fix(simple_metrics): ignore invalid simple_metrics_min_samples; tighten tests and docs (SIENTIAPDE-2189)

README.md

@@ -487,7 +487,7 @@ The **SimpleMetrics** workflow calculates regression metrics (RMSE, MSE, MAE, R
 
 #### Execution Flow
 1. **Data Loading**: Loads prediction vs target data via a JOIN query
-2. **Metrics Calculation**: Calculates configured regression metrics
+2. **Metrics Calculation**: Calculates configured regression metrics; alerts on `simple_metrics_thresholds` breaches (one alert per model, the message names the window) unless the window has fewer rows than `simple_metrics_min_samples`
 3. **Data Export**: Exports metrics to PostgreSQL
 
 #### Architecture Diagram
@@ -510,7 +510,12 @@ flowchart LR
   "data_table_name": "laborious_data",
   "target_table_name": "simple_metrics",
   "interval_minutes": 60,
-  "model_config": { "target": "temperature", "alias": "production" },
+  "model_config": {
+    "target": "temperature",
+    "alias": "production",
+    "simple_metrics_thresholds": { "rmse_max": 1.5 },
+    "simple_metrics_min_samples": 10
+  },
   "metrics": ["rmse", "mse", "mae", "r2"]
 }
 ```

laborious/workflows/simple_metrics.py

@@ -34,6 +34,14 @@ class SimpleMetrics:
         model_type = model_config.get('model_type')
         thresholds = model_config.get('simple_metrics_thresholds')
         min_samples = model_config.get('simple_metrics_min_samples')
+        # Untyped Mongo config: anything but a positive int (bool excluded) turns the guard off
+        # instead of raising in workflow code; falsy values mean "unset" and stay silent.
+        if type(min_samples) is not int or min_samples <= 0:
+            if min_samples:
+                workflow.logger.warning(
+                    f'Invalid simple_metrics_min_samples={min_samples!r}: min samples guard off'
+                )
+            min_samples = None
 
         # Built from internal Temporal workflow config, not user input (bandit B608).
         query = f"""

tests/laborious/workflows/test_simple_metrics.py

@@ -237,15 +237,16 @@ def _input(**model_config) -> dict:
 _ROW = {'timestamp': '2026-10-06 10:00:00+0000', 'prediction': 1.0, 'target': 2.0}
 
 
+_METRICS = [{'metric': 'rmse', 'value': 9.0}]
+
+
 @mark.asyncio
+@mark.parametrize('n_rows', [2, 4])
 @patch('laborious.workflows.simple_metrics.workflow', new_callable=AsyncMock)
-async def test_run_below_min_samples_drops_thresholds(workflow_mock, simple_metrics):
+async def test_run_below_min_samples_drops_thresholds(workflow_mock, simple_metrics, n_rows):
     workflow_mock.logger = MagicMock()
-    # load, export, and the adaptive-state call S2 (Task 8) adds.
-    workflow_mock.execute_activity_method = AsyncMock(side_effect=[[_ROW] * 2, None, None])
-    workflow_mock.execute_local_activity_method = AsyncMock(
-        return_value=[{'metric': 'rmse', 'value': 9.0}]
-    )
+    workflow_mock.execute_activity_method = AsyncMock(side_effect=[[_ROW] * n_rows, None])
+    workflow_mock.execute_local_activity_method = AsyncMock(return_value=_METRICS)
 
     await simple_metrics.run(
         _input(simple_metrics_thresholds={'rmse_max': 1.0}, simple_metrics_min_samples=5)
@@ -253,16 +254,19 @@ async def test_run_below_min_samples_drops_thresholds(workflow_mock, simple_metr
 
     activity_input = workflow_mock.execute_local_activity_method.call_args.args[1]
     assert activity_input['thresholds'] is None
+    # Persist, don't alert.
+    assert workflow_mock.execute_activity_method.call_count == 2
+    export_call = workflow_mock.execute_activity_method.call_args_list[1]
+    assert export_call.args[0] == Activities.export_data_to_postgres
+    assert export_call.args[1]['data'] == _METRICS
 
 
 @mark.asyncio
 @patch('laborious.workflows.simple_metrics.workflow', new_callable=AsyncMock)
 async def test_run_at_min_samples_keeps_thresholds(workflow_mock, simple_metrics):
     workflow_mock.logger = MagicMock()
-    workflow_mock.execute_activity_method = AsyncMock(side_effect=[[_ROW] * 5, None, None])
-    workflow_mock.execute_local_activity_method = AsyncMock(
-        return_value=[{'metric': 'rmse', 'value': 9.0}]
-    )
+    workflow_mock.execute_activity_method = AsyncMock(side_effect=[[_ROW] * 5, None])
+    workflow_mock.execute_local_activity_method = AsyncMock(return_value=_METRICS)
 
     await simple_metrics.run(
         _input(simple_metrics_thresholds={'rmse_max': 1.0}, simple_metrics_min_samples=5)
@@ -270,3 +274,23 @@ async def test_run_at_min_samples_keeps_thresholds(workflow_mock, simple_metrics
 
     activity_input = workflow_mock.execute_local_activity_method.call_args.args[1]
     assert activity_input['thresholds'] == {'rmse_max': 1.0}
+    assert workflow_mock.execute_activity_method.call_count == 2
+
+
+@mark.asyncio
+@mark.parametrize(
+    ('min_samples', 'warns'), [(None, False), (0, False), ('5', True), (5.0, True), (True, True)]
+)
+@patch('laborious.workflows.simple_metrics.workflow', new_callable=AsyncMock)
+async def test_run_guard_off_keeps_thresholds(workflow_mock, simple_metrics, min_samples, warns):
+    workflow_mock.logger = MagicMock()
+    workflow_mock.execute_activity_method = AsyncMock(side_effect=[[_ROW] * 2, None])
+    workflow_mock.execute_local_activity_method = AsyncMock(return_value=_METRICS)
+
+    await simple_metrics.run(
+        _input(simple_metrics_thresholds={'rmse_max': 1.0}, simple_metrics_min_samples=min_samples)
+    )
+
+    activity_input = workflow_mock.execute_local_activity_method.call_args.args[1]
+    assert activity_input['thresholds'] == {'rmse_max': 1.0}
+    assert workflow_mock.logger.warning.call_count == int(warns)