From 6c34be9ab1b27410b400e03f8281f03d4f956fa3 Mon Sep 17 00:00:00 2001 From: Po-Chuan Hsieh Date: Tue, 22 Aug 2023 00:52:01 +0800 Subject: [PATCH] textproc/py-tiktoken: Add py-tiktoken 0.4.0 tiktoken is a fast BPE tokeniser for use with OpenAI's models. --- textproc/Makefile | 1 + textproc/py-tiktoken/Makefile | 35 +++++++++++ textproc/py-tiktoken/Makefile.crates | 43 ++++++++++++++ textproc/py-tiktoken/distinfo | 89 ++++++++++++++++++++++++++++ textproc/py-tiktoken/pkg-descr | 1 + 5 files changed, 169 insertions(+) create mode 100644 textproc/py-tiktoken/Makefile create mode 100644 textproc/py-tiktoken/Makefile.crates create mode 100644 textproc/py-tiktoken/distinfo create mode 100644 textproc/py-tiktoken/pkg-descr diff --git a/textproc/Makefile b/textproc/Makefile index 126df16034c9..edfee41d6241 100644 --- a/textproc/Makefile +++ b/textproc/Makefile @@ -1593,6 +1593,7 @@ SUBDIR += py-texttable SUBDIR += py-textual SUBDIR += py-three-merge + SUBDIR += py-tiktoken SUBDIR += py-tinycss SUBDIR += py-tinycss2 SUBDIR += py-toml diff --git a/textproc/py-tiktoken/Makefile b/textproc/py-tiktoken/Makefile new file mode 100644 index 000000000000..ec107bbec15d --- /dev/null +++ b/textproc/py-tiktoken/Makefile @@ -0,0 +1,35 @@ +PORTNAME= tiktoken +PORTVERSION= 0.4.0 +CATEGORIES= textproc python +MASTER_SITES= PYPI +PKGNAMEPREFIX= ${PYTHON_PKGNAMEPREFIX} +DISTFILES= ${DISTNAME}${EXTRACT_SUFX} + +MAINTAINER= sunpoet@FreeBSD.org +COMMENT= Fast BPE tokeniser for use with OpenAI's models +WWW= https://github.com/openai/tiktoken + +LICENSE= MIT +LICENSE_FILE= ${WRKSRC}/LICENSE + +BUILD_DEPENDS= ${PYTHON_PKGNAMEPREFIX}setuptools>=62.4:devel/py-setuptools@${PY_FLAVOR} \ + ${PYTHON_PKGNAMEPREFIX}setuptools-rust>=1.5.2:devel/py-setuptools-rust@${PY_FLAVOR} \ + ${PYTHON_PKGNAMEPREFIX}wheel>=0:devel/py-wheel@${PY_FLAVOR} +RUN_DEPENDS= ${PYTHON_PKGNAMEPREFIX}regex>=2022.1.18:textproc/py-regex@${PY_FLAVOR} \ + ${PYTHON_PKGNAMEPREFIX}requests>=2.26.0:www/py-requests@${PY_FLAVOR} + +USES= cargo python +USE_PYTHON= autoplist concurrent pep517 + +CARGO_BUILD= no +CARGO_INSTALL= no + +OPTIONS_DEFINE= BLOBFILE +BLOBFILE_DESC= Blob file support + +BLOBFILE_RUN_DEPENDS= ${PYTHON_PKGNAMEPREFIX}blobfile>=2:net/py-blobfile@${PY_FLAVOR} + +post-install: + ${FIND} ${STAGEDIR}${PYTHON_SITELIBDIR} -name '*.so' -exec ${STRIP_CMD} {} + + +.include diff --git a/textproc/py-tiktoken/Makefile.crates b/textproc/py-tiktoken/Makefile.crates new file mode 100644 index 000000000000..1eaa2bb56a75 --- /dev/null +++ b/textproc/py-tiktoken/Makefile.crates @@ -0,0 +1,43 @@ +CARGO_CRATES= aho-corasick-1.0.4 \ + autocfg-1.1.0 \ + bit-set-0.5.3 \ + bit-vec-0.6.3 \ + bitflags-1.3.2 \ + bstr-1.6.0 \ + cfg-if-1.0.0 \ + fancy-regex-0.10.0 \ + indoc-1.0.9 \ + libc-0.2.147 \ + lock_api-0.4.10 \ + memchr-2.5.0 \ + memoffset-0.6.5 \ + once_cell-1.18.0 \ + parking_lot-0.12.1 \ + parking_lot_core-0.9.8 \ + proc-macro2-1.0.66 \ + pyo3-0.17.3 \ + pyo3-build-config-0.17.3 \ + pyo3-ffi-0.17.3 \ + pyo3-macros-0.17.3 \ + pyo3-macros-backend-0.17.3 \ + quote-1.0.33 \ + redox_syscall-0.3.5 \ + regex-1.9.3 \ + regex-automata-0.3.6 \ + regex-syntax-0.7.4 \ + rustc-hash-1.1.0 \ + scopeguard-1.2.0 \ + serde-1.0.183 \ + smallvec-1.11.0 \ + syn-1.0.109 \ + target-lexicon-0.12.11 \ + unicode-ident-1.0.11 \ + unindent-0.1.11 \ + windows-targets-0.48.5 \ + windows_aarch64_gnullvm-0.48.5 \ + windows_aarch64_msvc-0.48.5 \ + windows_i686_gnu-0.48.5 \ + windows_i686_msvc-0.48.5 \ + windows_x86_64_gnu-0.48.5 \ + windows_x86_64_gnullvm-0.48.5 \ + windows_x86_64_msvc-0.48.5 diff --git a/textproc/py-tiktoken/distinfo b/textproc/py-tiktoken/distinfo new file mode 100644 index 000000000000..568934741e8f --- /dev/null +++ b/textproc/py-tiktoken/distinfo @@ -0,0 +1,89 @@ +TIMESTAMP = 1692211619 +SHA256 (tiktoken-0.4.0.tar.gz) = 59b20a819969735b48161ced9b92f05dc4519c17be4015cfb73b65270a243620 +SIZE (tiktoken-0.4.0.tar.gz) = 25504 +SHA256 (rust/crates/aho-corasick-1.0.4.crate) = 6748e8def348ed4d14996fa801f4122cd763fff530258cdc03f64b25f89d3a5a +SIZE (rust/crates/aho-corasick-1.0.4.crate) = 171175 +SHA256 (rust/crates/autocfg-1.1.0.crate) = d468802bab17cbc0cc575e9b053f41e72aa36bfa6b7f55e3529ffa43161b97fa +SIZE (rust/crates/autocfg-1.1.0.crate) = 13272 +SHA256 (rust/crates/bit-set-0.5.3.crate) = 0700ddab506f33b20a03b13996eccd309a48e5ff77d0d95926aa0210fb4e95f1 +SIZE (rust/crates/bit-set-0.5.3.crate) = 14470 +SHA256 (rust/crates/bit-vec-0.6.3.crate) = 349f9b6a179ed607305526ca489b34ad0a41aed5f7980fa90eb03160b69598fb +SIZE (rust/crates/bit-vec-0.6.3.crate) = 19927 +SHA256 (rust/crates/bitflags-1.3.2.crate) = bef38d45163c2f1dde094a7dfd33ccf595c92905c8f8f4fdc18d06fb1037718a +SIZE (rust/crates/bitflags-1.3.2.crate) = 23021 +SHA256 (rust/crates/bstr-1.6.0.crate) = 6798148dccfbff0fae41c7574d2fa8f1ef3492fba0face179de5d8d447d67b05 +SIZE (rust/crates/bstr-1.6.0.crate) = 385877 +SHA256 (rust/crates/cfg-if-1.0.0.crate) = baf1de4339761588bc0619e3cbc0120ee582ebb74b53b4efbf79117bd2da40fd +SIZE (rust/crates/cfg-if-1.0.0.crate) = 7934 +SHA256 (rust/crates/fancy-regex-0.10.0.crate) = 0678ab2d46fa5195aaf59ad034c083d351377d4af57f3e073c074d0da3e3c766 +SIZE (rust/crates/fancy-regex-0.10.0.crate) = 79951 +SHA256 (rust/crates/indoc-1.0.9.crate) = bfa799dd5ed20a7e349f3b4639aa80d74549c81716d9ec4f994c9b5815598306 +SIZE (rust/crates/indoc-1.0.9.crate) = 13475 +SHA256 (rust/crates/libc-0.2.147.crate) = b4668fb0ea861c1df094127ac5f1da3409a82116a4ba74fca2e58ef927159bb3 +SIZE (rust/crates/libc-0.2.147.crate) = 686772 +SHA256 (rust/crates/lock_api-0.4.10.crate) = c1cc9717a20b1bb222f333e6a92fd32f7d8a18ddc5a3191a11af45dcbf4dcd16 +SIZE (rust/crates/lock_api-0.4.10.crate) = 26713 +SHA256 (rust/crates/memchr-2.5.0.crate) = 2dffe52ecf27772e601905b7522cb4ef790d2cc203488bbd0e2fe85fcb74566d +SIZE (rust/crates/memchr-2.5.0.crate) = 65812 +SHA256 (rust/crates/memoffset-0.6.5.crate) = 5aa361d4faea93603064a027415f07bd8e1d5c88c9fbf68bf56a285428fd79ce +SIZE (rust/crates/memoffset-0.6.5.crate) = 7686 +SHA256 (rust/crates/once_cell-1.18.0.crate) = dd8b5dd2ae5ed71462c540258bedcb51965123ad7e7ccf4b9a8cafaa4a63576d +SIZE (rust/crates/once_cell-1.18.0.crate) = 32969 +SHA256 (rust/crates/parking_lot-0.12.1.crate) = 3742b2c103b9f06bc9fff0a37ff4912935851bee6d36f3c02bcc755bcfec228f +SIZE (rust/crates/parking_lot-0.12.1.crate) = 40967 +SHA256 (rust/crates/parking_lot_core-0.9.8.crate) = 93f00c865fe7cabf650081affecd3871070f26767e7b2070a3ffae14c654b447 +SIZE (rust/crates/parking_lot_core-0.9.8.crate) = 32383 +SHA256 (rust/crates/proc-macro2-1.0.66.crate) = 18fb31db3f9bddb2ea821cde30a9f70117e3f119938b5ee630b7403aa6e2ead9 +SIZE (rust/crates/proc-macro2-1.0.66.crate) = 43575 +SHA256 (rust/crates/pyo3-0.17.3.crate) = 268be0c73583c183f2b14052337465768c07726936a260f480f0857cb95ba543 +SIZE (rust/crates/pyo3-0.17.3.crate) = 408939 +SHA256 (rust/crates/pyo3-build-config-0.17.3.crate) = 28fcd1e73f06ec85bf3280c48c67e731d8290ad3d730f8be9dc07946923005c8 +SIZE (rust/crates/pyo3-build-config-0.17.3.crate) = 29230 +SHA256 (rust/crates/pyo3-ffi-0.17.3.crate) = 0f6cb136e222e49115b3c51c32792886defbfb0adead26a688142b346a0b9ffc +SIZE (rust/crates/pyo3-ffi-0.17.3.crate) = 62854 +SHA256 (rust/crates/pyo3-macros-0.17.3.crate) = 94144a1266e236b1c932682136dc35a9dee8d3589728f68130c7c3861ef96b28 +SIZE (rust/crates/pyo3-macros-0.17.3.crate) = 7257 +SHA256 (rust/crates/pyo3-macros-backend-0.17.3.crate) = c8df9be978a2d2f0cdebabb03206ed73b11314701a5bfe71b0d753b81997777f +SIZE (rust/crates/pyo3-macros-backend-0.17.3.crate) = 49992 +SHA256 (rust/crates/quote-1.0.33.crate) = 5267fca4496028628a95160fc423a33e8b2e6af8a5302579e322e4b520293cae +SIZE (rust/crates/quote-1.0.33.crate) = 28090 +SHA256 (rust/crates/redox_syscall-0.3.5.crate) = 567664f262709473930a4bf9e51bf2ebf3348f2e748ccc50dea20646858f8f29 +SIZE (rust/crates/redox_syscall-0.3.5.crate) = 23404 +SHA256 (rust/crates/regex-1.9.3.crate) = 81bc1d4caf89fac26a70747fe603c130093b53c773888797a6329091246d651a +SIZE (rust/crates/regex-1.9.3.crate) = 252676 +SHA256 (rust/crates/regex-automata-0.3.6.crate) = fed1ceff11a1dddaee50c9dc8e4938bd106e9d89ae372f192311e7da498e3b69 +SIZE (rust/crates/regex-automata-0.3.6.crate) = 607174 +SHA256 (rust/crates/regex-syntax-0.7.4.crate) = e5ea92a5b6195c6ef2a0295ea818b312502c6fc94dde986c5553242e18fd4ce2 +SIZE (rust/crates/regex-syntax-0.7.4.crate) = 343365 +SHA256 (rust/crates/rustc-hash-1.1.0.crate) = 08d43f7aa6b08d49f382cde6a7982047c3426db949b1424bc4b7ec9ae12c6ce2 +SIZE (rust/crates/rustc-hash-1.1.0.crate) = 9331 +SHA256 (rust/crates/scopeguard-1.2.0.crate) = 