ajattix/packages/manga-ocr.scm

(define-module (manga-ocr))


(define-public manga-ocr
(package(define-module (manga-ocr)
  #:use-module (guix packages)
  #:use-module (guix download)
  #:use-module (guix build-system python)
  #:use-module (guix build-system pyproject)
  #:use-module (guix build-system cmake)
  #:use-module (gnu packages python-xyz)
  #:use-module (gnu packages xdisorg)
  #:use-module (gnu packages machine-learning)
  #:use-module (gnu packages python-web)
  #:use-module (gnu packages version-control)
  #:use-module ((guix licenses) #:prefix license:))

(define-public python-fugashi
  (package
   (name "python-fugashi")
   (version "1.3.0")
   (source
    (origin
     (method url-fetch)
     (uri (pypi-uri "fugashi" version))
     (sha256
      (base32 "1z61b7c71rz493fqzyg4dwzzhhv22dy6v2i3fivqjn8qwfnc1vhd"))))
   (build-system pyproject-build-system)
   (home-page "https://github.com/polm/fugashi")
   (synopsis "A Cython MeCab wrapper for fast, pythonic Japanese tokenization.")
   (description
    "This package provides a Cython @code{MeCab} wrapper for fast, pythonic Japanese
tokenization.")
   (license license:expat)))

(define-public python-jaconv
  (package
   (name "python-jaconv")
   (version "0.3.4")
   (source
    (origin
     (method url-fetch)
     (uri (pypi-uri "jaconv" version))
     (sha256
      (base32 "0v8xlfrmc26r15arb5w2nzdzyvzw606gmjgncanxpqmhy3rmaz4y"))))
   (build-system pyproject-build-system)
   (home-page "https://github.com/ikegami-yukino/jaconv")
   (synopsis
    "Pure-Python Japanese character interconverter for Hiragana, Katakana, Hankaku, Zenkaku and more")
   (description
    "Pure-Python Japanese character interconverter for Hiragana, Katakana, Hankaku,
Zenkaku and more")
   (license license:expat)))

(define-public python-transformers
  (package
   (name "python-transformers")
   (version "4.36.2")
   (source
    (origin
     (method url-fetch)
     (uri (pypi-uri "transformers" version))
     (sha256
      (base32 "1v5lq3a7gjp2v71vcb4w81b8anzwplmpsm0ya20k4ya7gs4qw1nq"))))
   (build-system pyproject-build-system)
   (propagated-inputs (list python-filelock
                            python-huggingface-hub
                            python-numpy
                            python-packaging
                            python-pyyaml
                            python-regex
                            python-requests
                            python-safetensors
                            python-tokenizers
                            python-tqdm))
   (native-inputs (list python-accelerate
                        python-av
                        python-beautifulsoup4
                        python-codecarbon
                        python-cookiecutter
                        python-datasets
                        python-decord
                        python-deepspeed
                        python-dill
                        python-evaluate
                        python-faiss-cpu
                        python-flax
                        python-fugashi
                        python-gitpython
                        python-hf-doc-builder
                        python-ipadic
                        python-isort
                        python-jax
                        python-jaxlib
                        python-kenlm
                        python-keras-nlp
                        python-librosa
                        python-nltk
                        python-onnxconverter-common
                        python-onnxruntime
                        python-onnxruntime-tools
                        python-optax
                        python-optuna
                        python-parameterized
                        python-phonemizer
                        python-pillow
                        python-protobuf
                        python-psutil
                        python-pyctcdecode
                        python-pydantic
                        python-pytest
                        python-pytest-timeout
                        python-pytest-xdist
                        python-ray
                        python-rhoknp
                        python-rjieba
                        python-rouge-score
                        python-ruff
                        python-sacrebleu
                        python-sacremoses
                        python-scikit-learn
                        python-sentencepiece
                        python-sigopt
                        python-sudachidict-core
                        python-sudachipy
                        python-tensorboard
                        python-tensorflow
                        python-tensorflow-text
                        python-tf2onnx
                        python-timeout-decorator
                        python-timm
                        python-tokenizers
                        python-torch
                        python-torchaudio
                        python-torchvision
                        python-unidic
                        python-unidic-lite
                        python-urllib3))
   (home-page "https://github.com/huggingface/transformers")
   (synopsis
    "State-of-the-art Machine Learning for JAX, PyTorch and TensorFlow")
   (description
    "State-of-the-art Machine Learning for JAX, @code{PyTorch} and @code{TensorFlow}")
   (license #f)))

