Quadra

Fixed

- Fix classification val_dataloader shuffling data when it shouldn't

Updated

- Remove black from pre-commit hooks
- Use ruff as the main formatter and linting tool
- Upgrade mypy version
- Upgrade mlflow version
- Apply new pre-commits to all tests
- Update most of typing to py310 style

Updated

- Update anomalib to v0.7.0+obx.1.3.2

Updated

- Update anomalib to v0.7.0+obx.1.3.1
- The optimal anomaly threshold is now computed as the average between the max good and min bad score when the F1 is 1

Updated

- Anomaly test task now exports results based on the normalized anomaly scores instead of the raw scores. The normalized anomaly scores and the optimal threshold are computed based on the training threshold of the model.

Updated

- Change the way anomaly scores are normalized by default, instead of using a [0-1] range with a 0.5 threshold, the scores are now normalized to a [0-1000] range with a threshold of 100, the new score represents the distance from the selected threshold, for example, a score of 200 means that the anomaly score is 100% of the threshold above the threshold itself, a score of 50 means that the anomaly score is 50% of the threshold below.
- Change the default normalization config name for anomaly from `min_max_normalization` to `score_normalization`.

Fixed

- Fix the output heatmaps and preditions of anomaly inference tasks not being saved properly when images belonged to
different classes but had the same name.