Picai-eval

What's Changed
* Bugfix for 21 by joeranbosma in https://github.com/DIAGNijmegen/picai_eval/pull/22
* Allow evaluating with no lesion candidates by joeranbosma in https://github.com/DIAGNijmegen/picai_eval/pull/23
* Add function to calculate the confidence intervals for the AUROC, AP and score

**Full Changelog**: https://github.com/DIAGNijmegen/picai_eval/compare/v1.4.8...v1.4.13

What's Changed
* Bugfix `lesion_TPR_at_FPR` by joeranbosma in https://github.com/DIAGNijmegen/picai_eval/pull/20


**Full Changelog**: https://github.com/DIAGNijmegen/picai_eval/compare/v1.4.7...v1.4.8

What's Changed
* Update documentation permutation test (AI vs AI) by joeranbosma in https://github.com/DIAGNijmegen/picai_eval/pull/11
* Evaluation progress bar by joeranbosma in https://github.com/DIAGNijmegen/picai_eval/pull/13
* Implement `lesion_TPR_at_FPR` by joeranbosma in https://github.com/DIAGNijmegen/picai_eval/pull/14
* Change image_utils.py to work with nnunetv2 predictions and ground truth labels with multi-class segmentations. by NataliaAlves13 in https://github.com/DIAGNijmegen/picai_eval/pull/17
* Account for new keyword in sklearn v1.5+ by joeranbosma in https://github.com/DIAGNijmegen/picai_eval/pull/19

New Contributors
* NataliaAlves13 made their first contribution in https://github.com/DIAGNijmegen/picai_eval/pull/17

**Full Changelog**: https://github.com/DIAGNijmegen/picai_eval/compare/v1.4.4...v1.4.7

* Improve robustness of initialising Metrics (https://github.com/DIAGNijmegen/picai_eval/pull/7)
* Evaluate softmax predictions from command line (https://github.com/DIAGNijmegen/picai_eval/pull/9)
* Single threaded evaluation (https://github.com/DIAGNijmegen/picai_eval/pull/10)
* Documentation/cleanup

Implemented linear sum assignment for lesion matching: Lesions candidates are matched to ground-truth lesions by optimising the number of matches first, and the total overlap second.

Improved saving of metrics to JSON.