Merge pull request #13 from FZJ-INM1-BDA/test_addNameDuplCheck

Check duplicate names and IDs in parcellations

.github/workflows/check-region-id-uniqueness.yml

@@ -0,0 +1,39 @@
+name: '[test] check region id uniqueness'
+
+on:
+  push:
+    paths:
+    - 'parcellations/**.json'
+
+jobs:
+  check_region_id_uniqueness:
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v4
+    - name: Set up Python 3.10
+      uses: actions/setup-python@v4
+      with:
+        python-version: '3.10'
+
+    - name: checkout siibra-python
+      uses: actions/checkout@v4
+      with:
+        repository: FZJ-INM1-BDA/siibra-python
+        path:  siibra-python-${{ github.run_id }}-${{ github.run_number }}
+        fetch-depth: 1
+        clean: True
+        ref: 'main'
+
+    - name: move siibra-python one up from workspace
+      run: mv siibra-python-${{ github.run_id }}-${{ github.run_number }} ../siibra-python
+
+    - name: Install siibra-python testing requirements
+      run: |
+          pip install -r ../siibra-python/requirements-test.txt
+          pip install -r ../siibra-python/requirements.txt
+    
+    - name: region id uniqueness test
+      run: |
+        export SIIBRA_USE_CONFIGURATION='./'
+        pytest ../siibra-python/e2e/core/test_parcellation.py::test_parc_id_uniqueness
+        pytest ../siibra-python/e2e/core/test_parcellation.py::test_region_id_uniqueness

.github/workflows/test-configuration.yml

@@ -91,3 +91,18 @@ jobs:
           pip install -r _ci/requirements.txt
     - name: Verify neuroglancer urls
       run: python _ci/verify_neuroglancer_urls.py
+
+  ensure_no_duplicated_region_name:
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v4
+    - name: Set up Python 3.10
+      uses: actions/setup-python@v4
+      with:
+        python-version: '3.10'
+    - name: Install requirements
+      run: |
+          pip install -U pip
+          pip install -r _ci/requirements.txt
+    - name: Verify no duplicated region name
+      run: python _ci/ensure_no_duplicated_region_name.py

_ci/ensure_no_duplicated_region_name.py

@@ -0,0 +1,77 @@
+from pathlib import Path
+import json
+from dataclasses import dataclass
+
+@dataclass
+class DuplicatedErr:
+    filename: str
+    key: str
+
+    def __eq__(self, other: "DuplicatedErr"):
+        return self.filename == other.filename and self.key == other.key
+
+
+class DuplicatedRegionName(Exception):
+    pass
+
+
+FAIL_FAST = False
+
+# There is no solution for the following 
+XFAIL_LIST = [
+    DuplicatedErr(filename='parcellations\\rat_waxholmv1_01.json', key='corticofugal pathways'),
+    DuplicatedErr(filename='parcellations\\rat_waxholmv1_01.json', key='medial lemniscus'),
+    DuplicatedErr(filename='parcellations\\rat_waxholmv1_01.json', key='facial nerve'),
+    DuplicatedErr(filename='parcellations\\rat_waxholmv1_01.json', key='spinal cord'),
+    DuplicatedErr(filename='parcellations\\rat_waxholmv2.json', key='corticofugal pathways'),
+    DuplicatedErr(filename='parcellations\\rat_waxholmv2.json', key='medial lemniscus'),
+    DuplicatedErr(filename='parcellations\\rat_waxholmv2.json', key='facial nerve'),
+    DuplicatedErr(filename='parcellations\\rat_waxholmv2.json', key='spinal cord'),
+    DuplicatedErr(filename='parcellations\\rat_waxholmv3.json', key='anterior commissure'),
+    DuplicatedErr(filename='parcellations\\rat_waxholmv3.json', key='medial lemniscus'),
+    DuplicatedErr(filename='parcellations\\rat_waxholmv3.json', key='facial nerve'),
+    DuplicatedErr(filename='parcellations\\rat_waxholmv3.json', key='lateral lemniscus'),
+    DuplicatedErr(filename='parcellations\\rat_waxholmv3.json', key='neocortex'),
+    DuplicatedErr(filename='parcellations\\rat_waxholmv3.json', key='thalamus'),
+    DuplicatedErr(filename='parcellations\\rat_waxholmv3.json', key='brainstem'),
+    DuplicatedErr(filename='parcellations\\rat_waxholmv3.json', key='ventricular system'),
+    DuplicatedErr(filename='parcellations\\rat_waxholmv3.json', key='spinal cord'),
+]
+
+def iterate_over_children(r):
+    yield r
+    for c in r.get("children", []):
+        for cc in iterate_over_children(c):
+            yield cc
+
+
+def main():
+    errors = []
+    _dir = Path("parcellations")
+    for _file in _dir.glob("*.json"):
+        reg_name_set = set()
+        with open(_file, "r", encoding="utf-8") as fp:
+            parc_json = json.load(fp)
+            all_regions = [
+                rc for r in parc_json.get("regions") for rc in iterate_over_children(r)
+            ]
+            for reg in all_regions:
+                reg_name = reg.get("name")
+                if reg_name.lower() in reg_name_set:
+                    error = DuplicatedErr(filename=str(_file), key=reg_name)
+                    if error in XFAIL_LIST:
+                        print(f"xfail: {error}")
+                        continue
+                    errors.append(error)
+                    if FAIL_FAST:
+                        raise DuplicatedRegionName(
+                            "\n".join([str(err) for err in errors])
+                        )
+                reg_name_set.add(reg_name.lower())
+
+    if len(errors) > 0:
+        raise DuplicatedRegionName("\n".join([str(err) for err in errors]))
+
+
+if __name__ == "__main__":
+    main()