Use defaultdict for _aim_metadata_dtype for pd.read_csv

Author: leehart

malariagen_data/anoph/sample_metadata.py

@@ -41,7 +41,8 @@ def __init__(
         # data resources, and so column names and dtype need to be
         # passed in as parameters.
         self._aim_metadata_columns: Optional[List[str]] = None
-        self._aim_metadata_dtype: Dict[str, Any] = dict()
+        #  `dtype` of `dict[str, Any]` is incompatible with `read_csv`
+        self._aim_metadata_dtype: defaultdict[str, Any] = dict()
         if isinstance(aim_metadata_dtype, Mapping):
             self._aim_metadata_columns = list(aim_metadata_dtype.keys())
             self._aim_metadata_dtype.update(aim_metadata_dtype)