94143f37725109f92c262ed2cf5e59bce7498c01bcc1502d7b9afe439a4e9f49 +SIZE (rust/crates/scopeguard-1.2.0.crate) = 11619 +SHA256 (rust/crates/serde-1.0.183.crate) = 32ac8da02677876d532745a130fc9d8e6edfa81a269b107c5b00829b91d8eb3c +SIZE (rust/crates/serde-1.0.183.crate) = 77210 +SHA256 (rust/crates/smallvec-1.11.0.crate) = 62bb4feee49fdd9f707ef802e22365a35de4b7b299de4763d44bfea899442ff9 +SIZE (rust/crates/smallvec-1.11.0.crate) = 34680 +SHA256 (rust/crates/syn-1.0.109.crate) = 72b64191b275b66ffe2469e8af2c1cfe3bafa67b529ead792a6d0160888b4237 +SIZE (rust/crates/syn-1.0.109.crate) = 237611 +SHA256 (rust/crates/target-lexicon-0.12.11.crate) = 9d0e916b1148c8e263850e1ebcbd046f333e0683c724876bb0da63ea4373dc8a +SIZE (rust/crates/target-lexicon-0.12.11.crate) = 24594 +SHA256 (rust/crates/unicode-ident-1.0.11.crate) = 301abaae475aa91687eb82514b328ab47a211a533026cb25fc3e519b86adfc3c +SIZE (rust/crates/unicode-ident-1.0.11.crate) = 42067 +SHA256 (rust/crates/unindent-0.1.11.crate) = e1766d682d402817b5ac4490b3c3002d91dfa0d22812f341609f97b08757359c +SIZE (rust/crates/unindent-0.1.11.crate) = 7700 +SHA256 (rust/crates/windows-targets-0.48.5.crate) = 9a2fa6e2155d7247be68c096456083145c183cbbbc2764150dda45a87197940c +SIZE (rust/crates/windows-targets-0.48.5.crate) = 6904 +SHA256 (rust/crates/windows_aarch64_gnullvm-0.48.5.crate) = 2b38e32f0abccf9987a4e3079dfb67dcd799fb61361e53e2882c3cbaf0d905d8 +SIZE (rust/crates/windows_aarch64_gnullvm-0.48.5.crate) = 418492 +SHA256 (rust/crates/windows_aarch64_msvc-0.48.5.crate) = dc35310971f3b2dbbf3f0690a219f40e2d9afcf64f9ab7cc1be722937c26b4bc +SIZE (rust/crates/windows_aarch64_msvc-0.48.5.crate) = 798483 +SHA256 (rust/crates/windows_i686_gnu-0.48.5.crate) = a75915e7def60c94dcef72200b9a8e58e5091744960da64ec734a6c6e9b3743e +SIZE (rust/crates/windows_i686_gnu-0.48.5.crate) = 844891 +SHA256 (rust/crates/windows_i686_msvc-0.48.5.crate) = 8f55c233f70c4b27f66c523580f78f1004e8b5a8b659e05a4eb49d4166cca406 +SIZE (rust/crates/windows_i686_msvc-0.48.5.crate) = 864300 +SHA256 (rust/crates/windows_x86_64_gnu-0.48.5.crate) = 53d40abd2583d23e4718fddf1ebec84dbff8381c07cae67ff7768bbf19c6718e +SIZE (rust/crates/windows_x86_64_gnu-0.48.5.crate) = 801619 +SHA256 (rust/crates/windows_x86_64_gnullvm-0.48.5.crate) = 0b7b52767868a23d5bab768e390dc5f5c55825b6d30b86c844ff2dc7414044cc +SIZE (rust/crates/windows_x86_64_gnullvm-0.48.5.crate) = 418486 +SHA256 (rust/crates/windows_x86_64_msvc-0.48.5.crate) = ed94fce61571a4006852b7389a063ab983c02eb1bb37b47f8272ce92d06d9538 +SIZE (rust/crates/windows_x86_64_msvc-0.48.5.crate) = 798412 diff --git a/textproc/py-tiktoken/pkg-descr b/textproc/py-tiktoken/pkg-descr new file mode 100644 index 000000000000..190f0b91352e --- /dev/null +++ b/textproc/py-tiktoken/pkg-descr @@ -0,0 +1 @@ +tiktoken is a fast BPE tokeniser for use with OpenAI's models.