ax.storage¶
JSON¶
ax.storage.json_store.decoder module¶
-
ax.storage.json_store.decoder.ax_class_from_json_dict(_class, object_json)[source]¶ Reinstantiates an Ax class registered in DECODER_REGISTRY from a JSON dict.
- Return type
-
ax.storage.json_store.decoder.experiment_from_json(object_json)[source]¶ Load Ax Experiment from JSON.
- Return type
-
ax.storage.json_store.decoder.generation_step_from_json(generation_step_json)[source]¶ Load generation step from JSON.
- Return type
-
ax.storage.json_store.decoder.generation_strategy_from_json(generation_strategy_json)[source]¶ Load generation strategy from JSON.
- Return type
-
ax.storage.json_store.decoder.generator_run_from_json(object_json)[source]¶ Load Ax GeneratorRun from JSON.
- Return type
-
ax.storage.json_store.decoder.multi_type_experiment_from_json(object_json)[source]¶ Load AE MultiTypeExperiment from JSON.
- Return type
-
ax.storage.json_store.decoder.object_from_json(object_json)[source]¶ Recursively load objects from a JSON-serializable dictionary.
- Return type
-
ax.storage.json_store.decoder.parameter_constraints_from_json(parameter_constraint_json, parameters)[source]¶ Load ParameterConstraints from JSON.
Order and SumConstraint are tied to a search space, and require that SearchSpace’s parameters to be passed in for decoding.
-
ax.storage.json_store.decoder.search_space_from_json(search_space_json)[source]¶ Load a SearchSpace from JSON.
This function is necessary due to the coupled loading of SearchSpace and parameter constraints.
- Return type
-
ax.storage.json_store.decoder.simple_benchmark_problem_from_json(object_json)[source]¶ Load a benchmark problem from JSON.
- Return type
ax.storage.json_store.decoders module¶
-
ax.storage.json_store.decoders.batch_trial_from_json(experiment, index, trial_type, status, time_created, time_completed, time_staged, time_run_started, abandoned_reason, run_metadata, generator_run_structs, runner, abandoned_arms_metadata, num_arms_created, status_quo, status_quo_weight_override, optimize_for_power, ttl_seconds=None, generation_step_index=None, properties=None)[source]¶ Load Ax BatchTrial from JSON.
Other classes don’t need explicit deserializers, because we can just use their constructors (see decoder.py). However, the constructor for Batch does not allow us to exactly recreate an existing object.
- Return type
-
ax.storage.json_store.decoders.class_from_json(json)[source]¶ Load any class registered in CLASS_DECODER_REGISTRY from JSON.
-
ax.storage.json_store.decoders.transform_type_from_json(object_json)[source]¶ Load the transform type from JSON.
-
ax.storage.json_store.decoders.trial_from_json(experiment, index, trial_type, status, time_created, time_completed, time_staged, time_run_started, abandoned_reason, run_metadata, generator_run, runner, num_arms_created, ttl_seconds=None, generation_step_index=None, properties=None)[source]¶ Load Ax trial from JSON.
Other classes don’t need explicit deserializers, because we can just use their constructors (see decoder.py). However, the constructor for Trial does not allow us to exactly recreate an existing object.
- Return type
ax.storage.json_store.encoder module¶
-
ax.storage.json_store.encoder.object_to_json(obj)[source]¶ Convert an Ax object to a JSON-serializable dictionary.
The root node passed to this function should always be an instance of a core Ax class or a JSON-compatible python builtin. The sub-fields of the input will then be recursively passed to this function.
e.g. if we pass an instance of Experiment, we will first fall through to the line object_dict = ENCODER_REGISTRY[_type](object), which will convert the Experiment to a (shallow) dictionary, where search subfield remains “unconverted”, i.e.: {“name”: <name: string>, “search_space”: <search space: SearchSpace>}. We then pass each item of the dictionary back into this function to recursively convert the entire object.
- Return type
ax.storage.json_store.encoders module¶
-
ax.storage.json_store.encoders.benchmark_problem_to_dict(benchmark_problem)[source]¶ Converts an Ax benchmark problem to a serializable dictionary.
-
ax.storage.json_store.encoders.botorch_model_to_dict(model)[source]¶ Convert Ax model to a dictionary.
-
ax.storage.json_store.encoders.botorch_modular_to_dict(class_type)[source]¶ Convert any class to a dictionary.
