.8.26" 1328 + source = "registry+https://github.com/rust-lang/crates.io-index" 1329 + checksum = "9ecf5b4cc5364572d7f4c329661bcc82724222973f2cab6f050a4e5c22f75181" 1330 + dependencies = [ 1331 + "proc-macro2", 1332 + "quote", 1333 + "syn", 1334 + ]
+27 -16
pkgs/development/python-modules/tokenizers/default.nix
··· 47 47 url = "https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-uncased-vocab.txt"; 48 48 hash = "sha256-B+ztN1zsFE0nyQAkHz4zlHjeyVj5L928VR8pXJkgOKM="; 49 49 }; 50 + "tokenizer-llama3.json" = fetchurl { 51 + url = "https://huggingface.co/Narsil/llama-tokenizer/resolve/main/tokenizer.json"; 52 + hash = "sha256-eePlImNfMXEwCRO7QhRkqH3mIiGCoFcLmyzLoqlksrQ="; 53 + }; 50 54 "big.txt" = fetchurl { 51 55 url = "https://norvig.com/big.txt"; 52 56 hash = "sha256-+gZsfUDw8gGsQUTmUqpiQw5YprOAXscGUPZ42lgE6Hs="; ··· 71 75 in 72 76 buildPythonPackage rec { 73 77 pname = "tokenizers"; 74 - version = "0.21.1"; 78 + version = "0.21.2"; 75 79 pyproject = true; 76 80 77 81 src = fetchFromGitHub { 78 82 owner = "huggingface"; 79 83 repo = "tokenizers"; 80 84 tag = "v${version}"; 81 - hash = "sha256-3S7ZCaZnnwyNjoZ4Y/q3ngQE2MIm2iyCCjYAkdMVG2A="; 85 + hash = "sha256-HO7Zg/yLY6yxjOo5Jf6Iu2zCreCyv7IaNrWtBKrspqQ="; 82 86 }; 83 87 84 - # TestUnigram.test_continuing_prefix_trainer_mismatch fails with: 85 - # Exception: No such file or directory (os error 2) 86 - # Fix submitted upstream: https://github.com/huggingface/tokenizers/pull/1747 88 + # Cargo.lock shipped with 0.21.2 is invalid: 89 + # error: no matching package found 90 + # searched package name: `ahash` 91 + # perhaps you meant: wasi 92 + # location searched: directory source `/build/tokenizers-0.21.2-vendor` (which is replacing registry `crates-io`) 93 + # required by package `tokenizers-python v0.21.2 (/build/source/bindings/python)` 94 + # 95 + # Hence, I (@GaetanLepage) re-generated the lockfile and embedded it here for now. 96 + # TODO: Try to switch back to `rustPlatform.fetchCargoVendor` at the next release. 87 97 postPatch = '' 88 - substituteInPlace tests/bindings/test_trainers.py \ 89 - --replace-fail '"data/' '"tests/data/' 98 + ln -sf '${./Cargo.lock}' Cargo.lock 90 99 ''; 91 - 92 - cargoDeps = rustPlatform.fetchCargoVendor { 93 - inherit 94 - pname 95 - version 96 - src 97 - sourceRoot 98 - ; 99 - hash = "sha256-I7LlBmeVY2rWI0ta6x311iAurQKuutsClrbUgkt9xWk="; 100 + cargoDeps = rustPlatform.importCargoLock { 101 + lockFile = ./Cargo.lock; 100 102 }; 103 + # cargoDeps = rustPlatform.fetchCargoVendor { 104 + # inherit 105 + # pname 106 + # version 107 + # src 108 + # sourceRoot 109 + # ; 110 + # hash = "sha256-EKiHjcXUjU8CWe2CB2EgAQlRcZebwe4EpD7P8lWbCjw="; 111 + # }; 101 112 102 113 sourceRoot = "${src.name}/bindings/python"; 103 114