(define-public python-unidic-lite
  (package
   (name "python-unidic-lite")
   (version "1.0.8")
   (source
    (origin
     (method url-fetch)
     (uri (pypi-uri "unidic-lite" version))
     (sha256
      (base32 "0idj4yp0sl27ylr2wzkybbh0wj7c843lp7cljw5d1m7xv5r4b7fv"))))
   (build-system pyproject-build-system)
   (home-page "https://github.com/polm/unidic-lite")
   (synopsis "A small version of UniDic packaged for Python")
   (description
    "This package provides a small version of @code{UniDic} packaged for Python")
   (license #f)))

(define-public python-huggingface-hub
  (package
   (name "python-huggingface-hub")
   (version "0.20.2")
   (source
    (origin
     (method url-fetch)
     (uri (pypi-uri "huggingface_hub" version))
     (sha256
      (base32 "1xhg41dj9g0f2qfz7vi17y5r074ji2ssg6yilg3k0433zz75yp11"))))
   (build-system pyproject-build-system)
   (propagated-inputs (list python-filelock
                            python-fsspec
                            python-packaging
                            python-pyyaml
                            python-requests
                            python-tqdm
                            python-typing-extensions))
   (native-inputs (list python-aiohttp
                        python-gradio
                        python-inquirerpy
                        python-jedi
                        python-jinja2
                        python-mypy
                        python-numpy
                        python-pillow
                        python-pydantic
                        python-pytest
                        python-pytest-asyncio
                        python-pytest-cov
                        python-pytest-env
                        python-pytest-rerunfailures
                        python-pytest-vcr
                        python-pytest-xdist
                        python-ruff
                        python-soundfile
                        python-types-pyyaml
                        python-types-requests
                        python-types-simplejson
                        python-types-toml
                        python-types-tqdm
                        python-types-urllib3
                        python-typing-extensions
                        python-urllib3))
   (home-page "https://github.com/huggingface/huggingface_hub")
   (synopsis
    "Client library to download and publish models, datasets and other repos on the huggingface.co hub")
   (description
    "Client library to download and publish models, datasets and other repos on the
huggingface.co hub")
   (license #f)))

(define-public python-safetensors
  (package
   (name "python-safetensors")
   (version "0.4.1")
   (source
    (origin
     (method url-fetch)
     (uri (pypi-uri "safetensors" version))
     (sha256
      (base32 "1j9nzg5665rjgvgzxgvrc3261rs88zlfzd1568iab0fsda76a113"))))
   (build-system pyproject-build-system)
   (native-inputs (list python-h5py
                        python-huggingface-hub
                        python-hypothesis
                        python-pytest
                        python-pytest-benchmark
                        python-safetensors
                        python-setuptools-rust))
   (home-page "")
   (synopsis "")
   (description "")
   (license #f)))

(define-public python-tokenizers
  (package
   (name "python-tokenizers")
   (version "0.15.0")
   (source
    (origin
     (method url-fetch)
     (uri (pypi-uri "tokenizers" version))
     (sha256
      (base32 "03jx7rmgapl2222z4m4bdz8r2qhjs7wgb4wylmypbgfankkydiqh"))))
   (build-system pyproject-build-system)
   (propagated-inputs (list python-huggingface-hub))
   (native-inputs (list python-black
                        python-datasets
                        python-numpy
                        python-pytest
                        python-requests
                        python-tokenizers))
   (home-page "")
   (synopsis "")
   (description "")
   (license #f)))

(define-public python-accelerate
  (package
   (name "python-accelerate")
   (version "0.26.1")
   (source
    (origin
     (method url-fetch)
     (uri (pypi-uri "accelerate" version))
     (sha256
      (base32 "12aga0dkmrj1nwvbwcx61b529f5b6gcg834pva3hsinrddmp2qxz"))))
   (build-system pyproject-build-system)
   (propagated-inputs (list python-huggingface-hub
                            python-numpy
                            python-packaging
                            python-psutil
                            python-pyyaml
                            python-safetensors
                            python-torch))
   (native-inputs (list python-bitsandbytes
                        python-black
                        python-comet-ml
                        python-datasets
                        python-deepspeed
                        python-dvclive
                        python-evaluate
                        python-hf-doc-builder
                        python-parameterized
                        python-pytest
                        python-pytest-subtests
                        python-pytest-xdist
                        python-rich
                        python-ruff
                        python-scikit-learn
                        python-scipy
                        python-tensorboard
                        python-timm
                        python-tqdm
                        python-transformers
                        python-urllib3
                        python-wandb))
   (home-page "https://github.com/huggingface/accelerate")
   (synopsis "Accelerate")
   (description "Accelerate")
   (license #f)))