-
ax.storage.json_store.encoders.choice_parameter_to_dict(parameter)[source]¶ Convert Ax choice parameter to a dictionary.
-
ax.storage.json_store.encoders.experiment_to_dict(experiment)[source]¶ Convert Ax experiment to a dictionary.
-
ax.storage.json_store.encoders.fixed_parameter_to_dict(parameter)[source]¶ Convert Ax fixed parameter to a dictionary.
-
ax.storage.json_store.encoders.generation_step_to_dict(generation_step)[source]¶ Converts Ax generation step to a dictionary.
-
ax.storage.json_store.encoders.generation_strategy_to_dict(generation_strategy)[source]¶ Converts Ax generation strategy to a dictionary.
-
ax.storage.json_store.encoders.generator_run_to_dict(generator_run)[source]¶ Convert Ax generator run to a dictionary.
-
ax.storage.json_store.encoders.multi_objective_to_dict(objective)[source]¶ Convert Ax objective to a dictionary.
-
ax.storage.json_store.encoders.multi_type_experiment_to_dict(experiment)[source]¶ Convert AE multitype experiment to a dictionary.
-
ax.storage.json_store.encoders.objective_to_dict(objective)[source]¶ Convert Ax objective to a dictionary.
-
ax.storage.json_store.encoders.observation_features_to_dict(obs_features)[source]¶ Converts Ax observation features to a dictionary
-
ax.storage.json_store.encoders.optimization_config_to_dict(optimization_config)[source]¶ Convert Ax optimization config to a dictionary.
-
ax.storage.json_store.encoders.order_parameter_constraint_to_dict(parameter_constraint)[source]¶ Convert Ax order parameter constraint to a dictionary.
-
ax.storage.json_store.encoders.outcome_constraint_to_dict(outcome_constraint)[source]¶ Convert Ax outcome constraint to a dictionary.
-
ax.storage.json_store.encoders.parameter_constraint_to_dict(parameter_constraint)[source]¶ Convert Ax sum parameter constraint to a dictionary.
-
ax.storage.json_store.encoders.range_parameter_to_dict(parameter)[source]¶ Convert Ax range parameter to a dictionary.
-
ax.storage.json_store.encoders.runner_to_dict(runner)[source]¶ Convert Ax synthetic runner to a dictionary.
-
ax.storage.json_store.encoders.scalarized_objective_to_dict(objective)[source]¶ Convert Ax objective to a dictionary.
-
ax.storage.json_store.encoders.search_space_to_dict(search_space)[source]¶ Convert Ax search space to a dictionary.
-
ax.storage.json_store.encoders.simple_experiment_to_dict(experiment)[source]¶ Convert AE simple experiment to a dictionary.
-
ax.storage.json_store.encoders.sum_parameter_constraint_to_dict(parameter_constraint)[source]¶ Convert Ax sum parameter constraint to a dictionary.
-
ax.storage.json_store.encoders.surrogate_to_dict(surrogate)[source]¶ Convert Ax surrogate to a dictionary.
ax.storage.json_store.load module¶
ax.storage.json_store.registry module¶
SQLAlchemy (MySQL / SQLite)¶
ax.storage.sqa_store.base_decoder module¶
-
class
ax.storage.sqa_store.decoder.Decoder(config)[source]¶ Bases:
objectClass that contains methods for loading an Ax experiment from SQLAlchemy.
Instantiate with an instance of Config to customize the functionality. For even more flexibility, create a subclass.
-
config¶ Metadata needed to save and load an experiment to SQLAlchemy.
-
abandoned_arm_from_sqa(abandoned_arm_sqa)[source]¶ Convert SQLAlchemy AbandonedArm to Ax AbandonedArm.
- Return type
-
experiment_from_sqa(experiment_sqa)[source]¶ Convert SQLAlchemy Experiment to Ax Experiment.
- Return type
-
generation_strategy_from_sqa(gs_sqa)[source]¶ Convert SQALchemy generation strategy to Ax GenerationStrategy.
- Return type
-
generator_run_from_sqa(generator_run_sqa)[source]¶ Convert SQLAlchemy GeneratorRun to Ax GeneratorRun.
- Return type
-
get_enum_name(value, enum)[source]¶ Given an enum value (int) and an enum (of ints), return the corresponding enum name. If the value is not present in the enum, throw an error.
-
metric_from_sqa(metric_sqa)[source]¶ Convert SQLAlchemy Metric to Ax Metric, Objective, or OutcomeConstraint.
