tracdap-runtime 0.5.30__py3-none-any.whl → 0.6.0__py3-none-any.whl

tracdap/rt/_exec/dev_mode.py

@@ -648,7 +648,8 @@ class DevModeTranslator:
             x_name = x_orig_path.name
 
             if x_storage.exists(str(x_orig_path.parent)):
-                existing_files = x_storage.ls(str(x_orig_path.parent))
+                listing = x_storage.ls(str(x_orig_path.parent))
+                existing_files = list(map(lambda stat: stat.file_name, listing))
             else:
                 existing_files = []
 

tracdap/rt/_impl/data.py

@@ -330,7 +330,7 @@ class DataMapping:
             DataConformance.check_duplicate_fields(table.schema.names, False)
 
         # Use Arrow's built-in function to convert to Pandas
-        df_from_arrow = table.to_pandas(
+        return table.to_pandas(
 
             # Mapping for arrow -> pandas types for core types
             types_mapper=cls.__ARROW_TO_PANDAS_TYPE_MAPPING.get,
@@ -346,33 +346,6 @@ class DataMapping:
             # This is a significant performance win for very wide datasets
             split_blocks=True)  # noqa
 
-        # Arrow 12 doesn't support the new precision handling for datetime values supported in Pandas 2
-        # However Arrow 13 dropped support for Python 3.7, which is a requirement for the TRAC 0.5.x series
-        # So to backport Pandas 2 support, special handling is needed for datetime fields when using Pandas 2
-        # This is not needed from TRAC 0.6 onward, which upgrades to Arrow 13 and drops Python 3.7 support
-        # Also it is not needed if the temporal objects flag is set, since it only affects NumPy datetime64
-
-        if cls.__PANDAS_MAJOR_VERSION == 2 and not temporal_objects_flag:
-            # Use table.schema, it is always present and has been normalized if a separate schema was supplied
-            return cls._fix_pandas_2_datetime_precision(df_from_arrow, table.schema)
-        else:
-            return df_from_arrow
-
-    @classmethod
-    def _fix_pandas_2_datetime_precision(cls, df: pd.DataFrame, schema: pa.Schema) -> pd.DataFrame:
-
-        for field in schema:
-            if pa.types.is_date(field.type):
-                dtype = cls.__PANDAS_DATE_TYPE
-                if df[field.name].dtype != dtype:
-                    df[field.name] = df[field.name].astype(dtype)
-            if pa.types.is_timestamp(field.type):
-                dtype = cls.__pandas_datetime_type(field.type.tz, field.type.unit)
-                if df[field.name].dtype != dtype:
-                    df[field.name] = df[field.name].astype(dtype)
-
-        return df
-
     @classmethod
     def pandas_to_arrow(cls, df: pd.DataFrame, schema: tp.Optional[pa.Schema] = None) -> pa.Table:
 

tracdap/rt/_impl/static_api.py

@@ -82,7 +82,11 @@ class StaticApiImpl(_StaticApiHook):
             param_type_descriptor = _meta.TypeDescriptor(param_type, None, None)
 
         if default_value is not None and not isinstance(default_value, _meta.Value):
-            default_value = _type_system.MetadataCodec.encode_value(default_value)
+            try:
+                default_value = _type_system.MetadataCodec.convert_value(default_value, param_type_descriptor)
+            except _ex.ETrac as e:
+                msg = f"Default value for parameter [{param_name}] does not match the declared type"
+                raise _ex.EModelValidation(msg) from e
 
         return _Named(param_name, _meta.ModelParameter(param_type_descriptor, label, default_value))