(define-public python-codecarbon
  (package
   (name "python-codecarbon")
   (version "2.3.2")
   (source
    (origin
     (method url-fetch)
     (uri (pypi-uri "codecarbon" version))
     (sha256
      (base32 "1lpn354r5vd5d0a1nzsqq4prsxbdkdw00wk70kx5f4h9jjg1sk2z"))))
   (build-system pyproject-build-system)
   (propagated-inputs (list python-arrow
                            python-click
                            python-fuzzywuzzy
                            python-pandas
                            python-prometheus-client
                            python-psutil
                            python-py-cpuinfo
                            python-pynvml
                            python-requests))
   (home-page "")
   (synopsis "")
   (description "")
   (license #f)))

(define-public python-datasets
  (package
   (name "python-datasets")
   (version "2.16.1")
   (source
    (origin
     (method url-fetch)
     (uri (pypi-uri "datasets" version))
     (sha256
      (base32 "105013rvzphk1mcxbhyh2wgdzjww673kn4m2zpj1skcqn7liacmd"))))
   (build-system pyproject-build-system)
   (propagated-inputs (list python-aiohttp
                            python-dill
                            python-filelock
                            python-fsspec
                            python-huggingface-hub
                            python-multiprocess
                            python-numpy
                            python-packaging
                            python-pandas
                            python-pyarrow
                            python-pyarrow-hotfix
                            python-pyyaml
                            python-requests
                            python-tqdm
                            python-xxhash))
   (native-inputs (list python-absl-py
                        python-accelerate
                        python-apache-beam
                        python-bert-score
                        python-elasticsearch
                        python-faiss-cpu
                        python-jax
                        python-jaxlib
                        python-jiwer
                        python-joblib
                        python-joblibspark
                        python-langdetect
                        python-librosa
                        python-lz4
                        python-mauve-text
                        python-nltk
                        python-pillow
                        python-py7zr
                        python-pyspark
                        python-pytest
                        python-pytest-datadir
                        python-pytest-xdist
                        python-rarfile
                        python-requests-file
                        python-rouge-score
                        python-ruff
                        python-s3fs
                        python-sacrebleu
                        python-sacremoses
                        python-scikit-learn
                        python-scipy
                        python-sentencepiece
                        python-seqeval
                        python-six
                        python-soundfile
                        python-spacy
                        python-sqlalchemy
                        python-tensorflow
                        python-tensorflow-macos
                        python-texttable
                        python-tiktoken
                        python-tldextract
                        python-toml
                        python-torch
                        python-transformers
                        python-typer
                        python-typing-extensions
                        python-werkzeug
                        python-zstandard))
   (home-page "https://github.com/huggingface/datasets")
   (synopsis "HuggingFace community-driven open-source library of datasets")
   (description
    "@code{HuggingFace} community-driven open-source library of datasets")
   (license license:asl2.0)))

(define-public python-decord
  (package
   (name "python-decord")
   (version "0.6.0")
   (source
    (origin
     (method url-fetch)
     (uri (string-append "https://github.com/dmlc/decord/releases/tag/v"
                         version ".tar.gz"))
     (sha256
      (base32 "1sxhk613ag1vk5hm6zxa0npjcn4gn0bgf1bq7id56qw44rlq1ki1"))))
   (build-system cmake-build-system)
   (synopsis "An efficient video loader for deep learning with smart shuffling that's super easy to digest")
   (description "Decord is a reverse procedure of Record. It provides convenient video slicing methods based on a thin wrapper on top of hardware accelerated video decoders")
   (home-page "https://github.com/dmlc/decord")
   (license license:epl1.0)))

