v1.6.0 (#38)

* create C++ mask module with pybind

* add new cpp dataset resolver

* fix rleMerge

* add mask tests

* remove cython masks

* remove old mask api from setup

* is_crowd or iscrowd

* createSubIndex for extra_calc only

* fix lint

* setup print_function

* create init_as_pycocotools

* add basic usage

* bump version

* create test for iscrowd

* add toUncompressedRLE

* update docstrings

* add c++ _summarize. but its slow

* rework summarize

* fix gt or dt is empty

* segm & pose compare

* create calc_auc on C++. faster

* prep calc auc

* segmToRle over C++

* segmToRle h,w -> w,h

* append detectron2 tests

* update tests

* update kp test

* create torchmetric test

* add coverage report

* fix

* make more tests

* use_deepcopy for loadRes

* add matched check

* rework math_matches with C++

* speedup List Comp

* fix mask code

* fix mistake gt-dt

* add key_check

* set filter is faster

* rework unordered_map with hash_pair

* stable_sort -> faster

* speedup vectors

* create rle SLOW erode

* speedup toBoundary kernel

* create rleToBoundary

* push_back -> emplace_back

* add basic boundary test

* add opencv skip test

* fix mask api

* fix mask api test

* check on C++ side

* fix new pybind11

* Separated work with RLE and masks

* fix lint

* add boundary IoU support

* add coco_mapping for mmdet

* fix bug

* clean docstrings

* create to_dict warpers

* fix test

* add async rle calc

* fix rle merge

* Fixed filter boundaries

* create show anns

* pathlib.PosixPath support

* display_matrix text color fix

* fix docstrings

* fix TypeError

* v1.6.0 fix

.github/workflows/unittest.yml

@@ -37,4 +37,4 @@ jobs:
           pip install "${source_file}[tests]"
 
       - name: Run test cases
-        run: pytest tests/*
+        run: cd tests && pytest --cov=faster_coco_eval .

README.md

@@ -31,6 +31,26 @@ pip install faster-coco-eval
 pip install faster-coco-eval[extra]
 ```
 
+### Basic usage
+
+```py
+import faster_coco_eval
+
+# Replace pycocotools with faster_coco_eval
+faster_coco_eval.init_as_pycocotools()
+
+from pycocotools.coco import COCO
+from pycocotools.cocoeval import COCOeval
+
+anno = COCO(str(anno_json))  # init annotations api
+pred = anno.loadRes(str(pred_json))  # init predictions api (must pass string, not Path)
+
+val = COCOeval(anno, pred, "bbox")
+val.evaluate()
+val.accumulate()
+val.summarize()
+
+```
 
 ## Faster-COCO-Eval base