omnilingual-asr/pyproject.toml at main · blackjaguar-ai/omnilingual-asr · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
[build-system]
requires = ["packaging~=23.1", "setuptools~=67.8", "wheel~=0.40"]
build-backend = "setuptools.build_meta"

[project]
name = "omnilingual_asr"
dynamic = ["version"]
description = "Omnilingual ASR Modeling Library"
authors = [{name = "Meta AI Research"}]
readme = "README.md"
license = {file = "LICENSE"}
keywords = [
    "speech recognition", "ASR", "automatic speech recognition",
    "multilingual", "omnilingual", "zero-shot", "few-shot",
    "speech-to-text", "audio processing", "wav2vec2", "fairseq2",
    "language models", "speech technology",  "CTC",
    "transformer models", "beam search",
    "audio transcription", "voice recognition"
]

# zip_safe = false
classifiers=[
    "Topic :: Scientific/Engineering",
    "Development Status :: 4 - Beta",
]

requires-python = ">=3.10"
dependencies = [
    "fairseq2[arrow]>=0.5.2,<=0.6",
    "pyarrow>=20.0.0",
    "torch",
    "torchaudio",
    "numba",
    "pandas",
    "numpy",
    "kenlm",
    "polars>=1.29.0",
]

[project.optional-dependencies]
  data = [
    "fire",
    "transformers",
    "datasets~=3.6.0",
    "ray>=2.49.0",
    "soundfile>=0.13.1",
    "librosa>=0.11.0",
    "unidecode",
    "uroman",
    "fire",
    "retrying",
  ]
  dev = [
    "black~=25.1",
    "flake8~=7.1",
    "flake8-pyi~=24.6",
    "flake8-pyproject~=1.2",
    "isort~=6.0",
    "mypy~=1.15.0",
    "pandas-stubs~=2.2.3",
    "pyarrow-stubs~=20.0",
    "pytest~=8.3",
    "shellcheck-py~=0.9",
    "types-tqdm~=4.67.0",
    "types-editdistance~=0.6",
    "types-psutil~=5.9",
    "types-setuptools~=75.8",
  ]

[project.entry-points."fairseq2.extension"]
setup = "omnilingual_asr:setup_fairseq2_extension"


[project.urls]
  Source = "https://github.com/facebookresearch/omnilingual-asr"
  Tracker = "https://github.com/facebookresearch/omnilingual-asr/issues"

[tool.setuptools.dynamic]
version = {attr = "omnilingual_asr.__version__"}

[tool.setuptools.package-data]
"omnilingual_asr" = ["py.typed"]
"omnilingual_asr.cards" = ["**/*.yaml"]
"omnilingual_asr.models.wav2vec2_llama" = ["*.parquet"]

[tool.flake8]
extend_ignore = ["E", "Y"]  # Black

[tool.isort]
profile = "black"

[tool.mypy]
disable_error_code = "type-abstract,typeddict-unknown-key"
disallow_untyped_calls = false
disallow_untyped_decorators = false
files = "src,tests"
python_version = "3.10"
show_error_codes = true
show_error_context = true
strict = true
warn_unused_configs = false
warn_unused_ignores = false

[tool.pytest.ini_options]
minversion = "7.1"
testpaths = ["tests"]