movefasta/nixpkgs
0
0
Fork 0
mirror of https://github.com/NixOS/nixpkgs.git synced 2025-07-13 21:50:33 +03:00
Code Activity

python/pyocr: Move package into python-modules

Browse source 
We already have a patch feeling lonely inside the python-modules
directory and to have everything at one place let's actually move pyocr
into its own dedicated directory so it's easier to patch it up (which
we're going to).

Right now, the package fails to build because of a few test failures, so
I haven't tested this apart from evaluating.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
This commit is contained in:
aszlig 2017-09-02 03:43:21 +02:00
parent 40b76c8809
commit 3086fc7f83
No known key found for this signature in database
GPG key ID: 1DE8E48E57DB5436
3 changed files with 65 additions and 60 deletions
Download patch file Download diff file Expand all files Collapse all files

64
pkgs/development/python-modules/pyocr/default.nix Normal file
View file

@ -0,0 +1,64 @@
{ lib, fetchFromGitHub, buildPythonPackage, pillow, six
, tesseract, cuneiform
}:
buildPythonPackage rec {
name = "pyocr-${version}";
version = "0.4.6";
# Don't fetch from PYPI because it doesn't contain tests.
src = fetchFromGitHub {
owner = "jflesch";
repo = "pyocr";
rev = version;
sha256 = "0amyhkkm400qzbw65ivyzrzxl2r7vxqgsgqm7ml95m7gwkwhnzz0";
};
patches = [ ./tesseract.patch ];
postPatch = ''
sed -i \
-e 's,^\(TESSERACT_CMD *= *\).*,\1"${tesseract}/bin/tesseract",' \
-e 's,^\(CUNEIFORM_CMD *= *\).*,\1"${cuneiform}/bin/cuneiform",' \
-e '/^CUNIFORM_POSSIBLE_PATHS *= *\[/,/^\]$/ {
c CUNIFORM_POSSIBLE_PATHS = ["${cuneiform}/share/cuneiform"]
}' src/pyocr/{tesseract,cuneiform}.py
sed -i -r \
-e 's,"libtesseract\.so\.3","${tesseract}/lib/libtesseract.so",' \
-e 's,^(TESSDATA_PREFIX *=).*,\1 "${tesseract}/share/tessdata",' \
src/pyocr/libtesseract/tesseract_raw.py
# Disable specific tests that are probably failing because of this issue:
# https://github.com/jflesch/pyocr/issues/52
for test in $disabledTests; do
file="''${test%%:*}"
fun="''${test#*:}"
echo "$fun = unittest.skip($fun)" >> "tests/tests_$file.py"
done
'';
disabledTests = [
"cuneiform:TestTxt.test_basic"
"cuneiform:TestTxt.test_european"
"cuneiform:TestTxt.test_french"
"cuneiform:TestWordBox.test_basic"
"cuneiform:TestWordBox.test_european"
"cuneiform:TestWordBox.test_french"
"libtesseract:TestBasicDoc.test_basic"
"libtesseract:TestDigitLineBox.test_digits"
"libtesseract:TestLineBox.test_japanese"
"libtesseract:TestTxt.test_japanese"
"libtesseract:TestWordBox.test_japanese"
"tesseract:TestDigitLineBox.test_digits"
"tesseract:TestTxt.test_japanese"
];
propagatedBuildInputs = [ pillow six ];
meta = {
homepage = "https://github.com/jflesch/pyocr";
description = "A Python wrapper for Tesseract and Cuneiform";
license = lib.licenses.gpl3Plus;
};
}

0
pkgs/development/python-modules/pyocr-tesseract.patch → pkgs/development/python-modules/pyocr/tesseract.patch
View file

61
pkgs/top-level/python-packages.nix
View file

@ -18113,66 +18113,7 @@ in {
};
};
pyocr = buildPythonPackage rec {
name = "pyocr-${version}";
version = "0.4.6";
# Don't fetch from PYPI because it doesn't contain tests.
src = pkgs.fetchFromGitHub {
owner = "jflesch";
repo = "pyocr";
rev = version;
sha256 = "0amyhkkm400qzbw65ivyzrzxl2r7vxqgsgqm7ml95m7gwkwhnzz0";
};
patches = [ ../development/python-modules/pyocr-tesseract.patch ];
postPatch = ''
sed -i \
-e 's,^\(TESSERACT_CMD *= *\).*,\1"${pkgs.tesseract}/bin/tesseract",' \
-e 's,^\(CUNEIFORM_CMD *= *\).*,\1"${pkgs.cuneiform}/bin/cuneiform",' \
-e '/^CUNIFORM_POSSIBLE_PATHS *= *\[/,/^\]$/ {
c CUNIFORM_POSSIBLE_PATHS = ["${pkgs.cuneiform}/share/cuneiform"]
}' src/pyocr/{tesseract,cuneiform}.py
sed -i -r \
-e 's,"libtesseract\.so\.3","${pkgs.tesseract}/lib/libtesseract.so",' \
-e 's,^(TESSDATA_PREFIX *=).*,\1 "${pkgs.tesseract}/share/tessdata",' \
src/pyocr/libtesseract/tesseract_raw.py
# Disable specific tests that are probably failing because of this issue:
# https://github.com/jflesch/pyocr/issues/52
for test in $disabledTests; do
file="''${test%%:*}"
fun="''${test#*:}"
echo "$fun = unittest.skip($fun)" >> "tests/tests_$file.py"
done
'';
disabledTests = [
"cuneiform:TestTxt.test_basic"
"cuneiform:TestTxt.test_european"
"cuneiform:TestTxt.test_french"
"cuneiform:TestWordBox.test_basic"
"cuneiform:TestWordBox.test_european"
"cuneiform:TestWordBox.test_french"
"libtesseract:TestBasicDoc.test_basic"
"libtesseract:TestDigitLineBox.test_digits"
"libtesseract:TestLineBox.test_japanese"
"libtesseract:TestTxt.test_japanese"
"libtesseract:TestWordBox.test_japanese"
"tesseract:TestDigitLineBox.test_digits"
"tesseract:TestTxt.test_japanese"
];
propagatedBuildInputs = [ self.pillow self.six ];
meta = {
homepage = "https://github.com/jflesch/pyocr";
description = "A Python wrapper for Tesseract and Cuneiform";
license = licenses.gpl3Plus;
};
};
pyocr = callPackage ../development/python-modules/pyocr { };
pyparsing = callPackage ../development/python-modules/pyparsing { };