- Return type
-
opt_config_and_tracking_metrics_from_sqa(metrics_sqa)[source]¶ Convert a list of SQLAlchemy Metrics to a a tuple of Ax OptimizationConfig and tracking metrics.
- Return type
-
parameter_constraint_from_sqa(parameter_constraint_sqa, parameters)[source]¶ Convert SQLAlchemy ParameterConstraint to Ax ParameterConstraint.
- Return type
-
parameter_from_sqa(parameter_sqa)[source]¶ Convert SQLAlchemy Parameter to Ax Parameter.
- Return type
-
ax.storage.sqa_store.base_encoder module¶
-
class
ax.storage.sqa_store.encoder.Encoder(config)[source]¶ Bases:
objectClass that contains methods for storing an Ax experiment to SQLAlchemy.
Instantiate with an instance of Config to customize the functionality. For even more flexibility, create a subclass.
-
config¶ Metadata needed to save and load an experiment to SQLAlchemy.
-
experiment_to_sqa(experiment)[source]¶ Convert Ax Experiment to SQLAlchemy.
In addition to creating and storing a new Experiment object, we need to create and store copies of the Trials, Metrics, Parameters, ParameterConstraints, and Runner owned by this Experiment.
- Return type
-
generation_strategy_to_sqa(generation_strategy, experiment_id)[source]¶ Convert an Ax GenerationStrategy to SQLAlchemy, preserving its state, so that the restored generation strategy can be resumed from the point at which it was interrupted and stored.
- Return type
-
generator_run_to_sqa(generator_run, weight=None)[source]¶ Convert Ax GeneratorRun to SQLAlchemy.
In addition to creating and storing a new GeneratorRun object, we need to create and store copies of the Arms, Metrics, Parameters, and ParameterConstraints owned by this GeneratorRun.
- Return type
-
get_enum_value(value, enum)[source]¶ Given an enum name (string) and an enum (of ints), return the corresponding enum value. If the name is not present in the enum, throw an error.
-
get_metric_type_and_properties(metric)[source]¶ Given an Ax Metric, convert its type into a member of MetricType enum, and construct a dictionary to be stored in the database properties json blob.
-
multi_objective_to_sqa(objective)[source]¶ Convert Ax Multi Objective to SQLAlchemy. Returns a parent SQAMetric, whose children are the SQAMetrics corresponding to metrics attribute of MultiObjective. The parent is used as a placeholder for storage purposes.
- Return type
-
optimization_config_to_sqa(optimization_config)[source]¶ Convert Ax OptimizationConfig to a list of SQLAlchemy Metrics.
-
outcome_constraint_to_sqa(outcome_constraint)[source]¶ Convert Ax OutcomeConstraint to SQLAlchemy.
- Return type
-
parameter_constraint_to_sqa(parameter_constraint)[source]¶ Convert Ax ParameterConstraint to SQLAlchemy.
- Return type
-
scalarized_objective_to_sqa(objective)[source]¶ Convert Ax Scalarized Objective to SQLAlchemy. Returns a parent SQAMetric, whose children are the SQAMetrics corresponding to metrics attribute of Scalarized Objective. The parent is used as a placeholder for storage purposes.
- Return type
-
search_space_to_sqa(search_space)[source]¶ Convert Ax SearchSpace to a list of SQLAlchemy Parameters and ParameterConstraints.
- Return type
-
ax.storage.sqa_store.db module¶
-
class
ax.storage.sqa_store.db.SQABase[source]¶ Bases:
objectMetaclass for SQLAlchemy classes corresponding to core Ax classes.
-
property
attributes¶ Return a list of the column attributes and relationship fields on this SQABase instance. Used for iterating over the fields to determine equality, perform updates, etc.
-
equals(other)¶
-
static
list_equals(l1, l2)[source]¶ Compare equality of two lists.
- Assumptions:
– The lists do not contain duplicates
Checking equality is then the same as checking that the lists are the same length, and that one is a subset of the other.
