lint and test fix

Author: peterdudfield

nowcasting_datamodel/models/forecast.py

@@ -7,7 +7,7 @@
 """
 
 import logging
-from datetime import datetime
+from datetime import datetime, timezone
 from typing import List, Optional
 
 import numpy as np
@@ -106,12 +106,21 @@ def create_partition(
 
 
 def default_horizon_minutes(context):
+    """Make a default horizon minutes for the ForecastValueSQLMixin"""
     columns = context.get_current_parameters()
     target_time = columns["target_time"]
     created_utc = columns["created_utc"]
 
     if created_utc is None:
-        created_utc = datetime.now(tz=target_time.tzinfo)
+        created_utc = datetime.now(tz=timezone.utc)
+
+    if isinstance(target_time, str):
+        target_time = datetime.fromisoformat(target_time)
+
+    # make sure they both have tz, or not
+    if target_time.tzinfo is None:
+        created_utc = created_utc.replace(tzinfo=None)
+
     delta = (target_time - created_utc).total_seconds() / 60.0  # convert to minutes
     return delta
 
@@ -130,9 +139,11 @@ class ForecastValueSQLMixin(CreatedMixin):
     # Want to keep it as json so that we can store different properties for different forecasts
     properties = Column(MutableDict.as_mutable(JSON), nullable=True)
 
-    # we want to store the forecast horizon. This is the difference between target_time and created_utc.
-    # By storing the value, and index, queries should be faster when looking for an N hour forecast.
-    # I tried to use  server_default = "CAST(EXTRACT(EPOCH FROM (target_time - created_utc))/60 as INT)"
+    # we want to store the forecast horizon.
+    # This is the difference between target_time and created_utc.
+    # By storing the value queries should be faster when looking for an N hour forecast.
+    # I tried to use
+    # server_default = "CAST(EXTRACT(EPOCH FROM (target_time - created_utc))/60 as INT)"
     # but this does not work.
     horizon_minutes = Column(
         Integer,