feat: improve type hints

.github/workflows/cruft.yml

@@ -61,6 +61,6 @@ jobs:
           title: "ci - update cruft"
           branch: "update-cruft"
           body: "🌲 Cruft updates"
-          token: ${{ secrets.PAT }}
+          token: ${{ secrets.PAT }} 
           commit-message: "ci: update cruft"
           labels: "dependencies" # This makes the PR exempt from the stale bot

src/timeseriesflattener/feature_cache/cache_to_disk.py

@@ -2,7 +2,6 @@
 import datetime as dt
 import os
 from pathlib import Path
-from typing import Optional
 
 import pandas as pd
 
@@ -17,7 +16,7 @@ class DiskCache(FeatureCache):
     def __init__(
         self,
         feature_cache_dir: Path,
-        prediction_times_df: Optional[pd.DataFrame] = None,
+        prediction_times_df: pd.DataFrame,
         pred_time_uuid_col_name: str = "pred_time_uuid",
         entity_id_col_name: str = "entity_id",
         timestamp_col_name: str = "timestamp",
@@ -131,7 +130,7 @@ def read_feature(self, feature_spec: TemporalSpec) -> pd.DataFrame:
 
         # Replace NaNs with fallback
         df[feature_spec.get_col_str()] = df[feature_spec.get_col_str()].fillna(
-            feature_spec.fallback,
+            feature_spec.fallback,  # type: ignore
         )
 
         return df
@@ -145,7 +144,7 @@ def write_feature(
         file_name = self._get_file_name(feature_spec=feature_spec)
 
         # Drop rows containing fallback, since it's non-informative
-        df = df[df[feature_spec.get_col_str()] != feature_spec.fallback].dropna()
+        df = df[df[feature_spec.get_col_str()] != feature_spec.fallback].dropna()  # type: ignore
 
         # Drop entity and timestamp columns if they exists
         for col in [self.entity_entity_id_col_name, self.timestamp_col_name]:

src/timeseriesflattener/feature_spec_objects.py

@@ -233,7 +233,7 @@ class Doc:
         description="""Optional kwargs for the values_loader.""",
     )
 
-    values_df: Optional[pd.DataFrame] = Field(
+    values_df: Optional[pd.DataFrame] = Field(  # type: ignore
         default=None,
         description="Dataframe with the values.",
     )
@@ -778,8 +778,8 @@ class Doc:
 
     def _check_loaders_are_valid(self):
         """Check that all loaders can be resolved from the data_loaders catalogue."""
-        invalid_loaders = list(
-            set(self.values_loader) - set(data_loaders.get_all().keys()),
+        invalid_loaders: list = list(
+            set(self.values_loader) - set(data_loaders.get_all().keys()),  # type: ignore
         )
         if len(invalid_loaders) != 0:
             # New line variable as f-string can't handle backslashes