(define-public python-deepspeed
  (package
   (name "python-deepspeed")
   (version "0.13.0")
   (source
    (origin
     (method url-fetch)
     (uri (pypi-uri "deepspeed" version))
     (sha256
      (base32 "018a6nzyr9wzzbpcqcxxp3q2jc0w343269a0bdvnrwqs0v5av478"))))
   (build-system pyproject-build-system)
   (propagated-inputs (list python-hjson
                            python-ninja
                            python-numpy
                            python-packaging
                            python-psutil
                            python-py-cpuinfo
                            python-pydantic
                            python-pynvml
                            python-torch
                            python-tqdm))
   (native-inputs (list python-accelerate
                        python-clang-format
                        python-coverage
                        python-deepspeed-kernels
                        python-docutils
                        python-future
                        python-importlib-metadata
                        python-mup
                        python-pre-commit
                        python-pytest
                        python-pytest-forked
                        python-pytest-randomly
                        python-pytest-xdist
                        python-recommonmark
                        python-sphinx
                        python-sphinx-rtd-theme
                        python-tensorboard
                        python-torchvision
                        python-transformers
                        python-wandb))
   (home-page "http://deepspeed.ai")
   (synopsis "DeepSpeed library")
   (description "@code{DeepSpeed} library")
   (license #f)))

(define-public python-evaluate
  (package
   (name "python-evaluate")
   (version "0.4.1")
   (source
    (origin
     (method url-fetch)
     (uri (pypi-uri "evaluate" version))
     (sha256
      (base32 "1nd3rmrq13c85471dyd8jsnbplgvaklhj1ca1mvpkvcw0prdj8fp"))))
   (build-system pyproject-build-system)
   (propagated-inputs (list python-datasets
                            python-dill
                            python-fsspec
                            python-huggingface-hub
                            python-importlib-metadata
                            python-multiprocess
                            python-numpy
                            python-packaging
                            python-pandas
                            python-requests
                            python-responses
                            python-tqdm
                            python-xxhash))
   (native-inputs (list python-absl-py
                        python-accelerate
                        python-bert-score
                        python-black
                        python-cer
                        python-charcut
                        python-flake8
                        python-isort
                        python-jiwer
                        python-mauve-text
                        python-nltk
                        python-pytest
                        python-pytest-datadir
                        python-pytest-xdist
                        python-pyyaml
                        python-requests-file
                        python-rouge-score
                        python-sacrebleu
                        python-sacremoses
                        python-scikit-learn
                        python-scipy
                        python-sentencepiece
                        python-seqeval
                        python-six
                        python-tensorflow
                        python-texttable
                        python-tldextract
                        python-toml
                        python-torch
                        python-transformers
                        python-trectools
                        python-unidecode
                        python-werkzeug))
   (home-page "https://github.com/huggingface/evaluate")
   (synopsis "HuggingFace community-driven open-source library of evaluation")
   (description
    "@code{HuggingFace} community-driven open-source library of evaluation")
   (license license:asl2.0)))

(define-public python-faiss-cpu
  (package
   (name "python-faiss-cpu")
   (version "1.7.4")
   (source
    (origin
     (method url-fetch)
     (uri (pypi-uri "faiss-cpu" version))
     (sha256
      (base32 "02nxhaf5wgrdbs8qn67rvw54j8iryw861gq36d1z96q71hdw6p96"))))
   (build-system pyproject-build-system)
   (home-page "https://github.com/kyamagu/faiss-wheels")
   (synopsis
    "A library for efficient similarity search and clustering of dense vectors.")
   (description
    "This package provides a library for efficient similarity search and clustering
of dense vectors.")
   (license license:expat)))

(define-public python-flax
  (package
   (name "python-flax")
   (version "0.7.5")
   (source
    (origin
     (method url-fetch)
     (uri (pypi-uri "flax" version))
     (sha256
      (base32 "12blliydx64xwvf7bj9ys0zyz4d27sp7iis88vfr9c8fsvpl647m"))))
   (build-system pyproject-build-system)
   (propagated-inputs (list python-jax
                            python-msgpack
                            python-numpy
                            python-optax
                            python-orbax-checkpoint
                            python-pyyaml
                            python-rich
                            python-tensorstore
                            python-typing-extensions))
   (native-inputs (list python-black
                        python-clu
                        python-einops
                        python-gymnasium
                        python-jaxlib
                        python-jraph
                        python-ml-collections
                        python-mypy
                        python-nbstripout
                        python-opencv-python
                        python-pytest
                        python-pytest-cov
                        python-pytest-custom-exit-code
                        python-pytest-xdist
                        python-pytype
                        python-sentencepiece
                        python-tensorflow
                        python-tensorflow-datasets
                        python-tensorflow-text
                        python-torch))
   (home-page "")
   (synopsis "Flax: A neural network library for JAX designed for flexibility")
   (description
    "Flax: A neural network library for JAX designed for flexibility")
   (license #f)))

