chiark - git - mdw - termux-packages/blob - packages/tesseract/build.sh

   1 TERMUX_PKG_HOMEPAGE=https://github.com/tesseract-ocr/tesseract
   2 TERMUX_PKG_DESCRIPTION="Tesseract is probably the most accurate open source OCR engine available"
   3 TERMUX_PKG_VERSION=3.04.01
   4 TERMUX_PKG_DEPENDS="libtool, libuuid, leptonica"
   5 TERMUX_PKG_SRCURL=https://github.com/tesseract-ocr/tesseract/archive/${TERMUX_PKG_VERSION}.tar.gz
   6 TERMUX_PKG_FOLDERNAME=tesseract-${TERMUX_PKG_VERSION}
   7
   8 termux_step_pre_configure() {
   9         # http://blog.matt-swain.com/post/26419042500/installing-tesseract-ocr-on-mac-os-x-lion
  10         export LIBLEPT_HEADERSDIR=${TERMUX_PREFIX}/include/leptonica
  11
  12         cd $TERMUX_PKG_SRCDIR
  13         perl -p -i -e 's|ADD_RT, true|ADD_RT, false|g' configure.ac
  14         ./autogen.sh
  15 }
  16
  17 termux_step_post_make_install() {
  18         # download english trained data
  19         cd "${TERMUX_PREFIX}/share/tessdata"
  20         rm -f eng.*
  21         for f in cube.{bigrams,fold,lm,nn,params,size,word-freq} tesseract_cube.nn traineddata; do
  22                 f=eng.$f
  23                 # From the tessdata README: "These language data files only work with
  24                 # Tesseract 4. They are based on the sources in tesseract-ocr/langdata on GitHub.
  25                 # Get language data files for Tesseract 3.04 or 3.05 from the 3.04 tree."
  26                 termux_download https://raw.githubusercontent.com/tesseract-ocr/tessdata/3.04.00/$f $f
  27         done
  28 }