Allow for missing data in input metadata table.

Warn that the column is missing.

allensdk/brain_observatory/behavior/behavior_project_cache/tables/metadata_table_schemas.py

@@ -12,48 +12,67 @@ class BehaviorSessionMetadataSchema(RaisingSchema):
     age_in_days = Int(required=True, description="Subject age")
     behavior_session_id = Int(
         required=True,
+        allow_none=True,
         description=(
             "Unique identifier for the "
             "behavior session to write into "
             "NWB format"
         ),
     )
     cre_line = String(
-        required=True, description="Genetic cre line of the subject."
+        required=True,
+        allow_none=True,
+        description="Genetic cre line of the subject."
     )
     date_of_acquisition = String(
         required=True,
+        allow_none=True,
         description=(
             "Date of acquisition of " "behavior session, in string " "format"
         ),
     )
     driver_line = List(
         String,
         required=True,
+        allow_none=True,
         cli_as_single_argument=True,
         description="Genetic driver line(s) of subject",
     )
     equipment_name = String(
-        required=True, description=("Name of the equipment used.")
+        required=True,
+        allow_none=True,
+        description=("Name of the equipment used.")
     )
     full_genotype = String(
-        required=True, description="Full genotype of subject"
+        required=True,
+        allow_none=True,
+        description="Full genotype of subject"
     )
     mouse_id = String(
         required=True,
+        allow_none=True,
         description="LabTracks ID of the subject. aka external_specimen_name.",
     )
     project_code = String(
         rquired=True,
+        allow_none=True,
         description="LabTracks ID of the subject. aka external_specimen_name.",
     )
     reporter_line = String(
-        required=True, description="Genetic reporter line(s) of subject"
+        required=True,
+        allow_none=True,
+        description="Genetic reporter line(s) of subject"
     )
     session_type = String(
-        required=True, description="Full name of session type."
+        required=True,
+        allow_none=True,
+        description="Full name of session type."
+    )
+    sex = String(
+        required=True,
+        allow_none=True,
+        description="Subject sex"
     )
-    sex = String(required=True, description="Subject sex")
 
     @mm.post_load
     def convert_date_time(self, data, **kwargs):

allensdk/brain_observatory/behavior/write_nwb/behavior/schemas.py

@@ -1,4 +1,6 @@
 import marshmallow as mm
+from warnings import warn
+import pandas as pd
 from allensdk.brain_observatory.argschema_utilities import (
     InputFile,
     RaisingSchema,
@@ -59,24 +61,70 @@ def _get_behavior_metadata(self, bs_row):
         """ """
         behavior_session_metadata = {}
 
-        behavior_session_metadata["age_in_days"] = bs_row["age_in_days"]
-        behavior_session_metadata["cre_line"] = bs_row["cre_line"]
-        behavior_session_metadata["date_of_acquisition"] = bs_row[
-            "date_of_acquisition"
-        ]
-        behavior_session_metadata["driver_line"] = sorted(
-            bs_row["driver_line"]
+        behavior_session_metadata["age_in_days"] = self._retrieve_column(
+            bs_row=bs_row, column_name="age_in_days"
+        )
+        behavior_session_metadata["cre_line"] = self._retrieve_column(
+            bs_row=bs_row, column_name="cre_line"
+        )
+        behavior_session_metadata["date_of_acquisition"] = self._retrieve_column(  # noqa: E501
+            bs_row=bs_row, column_name="date_of_acquisition"
+        )
+        behavior_session_metadata["driver_line"] = self._retrieve_column(
+            bs_row=bs_row, column_name="driver_line"
+        )
+        behavior_session_metadata["equipment_name"] = self._retrieve_column(
+            bs_row=bs_row, columns_name="equipment_name"
+        )
+        behavior_session_metadata["full_genotype"] = self._retrieve_column(
+            bs_row=bs_row, column_name="full_genotype"
+        )
+        behavior_session_metadata["mouse_id"] = self._retrieve_column(
+            bs_row=bs_row, column_name="mouse_id"
+        )
+        behavior_session_metadata["project_code"] = self._retrieve_column(
+            bs_row=bs_row, column_name="project_code"
+        )
+        behavior_session_metadata["reporter_line"] = self._retrieve_column(
+            bs_row=bs_row, column_name="reporter_line"
+        )
+        behavior_session_metadata["session_type"] = self._retrieve_column(
+            bs_row=bs_row, column_name="session_type"
+        )
+        behavior_session_metadata["sex"] = self._retrieve_column(
+            bs_row=bs_row, column_name="sex"
         )
-        behavior_session_metadata["equipment_name"] = bs_row["equipment_name"]
-        behavior_session_metadata["full_genotype"] = bs_row["full_genotype"]
-        behavior_session_metadata["mouse_id"] = bs_row["mouse_id"]
-        behavior_session_metadata["project_code"] = bs_row["project_code"]
-        behavior_session_metadata["reporter_line"] = bs_row["reporter_line"]
-        behavior_session_metadata["session_type"] = bs_row["session_type"]
-        behavior_session_metadata["sex"] = bs_row["sex"]
 
         return behavior_session_metadata
 
+    def _retrieve_column(self, bs_row: pd.Series, column_name: str):
+        """Pull a column safely, return None otherwise.
+
+        Parameters
+        ----------
+        bs_row : pd.Series
+            Row of a BehaviorSessionTable
+        column_name : str
+            Name of column to retrieve
+
+        Returns
+        -------
+        value : object
+            Value of column_name in bs_row, or None if column_name is not in
+            bs_row
+        """
+        if column_name not in bs_row.index:
+            warn(f"Warning, {column_name} not in metadata table. Unless this "
+                 "has been added it to skip_metadata_key or "
+                 "skip_stimulus_file_key as inpput, creating the NWB file "
+                 "fail.")
+            return None
+        else:
+            value = bs_row[column_name]
+            if isinstance(value, list):
+                value = sorted(value)
+            return value
+
 
 class BehaviorInputSchema(BaseInputSchema):
     behavior_session_id = Int(