Add exclude_detection_period_from_training flag to freshness anomaly tests

integration_tests/tests/test_freshness_anomalies.py

@@ -233,3 +233,81 @@ def test_first_metric_null(test_id, dbt_project: DbtProject):
             materialization="incremental",
         )
         assert result["status"] == "pass"
+
+
+@pytest.mark.skip_targets(["clickhouse"])
+def test_exclude_detection_from_training(test_id: str, dbt_project: DbtProject):
+    """
+    Test the exclude_detection_period_from_training flag functionality for freshness anomalies.
+
+    Scenario:
+    - 14 days of normal data with frequent updates (every 2 hours)
+    - 7 days of anomalous data (only 1 update per day at midnight) in detection period
+    - Without exclusion: anomaly gets included in training baseline, test passes (misses anomaly)
+    - With exclusion: anomaly excluded from training, test fails (detects anomaly)
+    """
+    utc_now = datetime.utcnow().replace(hour=0, minute=0, second=0, microsecond=0)
+
+    # Generate 14 days of normal data with frequent updates (every 2 hours)
+    normal_data = [
+        {TIMESTAMP_COLUMN: date.strftime(DATE_FORMAT)}
+        for date in generate_dates(
+            base_date=utc_now - timedelta(days=21),
+            step=timedelta(hours=2),
+            days_back=14,
+        )
+    ]
+
+    # Generate 7 days of anomalous data (only 1 update per day at midnight)
+    anomalous_data = [
+        {TIMESTAMP_COLUMN: date.strftime(DATE_FORMAT)}
+        for date in generate_dates(
+            base_date=utc_now - timedelta(days=7),
+            step=timedelta(hours=24),
+            days_back=7,
+        )
+    ]
+
+    all_data = normal_data + anomalous_data
+
+    # Test 1: WITHOUT exclusion (should pass - misses the anomaly because it's included in training)
+    test_args_without_exclusion = {
+        "timestamp_column": TIMESTAMP_COLUMN,
+        "training_period": {"period": "day", "count": 14},
+        "detection_period": {"period": "day", "count": 7},
+        "time_bucket": {"period": "day", "count": 1},
+        "sensitivity": 3,
+        # exclude_detection_period_from_training is not set (defaults to False/None)
+    }
+
+    test_result_without_exclusion = dbt_project.test(
+        test_id + "_without_exclusion",
+        TEST_NAME,
+        test_args_without_exclusion,
+        data=all_data,
+        test_vars={"custom_run_started_at": utc_now.isoformat()},
+    )
+
+    # This should PASS because the anomaly is included in training, making it part of the baseline
+    assert (
+        test_result_without_exclusion["status"] == "pass"
+    ), "Test should pass when anomaly is included in training"
+
+    # Test 2: WITH exclusion (should fail - detects the anomaly because it's excluded from training)
+    test_args_with_exclusion = {
+        **test_args_without_exclusion,
+        "exclude_detection_period_from_training": True,
+    }
+
+    test_result_with_exclusion = dbt_project.test(
+        test_id + "_with_exclusion",
+        TEST_NAME,
+        test_args_with_exclusion,
+        data=all_data,
+        test_vars={"custom_run_started_at": utc_now.isoformat()},
+    )
+
+    # This should FAIL because the anomaly is excluded from training, so it's detected as anomalous
+    assert (
+        test_result_with_exclusion["status"] == "fail"
+    ), "Test should fail when anomaly is excluded from training"

macros/edr/tests/test_event_freshness_anomalies.sql

@@ -1,4 +1,4 @@
-{% test event_freshness_anomalies(model, event_timestamp_column, update_timestamp_column, where_expression, anomaly_sensitivity, anomaly_direction, min_training_set_size, time_bucket, days_back, backfill_days, seasonality, sensitivity, ignore_small_changes, detection_delay, anomaly_exclude_metrics, detection_period, training_period) %}
+{% test event_freshness_anomalies(model, event_timestamp_column, update_timestamp_column, where_expression, anomaly_sensitivity, anomaly_direction, min_training_set_size, time_bucket, days_back, backfill_days, seasonality, sensitivity, ignore_small_changes, detection_delay, anomaly_exclude_metrics, detection_period, training_period, exclude_detection_period_from_training=false) %}
   {{ config(tags = ['elementary-tests']) }}
   {% if execute and elementary.is_test_command() and elementary.is_elementary_enabled() %}
     {% set model_relation = elementary.get_model_relation_for_test(model, elementary.get_test_model()) %}
@@ -32,7 +32,8 @@
       detection_delay=detection_delay,
       anomaly_exclude_metrics=anomaly_exclude_metrics,
       detection_period=detection_period,
-      training_period=training_period
+      training_period=training_period,
+      exclude_detection_period_from_training=exclude_detection_period_from_training
     )
   }}
 {% endtest %}

macros/edr/tests/test_freshness_anomalies.sql

@@ -1,4 +1,4 @@
-{% test freshness_anomalies(model, timestamp_column, where_expression, anomaly_sensitivity, anomaly_direction, min_training_set_size, time_bucket, days_back, backfill_days, seasonality, sensitivity, ignore_small_changes, detection_delay, anomaly_exclude_metrics, detection_period, training_period) %}
+{% test freshness_anomalies(model, timestamp_column, where_expression, anomaly_sensitivity, anomaly_direction, min_training_set_size, time_bucket, days_back, backfill_days, seasonality, sensitivity, ignore_small_changes, detection_delay, anomaly_exclude_metrics, detection_period, training_period, exclude_detection_period_from_training=false) %}
   {{ config(tags = ['elementary-tests']) }}
   {{ elementary.test_table_anomalies(
       model=model,
@@ -18,7 +18,8 @@
       detection_delay=detection_delay,
       anomaly_exclude_metrics=anomaly_exclude_metrics,
       detection_period=detection_period,
-      training_period=training_period
+      training_period=training_period,
+      exclude_detection_period_from_training=exclude_detection_period_from_training
     )
   }}
 {% endtest %}