encounter/adk-python
0
0
Fork 0
mirror of https://github.com/encounter/adk-python.git synced 2026-07-09 18:19:28 -07:00
Code Issues Packages Projects Releases Wiki Activity

feat: Bigquery detect_anomalies tool results sort by timestamp for better visualization

Browse Source 
Timestamp need to be ordered so that for better display and further visualization.

PiperOrigin-RevId: 829548481
This commit is contained in:
Google Team Member
2025-11-07 13:02:55 -08:00
committed by Copybara-Service
parent 51dee43f08
commit 9e22cc4022
2 changed files with 68 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/google/adk/tools/bigquery/query_tool.py
+10 -5
View File
@@ -1136,7 +1136,7 @@ def detect_anomalies(
history_data: str,
times_series_timestamp_col: str,
times_series_data_col: str,
horizon: Optional[int] = 10,
horizon: Optional[int] = 1000,
target_data: Optional[str] = None,
times_series_id_cols: Optional[list[str]] = None,
anomaly_prob_threshold: Optional[float] = 0.95,
@@ -1158,7 +1158,7 @@ def detect_anomalies(
times_series_data_col (str): The name of the column containing the
numerical values to be forecasted and anomaly detected.
horizon (int, optional): The number of time steps to forecast into the
future. Defaults to 10.
future. Defaults to 1000.
target_data (str, optional): The table id of the BigQuery table containing
the target time series data or a query statement that select the target
data.
@@ -1301,9 +1301,14 @@ def detect_anomalies(
OPTIONS ({options_str})
AS {history_data_source}
"""
order_by_id_cols = (
", ".join(col for col in times_series_id_cols) + ", "
if times_series_id_cols
else ""
)
anomaly_detection_query = f"""
SELECT * FROM ML.DETECT_ANOMALIES(MODEL {model_name}, STRUCT({anomaly_prob_threshold} AS anomaly_prob_threshold))
SELECT * FROM ML.DETECT_ANOMALIES(MODEL {model_name}, STRUCT({anomaly_prob_threshold} AS anomaly_prob_threshold)) ORDER BY {order_by_id_cols}{times_series_timestamp_col}
"""
if target_data:
trimmed_upper_target_data = target_data.strip().upper()
@@ -1312,10 +1317,10 @@ def detect_anomalies(
) or trimmed_upper_target_data.startswith("WITH"):
target_data_source = f"({target_data})"
else:
target_data_source = f"SELECT * FROM `{target_data}`"
target_data_source = f"(SELECT * FROM `{target_data}`)"
anomaly_detection_query = f"""
SELECT * FROM ML.DETECT_ANOMALIES(MODEL {model_name}, STRUCT({anomaly_prob_threshold} AS anomaly_prob_threshold), {target_data_source})
SELECT * FROM ML.DETECT_ANOMALIES(MODEL {model_name}, STRUCT({anomaly_prob_threshold} AS anomaly_prob_threshold), {target_data_source}) ORDER BY {order_by_id_cols}{times_series_timestamp_col}
"""
# Create a session and run the create model query.
tests/unittests/tools/bigquery/test_bigquery_query_tool.py
+58 -4
View File
@@ -1436,12 +1436,12 @@ def test_detect_anomalies_with_table_id(mock_uuid, mock_execute_sql):
expected_create_model_query = """
CREATE TEMP MODEL detect_anomalies_model_test_uuid
OPTIONS (MODEL_TYPE = 'ARIMA_PLUS', TIME_SERIES_TIMESTAMP_COL = 'ts_timestamp', TIME_SERIES_DATA_COL = 'ts_data', HORIZON = 10)
OPTIONS (MODEL_TYPE = 'ARIMA_PLUS', TIME_SERIES_TIMESTAMP_COL = 'ts_timestamp', TIME_SERIES_DATA_COL = 'ts_data', HORIZON = 1000)
AS (SELECT * FROM `test-dataset.test-table`)
"""
expected_anomaly_detection_query = """