(define-public python-hf-doc-builder
  (package
   (name "python-hf-doc-builder")
   (version "0.4.0")
   (source
    (origin
     (method url-fetch)
     (uri (pypi-uri "hf-doc-builder" version))
     (sha256
      (base32 "13al44b8d8c552kmj52dapkm7wnafdra1z09bijsw85xhry2gh13"))))
   (build-system pyproject-build-system)
   (propagated-inputs (list python-gitpython
                            python-gql
                            python-nbformat
                            python-packaging
                            python-pyyaml
                            python-requests
                            python-tqdm))
   (native-inputs (list python-black
                        python-flake8
                        python-isort
                        python-pytest
                        python-pytest-xdist
                        python-tokenizers
                        python-torch
                        python-transformers))
   (home-page "https://github.com/huggingface/doc-builder")
   (synopsis "Doc building utility")
   (description "Doc building utility")
   (license #f)))

(define-public python-ipadic
  (package
   (name "python-ipadic")
   (version "1.0.0")
   (source
    (origin
     (method url-fetch)
     (uri (pypi-uri "ipadic" version))
     (sha256
      (base32 "01qwlzlm0ipnfrj3l3b4gcsb2rc6k7c2iv8qmz51l4x6xhqkv4pm"))))
   (build-system pyproject-build-system)
   (home-page "https://github.com/polm/ipadic-py")
   (synopsis "IPAdic packaged for Python")
   (description "IPAdic packaged for Python")
   (license #f)))

(define-public python-jax
  (package
   (name "python-jax")
   (version "0.4.23")
   (source
    (origin
     (method url-fetch)
     (uri (pypi-uri "jax" version))
     (sha256
      (base32 "119njs28qjdirk04vcaq89dv3xi3jwrfvsmjj4w806wdfmd9l8ia"))))
   (build-system pyproject-build-system)
   (propagated-inputs (list python-importlib-metadata python-ml-dtypes
                            python-numpy python-opt-einsum python-scipy))
   (home-page "https://github.com/google/jax")
   (synopsis "Differentiate, compile, and transform Numpy code.")
   (description "Differentiate, compile, and transform Numpy code.")
   (license #f)))

(define-public manga-ocr
  (package
   (name "manga-ocr")
   (version "0.1.11")
   (source
    (origin
     (method url-fetch)
     (uri (pypi-uri "manga-ocr" version))
     (sha256
      (base32 "1sxhk613ag1vk5hm6zxa0npjcn4gn0bgf1bq7id56qw44rlq1ki1"))))
   (build-system pyproject-build-system)
   (inputs (list python-fire
                 python-fugashi
                 python-jaconv
                 python-loguru
                 python-numpy
                 python-pillow
                 python-pyperclip
                 python-pytorch
                 python-transformers
                 python-unidic-lite))
   (home-page "https://github.com/kha-white/manga-ocr")
   (synopsis "OCR for Japanese manga")
   (description "OCR for Japanese manga")
   (license #f)))

manga-ocr
  (name "manga-ocr")
  (version "0.1.11")
  (sourcet
   (origin
     (method url-fetch)
     (uri (pypi-uri "manga-ocr" version))
     (sha256
      (base32 "1sxhk613ag1vk5hm6zxa0npjcn4gn0bgf1bq7id56qw44rlq1ki1"))))
  (build-system pyproject-build-system)
  (propagated-inputs (list python-fire
                           python-fugashi
                           python-jaconv
                           python-loguru
                           python-numpy
                           python-pillow
                           python-pyperclip
                           python-torch
                           python-transformers
                           python-unidic-lite))
  (home-page "https://github.com/kha-white/manga-ocr")
  (synopsis "OCR for Japanese manga")
  (description "OCR for Japanese manga")
  (license #f)))