misc/py-jiwer: New port: Evaluate speech-to-text system with similarity measures

This commit is contained in:
Yuri Victorovich 2024-08-05 17:12:46 -07:00
parent d5bf0a77eb
commit 219e72a7c3
4 changed files with 45 additions and 0 deletions

View file

@ -439,6 +439,7 @@
SUBDIR += py-icoextract
SUBDIR += py-instructor
SUBDIR += py-ipyfastscape
SUBDIR += py-jiwer
SUBDIR += py-kartograph
SUBDIR += py-laspy
SUBDIR += py-lazrs

29
misc/py-jiwer/Makefile Normal file
View file

@ -0,0 +1,29 @@
PORTNAME= jiwer
#DISTVERSIONPREFIX= v
DISTVERSION= 3.0.4 # see https://github.com/jitsi/jiwer/issues/91
CATEGORIES= misc python # machine-learning
MASTER_SITES= PYPI # no tests
PKGNAMEPREFIX= ${PYTHON_PKGNAMEPREFIX}
MAINTAINER= yuri@FreeBSD.org
COMMENT= Evaluate speech-to-text system with similarity measures
WWW= https://github.com/jitsi/jiwer
LICENSE= APACHE20
LICENSE_FILE= ${WRKSRC}/LICENSE
BUILD_DEPENDS= ${PYTHON_PKGNAMEPREFIX}poetry-core>0:devel/py-poetry-core@${PY_FLAVOR}
RUN_DEPENDS= ${PYTHON_PKGNAMEPREFIX}click>=8.1.3:devel/py-click@${PY_FLAVOR} \
${PYTHON_PKGNAMEPREFIX}rapidfuzz>=3:devel/py-rapidfuzz@${PY_FLAVOR}
USES= python
USE_PYTHON= pep517 concurrent autoplist #pytest
#USE_GITHUB= yes
#GH_ACCOUNT= jitsi
TEST_ENV= ${MAKE_ENV} PYTHONPATH=${STAGEDIR}${PYTHONPREFIX_SITELIBDIR}
NO_ARCH= yes
.include <bsd.port.mk>

3
misc/py-jiwer/distinfo Normal file
View file

@ -0,0 +1,3 @@
TIMESTAMP = 1722887112
SHA256 (jiwer-3.0.4.tar.gz) = 2438acdc7ca22128fcab4be60db595809d2b5e73785b736de36dc3281a2a6ae8
SIZE (jiwer-3.0.4.tar.gz) = 17515

12
misc/py-jiwer/pkg-descr Normal file
View file

@ -0,0 +1,12 @@
JiWER is a simple and fast python package to evaluate an automatic speech
recognition system. It supports the following measures:
* word error rate (WER)
* match error rate (MER)
* word information lost (WIL)
* word information preserved (WIP)
* character error rate (CER)
These measures are computed with the use of the minimum-edit distance between
one or more reference and hypothesis sentences. The minimum-edit distance is
calculated using RapidFuzz, which uses C++ under the hood, and is therefore
faster than a pure python implementation.