diff options
Diffstat (limited to 'nixpkgs/pkgs/development/python-modules/tokenizers/default.nix')
| -rw-r--r-- | nixpkgs/pkgs/development/python-modules/tokenizers/default.nix | 88 |
1 files changed, 41 insertions, 47 deletions
diff --git a/nixpkgs/pkgs/development/python-modules/tokenizers/default.nix b/nixpkgs/pkgs/development/python-modules/tokenizers/default.nix index d8a731825feb..a008e15b3704 100644 --- a/nixpkgs/pkgs/development/python-modules/tokenizers/default.nix +++ b/nixpkgs/pkgs/development/python-modules/tokenizers/default.nix @@ -1,5 +1,6 @@ { lib , stdenv +, linkFarm , buildPythonPackage , cargo , datasets @@ -21,46 +22,48 @@ let # See https://github.com/huggingface/tokenizers/blob/main/bindings/python/tests/utils.py for details # about URLs and file names - robertaVocab = fetchurl { - url = "https://s3.amazonaws.com/models.huggingface.co/bert/roberta-base-vocab.json"; - sha256 = "0m86wpkfb2gdh9x9i9ng2fvwk1rva4p0s98xw996nrjxs7166zwy"; - }; - robertaMerges = fetchurl { - url = "https://s3.amazonaws.com/models.huggingface.co/bert/roberta-base-merges.txt"; - sha256 = "1idd4rvkpqqbks51i2vjbd928inw7slij9l4r063w3y5fd3ndq8w"; - }; - albertVocab = fetchurl { - url = "https://s3.amazonaws.com/models.huggingface.co/bert/albert-base-v1-tokenizer.json"; - sha256 = "1hra9pn8rczx7378z88zjclw2qsdrdwq20m56sy42s2crbas6akf"; - }; - bertVocab = fetchurl { - url = "https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-uncased-vocab.txt"; - sha256 = "18rq42cmqa8zanydsbzrb34xwy4l6cz1y900r4kls57cbhvyvv07"; - }; - norvigBig = fetchurl { - url = "https://norvig.com/big.txt"; - sha256 = "0yz80icdly7na03cfpl0nfk5h3j3cam55rj486n03wph81ynq1ps"; - }; - docPipelineTokenizer = fetchurl { - url = "https://s3.amazonaws.com/models.huggingface.co/bert/anthony/doc-pipeline/tokenizer.json"; - hash = "sha256-i533xC8J5CDMNxBjo+p6avIM8UOcui8RmGAmK0GmfBc="; - }; - docQuicktourTokenizer = fetchurl { - url = "https://s3.amazonaws.com/models.huggingface.co/bert/anthony/doc-quicktour/tokenizer.json"; - hash = "sha256-ipY9d5DR5nxoO6kj7rItueZ9AO5wq9+Nzr6GuEIfIBI="; - }; - openaiVocab = fetchurl { - url = "https://s3.amazonaws.com/models.huggingface.co/bert/openai-gpt-vocab.json"; - sha256 = "0y40gc9bixj5rxv674br1rxmxkd3ly29p80x1596h8yywwcrpx7x"; - }; - openaiMerges = fetchurl { - url = "https://s3.amazonaws.com/models.huggingface.co/bert/openai-gpt-merges.txt"; - sha256 = "09a754pm4djjglv3x5pkgwd6f79i2rq8ydg0f7c3q1wmwqdbba8f"; + test-data = linkFarm "tokenizers-test-data" { + "roberta-base-vocab.json" = fetchurl { + url = "https://s3.amazonaws.com/models.huggingface.co/bert/roberta-base-vocab.json"; + sha256 = "0m86wpkfb2gdh9x9i9ng2fvwk1rva4p0s98xw996nrjxs7166zwy"; + }; + "roberta-base-merges.txt" = fetchurl { + url = "https://s3.amazonaws.com/models.huggingface.co/bert/roberta-base-merges.txt"; + sha256 = "1idd4rvkpqqbks51i2vjbd928inw7slij9l4r063w3y5fd3ndq8w"; + }; + "albert-base-v1-tokenizer.json" = fetchurl { + url = "https://s3.amazonaws.com/models.huggingface.co/bert/albert-base-v1-tokenizer.json"; + sha256 = "1hra9pn8rczx7378z88zjclw2qsdrdwq20m56sy42s2crbas6akf"; + }; + "bert-base-uncased-vocab.txt" = fetchurl { + url = "https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-uncased-vocab.txt"; + sha256 = "18rq42cmqa8zanydsbzrb34xwy4l6cz1y900r4kls57cbhvyvv07"; + }; + "big.txt" = fetchurl { + url = "https://norvig.com/big.txt"; + sha256 = "0yz80icdly7na03cfpl0nfk5h3j3cam55rj486n03wph81ynq1ps"; + }; + "bert-wiki.json" = fetchurl { + url = "https://s3.amazonaws.com/models.huggingface.co/bert/anthony/doc-pipeline/tokenizer.json"; + hash = "sha256-i533xC8J5CDMNxBjo+p6avIM8UOcui8RmGAmK0GmfBc="; + }; + "tokenizer-wiki.json" = fetchurl { + url = "https://s3.amazonaws.com/models.huggingface.co/bert/anthony/doc-quicktour/tokenizer.json"; + hash = "sha256-ipY9d5DR5nxoO6kj7rItueZ9AO5wq9+Nzr6GuEIfIBI="; + }; + "openai-gpt-vocab.json" = fetchurl { + url = "https://s3.amazonaws.com/models.huggingface.co/bert/openai-gpt-vocab.json"; + sha256 = "0y40gc9bixj5rxv674br1rxmxkd3ly29p80x1596h8yywwcrpx7x"; + }; + "openai-gpt-merges.txt" = fetchurl { + url = "https://s3.amazonaws.com/models.huggingface.co/bert/openai-gpt-merges.txt"; + sha256 = "09a754pm4djjglv3x5pkgwd6f79i2rq8ydg0f7c3q1wmwqdbba8f"; + }; }; in buildPythonPackage rec { pname = "tokenizers"; - version = "0.14.1"; + version = "0.15.0"; format = "pyproject"; disabled = pythonOlder "3.7"; @@ -69,7 +72,7 @@ buildPythonPackage rec { owner = "huggingface"; repo = pname; rev = "v${version}"; - hash = "sha256-cq7dQLttNkV5UUhXujxKKMuzhD7hz+zTTKxUKlvz1s0="; + hash = "sha256-+yfX12eKtgZV1OQvPOlMVTONbpFuigHcl4SjoCIZkSk="; }; cargoDeps = rustPlatform.importCargoLock { @@ -107,16 +110,7 @@ buildPythonPackage rec { postUnpack = '' # Add data files for tests, otherwise tests attempt network access mkdir $sourceRoot/tests/data - ( cd $sourceRoot/tests/data - ln -s ${robertaVocab} roberta-base-vocab.json - ln -s ${robertaMerges} roberta-base-merges.txt - ln -s ${albertVocab} albert-base-v1-tokenizer.json - ln -s ${bertVocab} bert-base-uncased-vocab.txt - ln -s ${docPipelineTokenizer} bert-wiki.json - ln -s ${docQuicktourTokenizer} tokenizer-wiki.json - ln -s ${norvigBig} big.txt - ln -s ${openaiVocab} openai-gpt-vocab.json - ln -s ${openaiMerges} openai-gpt-merges.txt ) + ln -s ${test-data}/* $sourceRoot/tests/data/ ''; preCheck = '' |