NadeOCR

An easy and fast-to-use tool for scanning text anywhere with Google's Vision API and other third party services. This tool is mostly designed for Japanese learners, but should work for anyone else.

Features

• Compatible with Linux and Windows.
• Copy your scanned text immediately to your clipboard.
• Custom hotkeys for scanning quickly and anywhere on your screen.
• Supports a variety of OCR models, including Google's Vision API and Manga-OCR.
• Settings for removing new lines, copying line by line or using your own Regex filter.

Installation

Windows

• Download the latest zip file here.
• Decompress the file in the desired directory.
• Open the extracted folder and execute NadeOCR.exe.

Note: you must have your google credentiales file credentials.json in order to use Google provider. More information here.

PIP (Windows/Linux)

• To install NadeOCR, run in command line:

pip install nadeocr

• To execute NadeOCR, run in command line:

nadeocr

How to build

Distribution files

You can build and generate distribution archives by using this command in the project's root:

python -m build

Windows

You can also generate an executable for any Windows machine higher than Windows 8 by using PyInstaller with this configuration:

pyinstaller --noconfirm --onedir --windowed --icon "NadeOCR/nadeocr/resources/assets/icon.ico" --name "NadeOCR_v1.0.0" --clean --add-data "NadeOCR/nadeocr;nadeocr/" --additional-hooks-dir "NadeOCR/nadeocr/resources/hooks" --hidden-import "huggingface_hub.repository" --hidden-import "huggingface_hub.hf_api" --collect-data "torch" --copy-metadata "torch" --copy-metadata "tqdm" --copy-metadata "regex" --copy-metadata "requests" --copy-metadata "packaging" --copy-metadata "filelock" --copy-metadata "numpy" --copy-metadata "tokenizers" --copy-metadata "importlib" --collect-data "unidic_lite" --hidden-import "unidic_lite" --collect-data "manga_ocr" --collect-data "toml"  "NadeOCR/nadeocr/main.py"

Contributing

All contributions are appreciated. You are welcome to contribute whatever you think will be helpful, so feel free to create an issue or submit a pull request for review and discussion.

Please email me at jonathan.197ariza@gmail.com if you have any questions about the codebase.

Acknowledgements

This project wouldn't be possible without:

• Google's Vision API for detecting and recognising a wide variety of languages including, but not limited to, English, Japanese and Spanish.
• The awesome Manga-OCR model by Maciej Budyś for recognizing Japanese characters in manga.

Things to do

High priority

• Add support to PaddleOCR.
• Add multi-language (English) support to the PyQT GUI.
• Allow custom filters/colour/contrast for cropped images.
• Add Regex support.

Low priority

• Add Android support.
• Add support for running on startup (Windows/Linux).

License

This software is licensed under the GPLv3 (see [LICENSE) and uses third party libraries that are distributed under their own terms (see LICENSE-3RD-PARTY).