nixpkgs/pkgs/tools/text/ocrmypdf/default.nix
2019-12-15 23:15:07 -08:00

106 lines
2.4 KiB
Nix

{ fetchFromGitHub
, ghostscript
, img2pdf
, jbig2enc
, leptonica
, pngquant
, python3
, python3Packages
, qpdf
, stdenv
, tesseract4
, unpaper
}:
let
inherit (python3Packages) buildPythonApplication;
runtimeDeps = with python3Packages; [
ghostscript
jbig2enc
leptonica
pngquant
qpdf
tesseract4
unpaper
pillow
];
in buildPythonApplication rec {
pname = "ocrmypdf";
version = "9.2.0";
disabled = ! python3Packages.isPy3k;
src = fetchFromGitHub {
owner = "jbarlow83";
repo = "OCRmyPDF";
rev = "v${version}";
sha256 = "1mvc6x5nn242z65pxv39ch71vaikgi89bb0sjbfy2jbw91vk41xa";
};
nativeBuildInputs = with python3Packages; [
pytestrunner
setuptools
setuptools-scm-git-archive
setuptools_scm
];
propagatedBuildInputs = with python3Packages; [
cffi
chardet
img2pdf
pdfminer
pikepdf
pillow
reportlab
setuptools
tqdm
];
checkInputs = with python3Packages; [
pypdf2
pytest
pytest-helpers-namespace
pytest_xdist
pytestcov
pytestrunner
python-xmp-toolkit
setuptools
] ++ runtimeDeps;
postPatch = ''
substituteInPlace src/ocrmypdf/leptonica.py \
--replace "lept = ffi.dlopen(_libpath)" \
'lept = ffi.dlopen("${stdenv.lib.makeLibraryPath [leptonica]}/liblept${stdenv.hostPlatform.extensions.sharedLibrary}")'
'';
# The tests take potentially 20+ minutes, depending on machine
doCheck = false;
# These tests fail and it might be upstream problem... or packaging. :)
# development is happening on macos and the pinned test versions are
# significantly newer than nixpkgs has. Program still works...
# (to the extent I've used it) -- Kiwi
checkPhase = ''
export HOME=$TMPDIR
pytest -k 'not test_force_ocr_on_pdf_with_no_images \
and not test_tesseract_crash \
and not test_tesseract_crash_autorotate \
and not test_ghostscript_pdfa_failure \
and not test_gs_render_failure \
and not test_gs_raster_failure \
and not test_bad_utf8 \
and not test_old_unpaper'
'';
makeWrapperArgs = [ "--prefix PATH : ${stdenv.lib.makeBinPath [ ghostscript jbig2enc pngquant qpdf tesseract4 unpaper ]}" ];
meta = with stdenv.lib; {
homepage = "https://github.com/jbarlow83/OCRmyPDF";
description = "Adds an OCR text layer to scanned PDF files, allowing them to be searched";
license = licenses.gpl3;
platforms = platforms.linux;
maintainers = [ maintainers.kiwi ];
};
}