- Return type
-
static
list_update(l1, l2)[source]¶ - Given an existing list (l1) and an new version (l2):
– update the existing items in l1 that have matching items in l2 – delete existing items in l1 that don’t have matching items in l2 – add items in l2 that don’t exist in l1
- e.g. list_update([1,2,3], [1,5]) => [1,5]
- list_update([Arm(name=”0_0”)], [Arm(name=”0_0”), Arm(name=”0_1”)]) =>
[Arm(name=”0_0”), Arm(name=”0_1”)] where Arm(name=”0_0”) has been updated, not replaced, so that we don’t delete/recreate the DB row
- Return type
List[~T]
-
property
-
ax.storage.sqa_store.db.create_all_tables(engine)[source]¶ Create all tables that inherit from Base.
- Parameters
engine (
Engine) – a SQLAlchemy engine with a connection to a MySQL or SQLite DB.
Note
In order for all tables to be correctly created, all modules that define a mapped class that inherits from Base must be imported.
- Return type
None
-
ax.storage.sqa_store.db.create_mysql_engine_from_creator(creator, echo=False, pool_recycle=10, **kwargs)[source]¶ Create a SQLAlchemy engine with the MySQL dialect given a creator function.
- Parameters
- Returns
SQLAlchemy engine with connection to MySQL DB.
- Return type
Engine
-
ax.storage.sqa_store.db.create_mysql_engine_from_url(url, echo=False, pool_recycle=10, **kwargs)[source]¶ Create a SQLAlchemy engine with the MySQL dialect given a database url.
- Parameters
url (
str) – a database url that can include username, password, hostname, database name as well as optional keyword arguments for additional configuration. e.g. dialect+driver://username:password@host:port/database.echo (
bool) – if True, set engine to be verbose.pool_recycle (
int) – number of seconds after which to recycle connections. -1 means no timeout. Default is 10 seconds.**kwargs – keyword args passed to create_engine
- Returns
SQLAlchemy engine with connection to MySQL DB.
- Return type
Engine
-
ax.storage.sqa_store.db.create_test_engine(path=None, echo=True)[source]¶ Creates a SQLAlchemy engine object for use in unit tests.
-
ax.storage.sqa_store.db.get_engine()[source]¶ Fetch a SQLAlchemy engine, if already initialized.
If not initialized, need to either call init_engine_and_session_factory or get_session explicitly.
- Returns
an instance of a SQLAlchemy engine with a connection to a DB.
- Return type
Engine
-
ax.storage.sqa_store.db.get_session()[source]¶ Fetch a SQLAlchemy session with a connection to a DB.
Unless init_engine_and_session_factory is called first with custom args, this will automatically initialize a connection to xdb.adaptive_experiment.
- Returns
an instance of a SQLAlchemy session.
- Return type
Session
-
ax.storage.sqa_store.db.init_engine_and_session_factory(url=None, creator=None, echo=False, force_init=False, **kwargs)[source]¶ Initialize the global engine and SESSION_FACTORY for SQLAlchemy.
The initialization needs to only happen once. Note that it is possible to re-initialize the engine by setting the force_init flag to True, but this should only be used if you are absolutely certain that you know what you are doing.
- Parameters
url (
Optional[str]) – a database url that can include username, password, hostname, database name as well as optional keyword arguments for additional configuration. e.g. dialect+driver://username:password@host:port/database. Either this argument or creator argument must be specified.creator (
Optional[Callable]) – a callable which returns a DBAPI connection. Either this argument or url argument must be specified.echo (
bool) – if True, logging for engine is enabled.force_init (
bool) – if True, allows re-initializing engine and session factory.**kwargs – keyword arguments passed to create_mysql_engine_from_creator
- Return type
None
-
ax.storage.sqa_store.db.init_test_engine_and_session_factory(tier_or_path=None, echo=False, force_init=False, **kwargs)[source]¶ Initialize the global engine and SESSION_FACTORY for SQLAlchemy, using an in-memory SQLite database.
The initialization needs to only happen once. Note that it is possible to re-initialize the engine by setting the force_init flag to True, but this should only be used if you are absolutely certain that you know what you are doing.
ax.storage.sqa_store.json module¶
-
class
ax.storage.sqa_store.json.JSONEncodedMediumText(object_pairs_hook=None, *args, **kwargs)[source]¶ Bases:
ax.storage.sqa_store.json.JSONEncodedObjectClass for JSON-encoding objects in SQLAlchemy, backed by MEDIUMTEXT (MySQL).
See description in JSONEncodedObject.
-
impl: sqlalchemy.sql.sqltypes.VARCHAR = Text(length=16777215)¶
-
-
class
ax.storage.sqa_store.json.JSONEncodedObject(object_pairs_hook=None, *args, **kwargs)[source]¶ Bases:
sqlalchemy.sql.type_api.TypeDecoratorClass for JSON-encoding objects in SQLAlchemy.
