tjh.dev
1{
2 lib,
3 fetchFromGitHub,
4 buildPythonPackage,
5 fire,
6 fugashi,
7 jaconv,
8 loguru,
9 numpy,
10 pillow,
11 pyperclip,
12 torch,
13 transformers,
14 unidic-lite,
15 pythonOlder,
16}:
17buildPythonPackage rec {
18 pname = "manga-ocr";
19 version = "0.1.11";
20 disabled = pythonOlder "3.7";
21
22 src = fetchFromGitHub {
23 owner = "kha-white";
24 repo = pname;
25 rev = "refs/tags/v${version}";
26 hash = "sha256-cLmgHBt6HvhY6Hb9yQ425Gk181axnMr+Mp2LxSmPoDg=";
27 };
28
29 preBuild = ''
30 # remove subproject dedicated to model training
31 rm -rf manga_ocr_dev
32 # copy assets/example.jpg inside the package
33 # required by https://github.com/kha-white/manga-ocr/blob/ba1b0d94a8ef6676b618ba4e5ffe8ce2ab655270/manga_ocr/ocr.py#L27-L30
34 # see also package_data.patch
35 mkdir manga_ocr/assets
36 cp assets/example.jpg manga_ocr/assets/example.jpg
37 '';
38
39 patches = [
40 # instruct setuptool to copy assets/example.jpg to package when building wheel
41 ./package_data.patch
42 ];
43
44 propagatedBuildInputs = [
45 # taken from requirements.txt
46 fire
47 fugashi
48 jaconv
49 loguru
50 numpy
51 pillow
52 pyperclip
53 torch
54 transformers
55 unidic-lite
56 ];
57
58 meta = with lib; {
59 description = "Optical character recognition for Japanese text, with the main focus being Japanese manga";
60 homepage = "https://github.com/kha-white/manga-ocr";
61 changelog = "https://github.com/kha-white/manga-ocr/releases/tag/${version}";
62 license = licenses.asl20;
63 maintainers = with maintainers; [ laurent-f1z1 ];
64 };
65}