Date: Sun, 6 Aug 2023 15:36:39 GMT From: "Jason W. Bacon" <jwb@FreeBSD.org> To: ports-committers@FreeBSD.org, dev-commits-ports-all@FreeBSD.org, dev-commits-ports-main@FreeBSD.org Subject: git: 21e0240b8fe6 - main - devel/py-cykhash: Cython equivalent to khash-sets/maps Message-ID: <202308061536.376FadeH042904@gitrepo.freebsd.org>
next in thread | raw e-mail | index | archive | help
The branch main has been updated by jwb: URL: https://cgit.FreeBSD.org/ports/commit/?id=21e0240b8fe6a1552665bdc2e52c1b198e4df49c commit 21e0240b8fe6a1552665bdc2e52c1b198e4df49c Author: Jason W. Bacon <jwb@FreeBSD.org> AuthorDate: 2023-08-06 15:34:43 +0000 Commit: Jason W. Bacon <jwb@FreeBSD.org> CommitDate: 2023-08-06 15:36:37 +0000 devel/py-cykhash: Cython equivalent to khash-sets/maps Cykhash is a cython equivalent to khash-sets/maps, efficient implementation of isin and unique Benefits: Brings functionality of khash to Python and Cython and can be used seamlessly in numpy or pandas. Numpy's world is lacking the concept of a (hash-)set. This shortcoming is fixed and efficient (memory- and speedwise compared to pandas') unique and isin are implemented. Python-set/dict have big memory-footprint. For some datatypes the overhead can be reduced by using khash by factor 4-8. --- devel/py-cykhash/Makefile | 20 ++++++++++++++++++++ devel/py-cykhash/distinfo | 3 +++ devel/py-cykhash/pkg-descr | 14 ++++++++++++++ 3 files changed, 37 insertions(+) diff --git a/devel/py-cykhash/Makefile b/devel/py-cykhash/Makefile new file mode 100644 index 000000000000..765dcbde2035 --- /dev/null +++ b/devel/py-cykhash/Makefile @@ -0,0 +1,20 @@ +PORTNAME= cykhash +DISTVERSION= 2.0.1 +CATEGORIES= devel python +MASTER_SITES= CHEESESHOP +PKGNAMEPREFIX= ${PYTHON_PKGNAMEPREFIX} + +MAINTAINER= jwb@FreeBSD.org +COMMENT= Cython equivalent to khash-sets/maps +WWW= https://pypi.python.org/project/cykhash/ + +LICENSE= MIT +LICENSE_FILE= ${WRKSRC}/LICENSE + +USES= python +USE_PYTHON= autoplist cython distutils + +post-stage: + @${STRIP_CMD} ${STAGEDIR}${PYTHON_SITELIBDIR}/cykhash/*.so + +.include <bsd.port.mk> diff --git a/devel/py-cykhash/distinfo b/devel/py-cykhash/distinfo new file mode 100644 index 000000000000..07e7db282c6b --- /dev/null +++ b/devel/py-cykhash/distinfo @@ -0,0 +1,3 @@ +TIMESTAMP = 1691328170 +SHA256 (cykhash-2.0.1.tar.gz) = b4794bc9f549114d8cf1d856d9f64e08ff5f246bf043cf369fdb414e9ceb97f7 +SIZE (cykhash-2.0.1.tar.gz) = 44895 diff --git a/devel/py-cykhash/pkg-descr b/devel/py-cykhash/pkg-descr new file mode 100644 index 000000000000..dc8460a45675 --- /dev/null +++ b/devel/py-cykhash/pkg-descr @@ -0,0 +1,14 @@ +Cykhash is a cython equivalent to khash-sets/maps, efficient +implementation of isin and unique + +Benefits: + + Brings functionality of khash to Python and Cython and can be used + seamlessly in numpy or pandas. + + Numpy's world is lacking the concept of a (hash-)set. This + shortcoming is fixed and efficient (memory- and speedwise compared + to pandas') unique and isin are implemented. + + Python-set/dict have big memory-footprint. For some datatypes the + overhead can be reduced by using khash by factor 4-8.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?202308061536.376FadeH042904>