SIENTIAPDE-1646

Simplify merge in retrain: expose timestamp explicitly for join

Refactor evaluation_data merge to reset both frames' index into a
"timestamp" column before joining, instead of merging on index then
copying it out afterward. Same validate='one_to_one' safety check,
fewer intermediate steps.

laborious/activities/mlflow.py

@@ -531,23 +531,19 @@ class MLFlow(SientiaMonitoring):
             # Keep heavy model fitting outside MLflow run timing.
             prediction_data = wrapper.retrain(data).rename(columns={target: 'prediction'})
 
-            # Merge prediction data with retrain data. Both sides are indexed by the pivoted
+            # Both frames are indexed by the pivoted timestamp: expose it as an explicit column
-            # timestamp, so one row per timestamp on each side; `validate` fails loudly instead
+            # and join on it. `validate` fails loudly instead of silently multiplying rows if the
-            # of silently multiplying rows if the wrapper ever returns duplicated timestamps.
+            # wrapper ever returns duplicated timestamps.
-            evaluation_data = pd.merge(
+            evaluation_data = (
-                data,
+                data.reset_index(names='timestamp')
-                prediction_data,
+                .merge(
-                left_index=True,
+                    prediction_data.reset_index(names='timestamp'),
-                right_index=True,
+                    on='timestamp',
-                how='left',
+                    how='left',
-                validate='one_to_one',
+                    validate='one_to_one',
-            )
+                )
-
+                .rename(columns={target: 'target'})
-            # Rename target column to "target", then move the index into a "timestamp" column
+                .sort_values(by='timestamp', ascending=True, ignore_index=True)
-            evaluation_data = evaluation_data.rename(columns={target: 'target'})
-            evaluation_data['timestamp'] = evaluation_data.index
-            evaluation_data = evaluation_data.reset_index(drop=True).sort_values(
-                by='timestamp', ascending=True
             )
 
             run_name = f'{model_name}-retrain-{datetime.now().strftime("%Y%m%d%H%M%S")}'