2.5 KiB
Kapitanbooru Uploader
Kapitanbooru Uploader is a GUI application for uploading images to KapitanBooru. It provides features such as tag management, automatic tagging using wdtagger, and more.
Features
- Image Upload: Easily upload images to KapitanBooru.
- Tag Management: Manage tags for images, including automatic tagging using wdtagger.
- Cache Management: Cache results from wdtagger to speed up processing.
- Settings: Configure application settings such as default tags, cache expiry, and more.
Installation
-
Clone the repository:
git clone https://git.mlesniak.pl/kapitan/kapitanbooru-uploader.git cd kapitanbooru-uploader
-
Standard Installation (Supports most NVIDIA GPUs):
pip install .
-
For RTX 50xx (Blackwell GPUs) with CUDA 12 Nightly Support:
Requires thecuda12-nightly
extra and PyTorch nightly binaries:pip install ".[cuda12-nightly]" --extra-index-url https://download.pytorch.org/whl/nightly/cu128 --pre
Using
pipx
for isolated installation:pipx install "kapitanbooru-uploader[cuda12-nightly]" --pip-args "--extra-index-url https://download.pytorch.org/whl/nightly/cu128 --pre"
-
Verify Python version:
python --version # Should be >= 3.13
⚠️ Notes:
- The
cuda12-nightly
extra depends on pre-release PyTorch packages from the specified index. - Private PyPI repository (for
wdtagger[cuda12-nightly]
) must already be configured in your environment (no URL needed here).
Usage
-
Run the application:
python -m kapitanbooru_uploader.main
-
Select the folder containing the images you want to upload.
-
Manage tags and upload images using the GUI.
Configuration
Configuration settings can be found in the settings.json
file located in the application data directory. You can modify settings such as username, password, base URL, default tags, and more.
Development
Debugging
You can use the provided VSCode launch configurations to debug the application. Open the .vscode/launch.json
file and select the appropriate configuration.
Contributing
Contributions are welcome! Please fork the repository and submit a pull request.
License
This project is licensed under the MIT License. See the LICENSE
file for details.
Contact
For any questions or issues, please contact Michał Leśniak at kapitan@mlesniak.pl.