From d95f76d47a05e324153cef13dfaac2f4dfc2d3b8 Mon Sep 17 00:00:00 2001 From: Shreeshrii Date: Sun, 14 Jun 2015 18:21:01 +0530 Subject: [PATCH] Use Tesseract-ocr Traineddata from github Tesseract OCR project has moved to gothub and all new changes are being made here only. --- mingw-w64-tesseract-ocr-eng/PKGBUILD | 34 ++++++++++++++++++++-------- 1 file changed, 25 insertions(+), 9 deletions(-) diff --git a/mingw-w64-tesseract-ocr-eng/PKGBUILD b/mingw-w64-tesseract-ocr-eng/PKGBUILD index f4d5f74a86..caf5228ff3 100644 --- a/mingw-w64-tesseract-ocr-eng/PKGBUILD +++ b/mingw-w64-tesseract-ocr-eng/PKGBUILD @@ -1,27 +1,43 @@ # Maintainer: Alexey Pavlov # Maintainer: Ray Donnelly -# The traineddata files from the source should be unpacked in tesseract ocr's tessdata directory -# this source tar.gz does not contain any makefile +# The traineddata and related files for english should be placed in tesseract ocr's tessdata directory + _realname=tesseract-ocr-eng pkgname="${MINGW_PACKAGE_PREFIX}-${_realname}" -pkgver=3.03.298e31465a44 +pkgver=3.04.rc1.master pkgrel=1 -pkgdesc="English language data (3.02) for Tesseract OCR engine (mingw-w64)" +pkgdesc="English language data for Tesseract OCR engine (mingw-w64)" arch=('any') -url="https://code.google.com/p/tesseract-ocr/" +url="https://github.com/tesseract-ocr/tessdata" license=("Apache License 2.0") makedepends=("${MINGW_PACKAGE_PREFIX}-gcc" "${MINGW_PACKAGE_PREFIX}-pkg-config") depends=(${MINGW_PACKAGE_PREFIX}-gcc-libs ${MINGW_PACKAGE_PREFIX}-tesseract-ocr ${MINGW_PACKAGE_PREFIX}-zlib) options=('!libtool' 'strip') -source=("https://tesseract-ocr.googlecode.com/files/tesseract-ocr-3.02.eng.tar.gz" +source=("https://github.com/tesseract-ocr/tessdata/blob/master/eng.traineddata" + "https://github.com/tesseract-ocr/tessdata/blob/master/eng.cube.bigrams" + "https://github.com/tesseract-ocr/tessdata/blob/master/eng.cube.fold" + "https://github.com/tesseract-ocr/tessdata/blob/master/eng.cube.lm" + "https://github.com/tesseract-ocr/tessdata/blob/master/eng.cube.nn" + "https://github.com/tesseract-ocr/tessdata/blob/master/eng.cube.params" + "https://github.com/tesseract-ocr/tessdata/blob/master/eng.cube.size" + "https://github.com/tesseract-ocr/tessdata/blob/master/eng.cube.word-freq" + "https://github.com/tesseract-ocr/tessdata/blob/master/eng.tesseract_cube.nn" ) -sha256sums=('989ed4c3a5b246d7353893e466c353099d8b73a1' +sha256sums=('SKIP' +'SKIP' +'SKIP' +'SKIP' +'SKIP' +'SKIP' +'SKIP' +'SKIP' +'SKIP' ) prepare() { - cd "$srcdir/${_realname}-298e31465a44" + cd "$srcdir/${_realname}-master" } build() { @@ -30,7 +46,7 @@ build() { cd "${srcdir}/build-${MINGW_CHOST}" - ${srcdir}/${_realname}-298e31465a44/configure \ + ${srcdir}/${_realname}-master/configure \ --build=${MINGW_CHOST} \ --host=${MINGW_CHOST} \ --target=${MINGW_CHOST} \