component add tesseract 4.0 and dependancys to slapos

This commit contain tesseract4.0 and its up do date leptonica
dependancy as well as a python wrapper that use tesseract API to do text
recognition.

Currently, the wrapper cannot initialise the api as intended and the
traineddata emplacement is not set (an new environnement variable should
be introduced in zope, but it should be possible to "force" a new path
with a patch that change the tesserocr _DEFAULT_PATH value.

component/ocr-leptonica/buildout.cfg

+[buildout]
+extends =
+
+parts = ocr-leptonica
+
+[ocr-leptonica]
+recipe = slapos.recipe.cmmi
+url = http://www.leptonica.com/source/leptonica-1.74.1.tar.gz
+configure-options =
+  --disable-static
\ No newline at end of file

component/ocr-tesseract/buildout.cfg

+[buildout]
+extends =
+  ../ocr-leptonica/buildout.cfg
+  ../pkgconfig/buildout.cfg
+  ../libtool/buildout.cfg
+  ../autoconf/buildout.cfg
+  ../automake/buildout.cfg
+  ../cmake/buildout.cfg
+
+parts = ocr-tesseract
+
+[ocr-tesseract-eng-traineddata]
+recipe = slapos.recipe.build:download
+strip-top-level-dir = false
+filename = eng.traineddata
+url = https://github.com/tesseract-ocr/tessdata/raw/4.00/eng.traineddata
+md5sum = 7af2ad02d11702c7092a5f8dd044d52f
+
+[tesseract-env]
+PATH=${pkg-config:location}/bin:${autoconf:location}/bin:${automake:location}/bin:${libtool:location}/bin:%(PATH)s
+LIBTOOL=${libtool:location}/bin/libtool
+CPPFLAGS=-I${ocr-leptonica:location}/include
+LDFLAGS =-L${ocr-leptonica:location}/lib -Wl,-rpath=${ocr-leptonica:location}/lib
+PKG_CONFIG_PATH=${ocr-leptonica:location}/lib/pkgconfig
+TESSDATA_PREFIX=${ocr-tesseract-eng-traineddata:location}
+
+[ocr-tesseract]
+recipe = slapos.recipe.cmmi
+url = https://github.com/tesseract-ocr/tesseract/tarball/84db453d3ac002e141a22cd04b41edaac3c50386/tesseract-4.0.tar.gz
+environment-section = tesseract-env
+patch-options = -p1
+patches = ${:_profile_base_location_}/tesseract_cmake.patch
+make-options =
+  -C build
+pre-configure =
+  mkdir build
+configure-command =
+  cd build
+  ${cmake:location}/bin/cmake -DCMAKE_INSTALL_PREFIX=${buildout:parts-directory}/${:_buildout_section_name_} ..
\ No newline at end of file

