Merge pull request #1207 from khushthecoder/fix/issue-1206-surveillance-flags-graceful-fallback

fix: replace NotImplementedError in _surveillance_flags with graceful fallback

Author: jonbrenas

malariagen_data/anoph/base.py

@@ -1,4 +1,5 @@
 import os
+import warnings
 
 import json
 from contextlib import nullcontext
@@ -496,7 +497,20 @@ def client_location(self) -> str:
         return location
 
     def _surveillance_flags(self, sample_sets: List[str]):
-        raise NotImplementedError("Subclasses must implement `_surveillance_flags`.")
+        """Return surveillance flags for sample sets. Subclasses should override to
+        load real data; this base implementation returns empty data and warns.
+        """
+        warnings.warn(
+            "Surveillance flags not implemented for this resource; returning empty data.",
+            UserWarning,
+            stacklevel=2,
+        )
+        return pd.DataFrame(
+            {
+                "sample_id": pd.Series(dtype="object"),
+                "is_surveillance": pd.Series(dtype="boolean"),
+            }
+        )
 
     def _release_has_unrestricted_data(self, *, release: str):
         """Return `True` if the specified release has any unrestricted data. Otherwise return `False`."""

tests/anoph/test_base.py

@@ -411,3 +411,28 @@ def test_sample_sets_no_terms_of_use(ag3_sim_fixture):
     finally:
         for mp, bp in zip(manifest_paths, backups):
             shutil.move(bp, mp)
+
+
+class TestSurveillanceFlagsBaseFallback:
+    """Tests for issue #1206: base _surveillance_flags graceful fallback."""
+
+    def test_surveillance_flags_base_returns_empty_and_warns(self, ag3_sim_api):
+        """Base implementation returns empty DataFrame with correct schema and warns."""
+        with pytest.warns(UserWarning, match="Surveillance flags not implemented"):
+            df = ag3_sim_api._surveillance_flags(sample_sets=["AG1000G-AO"])
+
+        assert isinstance(df, pd.DataFrame)
+        assert list(df.columns) == ["sample_id", "is_surveillance"]
+        assert df["sample_id"].dtype == object
+        assert pd.api.types.is_bool_dtype(df["is_surveillance"])
+        assert len(df) == 0
+
+    def test_sample_set_has_surveillance_data_returns_false_when_fallback(
+        self, ag3_sim_api
+    ):
+        """_sample_set_has_surveillance_data returns False when base fallback is used."""
+        with pytest.warns(UserWarning, match="Surveillance flags not implemented"):
+            result = ag3_sim_api._sample_set_has_surveillance_data(
+                sample_set="AG1000G-AO"
+            )
+        assert not result