Migration notes
Existing users have to migrate their `config.json` to the updated `config.toml` file. An example file `config_sample.toml` is provided with all options.
It's also recommended to clone the repo from scratch instead of pulling the changes since I altered the git history.
Changelog
* Rename project from szurubooru-scripts to szuru-toolkit (thanks to Manicarus for the name suggestion)
* Rename upload-images to upload-media
* This project now uses Poetry to manage its dependencies and packaging
* Introduce pre-commit to check commits before pushing
* This project is now available on PyPI: https://pypi.org/project/szuru-toolkit
* If the package is installed via pip/PyPi, the scripts `auto-tagger`, `upload-media` and `create-tags` get added to pips `bin` path
* Scripts can now be invoked simply with e.g. `auto-tagger "date:today"`. See the updated `README` for more examples.
* The auto-tagger script is now faster with large queries, since results are now being yielded. Before, all results were being parsed before tagging started.
* The auto-tagger script has two new command line arguments: `--add-tags` and `--remove-tags`
* As the names suggest, you can add/remove your own tags which will be applied to all posts matching your query. Tags specified with `--remove-tags` will be applied after tagging has finished, so you can remove potential unwanted tags returned from SauceNAO/Deepbooru.
* Use a TOML based config file (`config.toml`) now instead of a JSON based one (`config.json`)
* See the `README` for full reference
* Implement logging (to a log file) with loguru. Can be configured in `config.toml`. (6)
* Improved error handling (6)
* Fixed issues with unused Deepbooru/Tensorflow imports (9)
* Added support for tagging with Konachan and Yandere
* If the post is being tagged with Deepbooru and has relations, copy `series`, `artist` and `character` tags from these relation posts (useful for Fanbox sets where the main post is uploaded to a booru but not the variations)
* Added support to tag posts with SauceNAO and Deepbooru (option `deepbooru_forced`)
* Szuru token does now not have to be converted to base64 beforehand (2, 3)
* Start using pyszuru package for interacting with the szuru API
* Majority of the szuru API calls are still using my own implementation because of speed issues
* Code and wording improvements
* If the SauceNAO API limit has been hit, the script continues to tag with Deepbooru if `deepbooru_enabled` is set