component/ocr-tesseract/tesseract_cmake.patch

+diff --git a/CMakeLists.txt b/CMakeLists.txt
+index 079d8482..2b69f212 100644
+--- a/CMakeLists.txt
++++ b/CMakeLists.txt
+@@ -41,7 +41,7 @@ set_property(GLOBAL PROPERTY PREDEFINED_TARGETS_FOLDER "CMake Targets")
+ project(tesseract C CXX)
+ 
+ set(VERSION_MAJOR 4)
+-set(VERSION_MINOR 00)
++set(VERSION_MINOR 00.00)
+ set(VERSION_PLAIN ${VERSION_MAJOR}.${VERSION_MINOR})
+ 
+ set(MINIMUM_LEPTONICA_VERSION 1.74)
+@@ -156,8 +156,8 @@ include_directories(wordrec)
+ # LIBRARY tesseract
+ ########################################
+ 
+-string(SUBSTRING ${VERSION_MINOR} 0 1 VERSION_MINOR_0)
+-string(SUBSTRING ${VERSION_MINOR} 1 1 VERSION_MINOR_1)
++string(SUBSTRING ${VERSION_MINOR} 0 2 VERSION_MINOR_0)
++string(SUBSTRING ${VERSION_MINOR} 3 2 VERSION_MINOR_1)
+ 
+ file(GLOB tesseract_src
+     arch/*.cpp
+@@ -224,10 +224,8 @@ endif()
+ target_link_libraries           (libtesseract ${LIB_Ws2_32} ${LIB_pthread})
+ set_target_properties           (libtesseract PROPERTIES VERSION ${VERSION_MAJOR}.${VERSION_MINOR_0}.${VERSION_MINOR_1})
+ set_target_properties           (libtesseract PROPERTIES SOVERSION ${VERSION_MAJOR}.${VERSION_MINOR_0}.${VERSION_MINOR_1})
+-if (WIN32)
+-set_target_properties           (libtesseract PROPERTIES OUTPUT_NAME tesseract${VERSION_MAJOR}${VERSION_MINOR})
+-set_target_properties           (libtesseract PROPERTIES DEBUG_OUTPUT_NAME tesseract${VERSION_MAJOR}${VERSION_MINOR}d)
+-endif()
++set_target_properties           (libtesseract PROPERTIES OUTPUT_NAME tesseract)
++set_target_properties           (libtesseract PROPERTIES DEBUG_OUTPUT_NAME tesseractd)
+ 
+ if (NOT CPPAN_BUILD)
+     target_link_libraries       (libtesseract ${Leptonica_LIBRARIES})
+diff --git a/tesseract.pc.cmake b/tesseract.pc.cmake
+index f9f64f6c..475236f1 100644
+--- a/tesseract.pc.cmake
++++ b/tesseract.pc.cmake
+@@ -7,6 +7,6 @@ Name: @tesseract_NAME@
+ Description: An OCR Engine that was developed at HP Labs between 1985 and 1995... and now at Google.
+ URL: https://github.com/tesseract-ocr/tesseract
+ Version: @tesseract_VERSION@
+-Libs: -L${libdir} -l@tesseract_OUTPUT_NAME@
++Libs: -L${libdir} -ltesseract
+ Libs.private:
+ Cflags: -I${includedir} -I${includedir}/tesseract

component/ocr-tesserocr/buildout.cfg

+[buildout]
+extends =
+  ../ocr-tesseract/buildout.cfg
+  ../ocr-leptonica/buildout.cfg
+  ../cython/buildout.cfg
+
+parts =
+  tesserocr-patch
+  ocr-tesserocr
+
+[tesserocr-repository]
+recipe = slapos.recipe.build:gitclone
+repository = https://github.com/sirfz/tesserocr
+git-executable = ${git:location}/bin/git
+branch = tesseract4
+revision = 61b85da
+
+[tesserocr-env]
+PATH=${pkg-config:location}/bin:${ocr-tesseract:location}/bin:%(PATH)s
+PKG_CONFIG_PATH=${ocr-tesseract:location}/lib/pkgconfig:${ocr-leptonica:location}/lib/pkgconfig
+CPPFLAGS = -I${ocr-leptonica:location}/include -I${ocr-tesseract:location}/include
+LDFLAGS =-L${ocr-leptonica:location}/lib -Wl,-rpath=${ocr-leptonica:location}/lib -L${ocr-tesseract:location}/lib -Wl,-rpath=${ocr-tesseract:location}/lib
+
+[ocr-tesserocr]
+recipe = zc.recipe.egg:develop
+setup = ${tesserocr-repository:location}
+egg = tesserocr
+setup-eggs =
+  ${cython:egg}
+environment = tesserocr-env
\ No newline at end of file

stack/erp5/buildout.cfg

@@ -60,6 +60,7 @@ extends =
   ../../component/findutils/buildout.cfg
   ../../component/userhosts/buildout.cfg
   ../../component/postfix/buildout.cfg
+  ../../component/ocr-tesserocr/buildout.cfg
   ../../software/ipython_notebook/software.cfg
   ../../software/neoppod/software-common.cfg
 # keep neoppod extends last
@@ -140,9 +141,12 @@ parts +=
   ipython-notebook
   instance-jupyter
   
+# Ocr
+  ocr-tesserocr
+
 # override python2.7 to add SlapOS libstdc++ in RPATH.
 [python2.7]
-extra-ldflags = -Wl,-rpath=${gcc:location}/lib -Wl,-rpath=${gcc:location}/lib64
+extra-ldflags = -Wl,-rpath=${gcc:location}/lib -Wl,-rpath=${gcc:location}/lib64 -Wl,-rpath=${ocr-leptonica:location}/lib
 
 # override instance-jupyter not to render into default template.cfg
 [instance-jupyter]
@@ -453,6 +457,7 @@ eggs = ${neoppod:eggs}
   ${pycrypto-python:egg}
   ${scipy:egg}
   ${scikit-learn:egg}
+  ${ocr-tesserocr:egg}
   lock_file
   astor
   PyStemmer