3a00ffa77e
part). These patches, released under a BSD license, seem to improve the accuracy of language detection, especially those that don't have a Latin script.
45 lines
1.3 KiB
Makefile
45 lines
1.3 KiB
Makefile
# New ports collection makefile for: libtextcat
|
|
# Date created: Sat 18 nov 2007
|
|
# Whom: thierry@pompo.net
|
|
#
|
|
# $FreeBSD$
|
|
#
|
|
|
|
PORTNAME= libtextcat
|
|
PORTVERSION= 2.2
|
|
PORTREVISION= 1
|
|
CATEGORIES= textproc
|
|
MASTER_SITES= http://software.wise-guys.nl/download/
|
|
|
|
#PATCH_SITES= http://cvs.fedora.redhat.com/viewcvs/*checkout*/rpms/libtextcat/devel/
|
|
#PATCHFILES= ${PORTNAME}-${PORTVERSION}-OOo.patch
|
|
#PATCH_DIST_STRIP= -p1
|
|
|
|
MAINTAINER= thierry@FreeBSD.org
|
|
COMMENT= Language guessing by N-Gram-Based Text Categorization
|
|
|
|
USE_AUTOTOOLS= libtool:15
|
|
GNU_CONFIGURE= yes
|
|
CONFIGURE_TARGET= --build=${ARCH}-portbld-freebsd${OSREL}
|
|
USE_LDCONFIG= yes
|
|
|
|
PORTDOCS= LICENSE README TODO
|
|
|
|
post-install:
|
|
${INSTALL_DATA} ${WRKSRC}/src/textcat.h ${PREFIX}/include/
|
|
${MKDIR} ${DATADIR}/LM
|
|
@${ECHO_MSG} "Installing language models provided in Gertjan van Noord's TextCat package"
|
|
(cd ${WRKSRC}/langclass/LM && \
|
|
${FIND} . -name "*.lm" -exec ${INSTALL_DATA} "{}" "${DATADIR}/LM/{}" \;)
|
|
${INSTALL_DATA} ${WRKSRC}/langclass/conf.txt "${DATADIR}"
|
|
${MKDIR} ${DOCSDIR}
|
|
${INSTALL_DATA} ${PORTDOCS:S|^|${WRKSRC}/|} ${DOCSDIR}
|
|
|
|
regression-test:
|
|
(cd ${WRKSRC}/langclass/ && \
|
|
for t in `${LS} ShortTexts/*.txt` ; do \
|
|
${ECHO_MSG} "Analyzing $$t..." ; \
|
|
../src/testtextcat conf.txt < $$t ; \
|
|
done)
|
|
|
|
.include <bsd.port.mk>
|