mlos_bench.storage.sql.trial_data

An interface to access the benchmark trial data stored in SQL DB using the TrialData interface.

Classes

An interface to access the trial data stored in the SQL DB.

Module Contents

class mlos_bench.storage.sql.trial_data.TrialSqlData(*, engine: sqlalchemy.engine.Engine, schema: mlos_bench.storage.sql.schema.DbSchema, experiment_id: str, trial_id: int, config_id: int, ts_start: datetime.datetime, ts_end: datetime.datetime | None, status: mlos_bench.environments.status.Status, trial_runner_id: int | None = None)[source]

Bases: mlos_bench.storage.base_trial_data.TrialData

An interface to access the trial data stored in the SQL DB.

Parameters:

engine (sqlalchemy.engine.Engine)
schema (mlos_bench.storage.sql.schema.DbSchema)
experiment_id (str)
trial_id (int)
config_id (int)
ts_start (datetime.datetime)
ts_end (datetime.datetime | None)
status (mlos_bench.environments.status.Status)
trial_runner_id (int | None)

property metadata_df: pandas.DataFrame[source]

Retrieve the trials’ metadata params.

Note: this corresponds to the Trial object’s “config” property.

Return type:: pandas.DataFrame

property results_df: pandas.DataFrame[source]

Retrieve the trials’ results from the storage.

Return type:: pandas.DataFrame

property telemetry_df: pandas.DataFrame[source]

Retrieve the trials’ telemetry from the storage.

Return type:: pandas.DataFrame

property tunable_config: mlos_bench.storage.base_tunable_config_data.TunableConfigData[source]

Retrieve the trial’s tunable configuration from the storage.

Note: this corresponds to the Trial object’s “tunables” property.

Return type:: mlos_bench.storage.base_tunable_config_data.TunableConfigData

property tunable_config_trial_group: mlos_bench.storage.base_tunable_config_trial_group_data.TunableConfigTrialGroupData[source]

Retrieve the trial’s tunable config group configuration data from the storage.

Return type:: mlos_bench.storage.base_tunable_config_trial_group_data.TunableConfigTrialGroupData