WIP: amend data types

Author: leehart

malariagen_data/anoph/base.py

@@ -405,14 +405,14 @@ def _relevant_releases(self) -> Tuple[str, ...]:
 
         if self._cache_releases is None:
             # Start a list of the relevant releases.
-            relevant_releases = []
+            relevant_releases = []  # type: List[str]
 
             # Get the available releases, which depends on the `pre` setting.
             available_releases = self._available_releases
 
             # If there are no criteria, then all available releases are relevant.
             if not self._unrestricted_use_only and not self._surveillance_use_only:
-                relevant_releases = available_releases
+                relevant_releases = list(available_releases)
 
             elif self._unrestricted_use_only and not self._surveillance_use_only:
                 # Get the releases with unrestricted data.
@@ -494,6 +494,9 @@ def client_location(self) -> str:
             location = "unknown"
         return location
 
+    def _surveillance_flags(self, sample_sets: List[str]):
+        raise NotImplementedError("Subclasses must implement `_surveillance_flags`.")
+
     def _release_has_unrestricted_data(self, *, release: str):
         """Return `True` if the specified release has any unrestricted data. Otherwise return `False`."""
 
@@ -526,7 +529,6 @@ def _release_has_surveillance_data(self, *, release: str):
         sample_sets = sample_sets_manifest_df["sample_set"].to_list()
 
         # Determine whether any of the sample sets have surveillance data.
-        # Note: rather than using `_surveillance_flags`, to avoid unnecessary processing, we only need to find one sample set.
         release_has_surveillance_data = False
         for sample_set in sample_sets:
             if self._sample_set_has_surveillance_data(sample_set=sample_set):

malariagen_data/anoph/sample_metadata.py

@@ -90,13 +90,13 @@ def _parse_metadata_paths(
         self,
         path_template: str,
         parse_metadata_func: Callable[[str, Union[bytes, Exception]], pd.DataFrame],
-        sample_sets: Optional[base_params.sample_sets] = None,
+        sample_sets: List[str],
         aim_analysis: Optional[str] = None,
         cohorts_analysis: Optional[str] = None,
     ) -> pd.DataFrame:
-        # Warning: don't use `_prep_sample_sets_param` in this function because that can cause a circular dependency, eventually raising a RecursionError.
-        # For instance, `_prep_sample_sets_param` uses `_relevant_sample_sets`, which uses `_surveillance_flags, which uses `_parse_metadata_paths`.
-        # Instead, use `_prep_sample_sets_param` to prepare `sample_sets` before passing it to this function.
+        # Note: we don't use `_prep_sample_sets_param` in this function because that can cause a circular dependency, eventually raising a `RecursionError`.
+        # For instance, `_prep_sample_sets_param` uses `_relevant_sample_sets`, which uses `_surveillance_flags`, which uses `_parse_metadata_paths`.
+        # Instead, use `_prep_sample_sets_param` to prepare `sample_sets` as a `List[str]` before passing it to this function.
 
         # Obtain paths for all files we need to fetch.
         file_paths: Mapping[str, str] = self._metadata_paths(
@@ -410,10 +410,10 @@ def _parse_surveillance_flags(
         `is_surveillance` indicates whether the sample can be used for surveillance,
         """,
     )
-    def _surveillance_flags(self, sample_sets: base_params.sample_sets) -> pd.DataFrame:
-        # Warning: don't use `_prep_sample_sets_param` here, because `_prep_sample_sets_param` uses `_relevant_sample_sets`,
-        # which uses this function, which would cause a RecursionError due to cyclic dependency.
-        # Instead, prepare the `sample_sets` parameter before calling this function.
+    def _surveillance_flags(self, sample_sets: List[str]) -> pd.DataFrame:
+        # Note: we don't use `_prep_sample_sets_param` in this function because that can cause a circular dependency, eventually raising a `RecursionError`.
+        # For instance, `_prep_sample_sets_param` uses `_relevant_sample_sets`, which uses `_surveillance_flags`.
+        # Instead, use `_prep_sample_sets_param` to prepare `sample_sets` as a `List[str]` before passing it to this function.
 
         return self._parse_metadata_paths(
             path_template="{release_path}/metadata/general/{sample_set}/surveillance.flags.csv",