SELECT * FROM ML.DETECT_ANOMALIES(MODEL detect_anomalies_model_test_uuid, STRUCT(0.95 AS anomaly_prob_threshold))
SELECT * FROM ML.DETECT_ANOMALIES(MODEL detect_anomalies_model_test_uuid, STRUCT(0.95 AS anomaly_prob_threshold)) ORDER BY ts_timestamp
"""
assert mock_execute_sql.call_count == 2
@@ -1497,7 +1497,7 @@ def test_detect_anomalies_with_custom_params(mock_uuid, mock_execute_sql):
"""
expected_anomaly_detection_query = """
SELECT * FROM ML.DETECT_ANOMALIES(MODEL detect_anomalies_model_test_uuid, STRUCT(0.8 AS anomaly_prob_threshold))
SELECT * FROM ML.DETECT_ANOMALIES(MODEL detect_anomalies_model_test_uuid, STRUCT(0.8 AS anomaly_prob_threshold)) ORDER BY dim1, dim2, ts_timestamp
"""
assert mock_execute_sql.call_count == 2
@@ -1555,7 +1555,61 @@ def test_detect_anomalies_on_target_table(mock_uuid, mock_execute_sql):
"""
expected_anomaly_detection_query = """
SELECT * FROM ML.DETECT_ANOMALIES(MODEL detect_anomalies_model_test_uuid, STRUCT(0.8 AS anomaly_prob_threshold), (SELECT * FROM `test-dataset.target-table`))
SELECT * FROM ML.DETECT_ANOMALIES(MODEL detect_anomalies_model_test_uuid, STRUCT(0.8 AS anomaly_prob_threshold), (SELECT * FROM `test-dataset.target-table`)) ORDER BY dim1, dim2, ts_timestamp
"""
assert mock_execute_sql.call_count == 2
mock_execute_sql.assert_any_call(
project_id="test-project",
query=expected_create_model_query,
credentials=mock_credentials,
settings=mock_settings,
tool_context=mock_tool_context,
caller_id="detect_anomalies",
)
mock_execute_sql.assert_any_call(
project_id="test-project",
query=expected_anomaly_detection_query,
credentials=mock_credentials,
settings=mock_settings,
tool_context=mock_tool_context,
caller_id="detect_anomalies",
)
# detect_anomalies calls execute_sql twice. We need to test that
# the queries are properly constructed and call execute_sql with the correct
# parameters exactly twice.
@mock.patch("google.adk.tools.bigquery.query_tool._execute_sql", autospec=True)
@mock.patch("uuid.uuid4", autospec=True)
def test_detect_anomalies_with_str_table_id(mock_uuid, mock_execute_sql):
"""Test time series anomaly detection tool invocation with a table id."""
mock_credentials = mock.MagicMock(spec=Credentials)
mock_settings = BigQueryToolConfig(write_mode=WriteMode.PROTECTED)
mock_tool_context = mock.create_autospec(ToolContext, instance=True)
mock_uuid.return_value = "test_uuid"
mock_execute_sql.return_value = {"status": "SUCCESS"}
history_data_query = "SELECT * FROM `test-dataset.test-table`"
detect_anomalies(
project_id="test-project",
history_data=history_data_query,
times_series_timestamp_col="ts_timestamp",
times_series_data_col="ts_data",
target_data="test-dataset.target-table",
credentials=mock_credentials,
settings=mock_settings,
tool_context=mock_tool_context,
)
expected_create_model_query = """
CREATE TEMP MODEL detect_anomalies_model_test_uuid
OPTIONS (MODEL_TYPE = 'ARIMA_PLUS', TIME_SERIES_TIMESTAMP_COL = 'ts_timestamp', TIME_SERIES_DATA_COL = 'ts_data', HORIZON = 1000)
AS (SELECT * FROM `test-dataset.test-table`)
"""
expected_anomaly_detection_query = """
SELECT * FROM ML.DETECT_ANOMALIES(MODEL detect_anomalies_model_test_uuid, STRUCT(0.95 AS anomaly_prob_threshold), (SELECT * FROM `test-dataset.target-table`)) ORDER BY ts_timestamp
"""
assert mock_execute_sql.call_count == 2