Represents an object that is automatically marshalled and unmarshalled to/from the corresponding JSON string. By itself, this data structure does not track any changes.
-
impl: sqlalchemy.sql.sqltypes.VARCHAR = VARCHAR(length=4096)¶
-
process_bind_param(value, dialect)[source]¶ Receive a bound parameter value to be converted.
Subclasses override this method to return the value that should be passed along to the underlying
TypeEngineobject, and from there to the DBAPIexecute()method.The operation could be anything desired to perform custom behavior, such as transforming or serializing data. This could also be used as a hook for validating logic.
This operation should be designed with the reverse operation in mind, which would be the process_result_value method of this class.
-
process_result_value(value, dialect)[source]¶ Receive a result-row column value to be converted.
Subclasses should implement this method to operate on data fetched from the database.
Subclasses override this method to return the value that should be passed back to the application, given a value that is already processed by the underlying
TypeEngineobject, originally from the DBAPI cursor methodfetchone()or similar.The operation could be anything desired to perform custom behavior, such as transforming or serializing data. This could also be used as a hook for validating logic.
- Parameters
This operation should be designed to be reversible by the “process_bind_param” method of this class.
- Return type
-
-
class
ax.storage.sqa_store.json.JSONEncodedText(object_pairs_hook=None, *args, **kwargs)[source]¶ Bases:
ax.storage.sqa_store.json.JSONEncodedObjectClass for JSON-encoding objects in SQLAlchemy, backed by TEXT (MySQL).
See description in JSONEncodedObject.
-
impl¶ alias of
sqlalchemy.sql.sqltypes.Text
-
ax.storage.sqa_store.load module¶
-
ax.storage.sqa_store.load.load_experiment(experiment_name, config=None)[source]¶ Load experiment by name (uses default SQAConfig).
- Return type
-
ax.storage.sqa_store.load.load_generation_strategy_by_experiment_name(experiment_name, config=None)[source]¶ Finds a generation strategy attached to an experiment specified by a name and restores it from its corresponding SQA object.
- Return type
ax.storage.sqa_store.save module¶
-
ax.storage.sqa_store.save.save_experiment(experiment, config=None)[source]¶ Save experiment (using default SQAConfig).
- Return type
None
-
ax.storage.sqa_store.save.save_generation_strategy(generation_strategy, config=None)[source]¶ Save generation strategy (using default SQAConfig if no config is specified). If the generation strategy has an experiment set, the experiment will be saved first.
- Return type
- Returns
The ID of the saved generation strategy.
-
ax.storage.sqa_store.save.save_new_trial(experiment, trial, config=None)[source]¶ Add new trial to the experiment (using default SQAConfig).
- Return type
None
ax.storage.sqa_store.structs module¶
-
class
ax.storage.sqa_store.structs.DBSettings[source]¶ Bases:
tupleDefines behavior for loading/saving experiment to/from db. Either creator or url must be specified as a way to connect to the SQL db.
-
property
creator¶ Alias for field number 0
-
property
decoder¶ Alias for field number 1
-
property
encoder¶ Alias for field number 2
-
property
url¶ Alias for field number 3
-
property
ax.storage.sqa_store.sqa_classes module¶
-
class
ax.storage.sqa_store.sqa_classes.SQAAbandonedArm(*args, **kwargs)[source]¶ Bases:
sqlalchemy.ext.declarative.declarative_base-
abandoned_reason: Optional[str] = Column(None, String(length=255), table=None)¶
-
id: int = Column(None, Integer(), table=None, primary_key=True, nullable=False)¶
-
immutable_fields= ['name']¶
-
name: str = Column(None, String(length=100), table=None, nullable=False)¶
-
time_abandoned: datetime.datetime = Column(None, IntTimestamp(), table=None, nullable=False, default=ColumnDefault(<function datetime.now>))¶
-
trial_id: int = Column(None, Integer(), ForeignKey('trial_v2.id'), table=None)¶
-
unique_id= 'name'¶
-
-
class
ax.storage.sqa_store.sqa_classes.SQAArm(*args, **kwargs)[source]¶ Bases:
sqlalchemy.ext.declarative.declarative_base-
generator_run_id: int = Column(None, Integer(), ForeignKey('generator_run_v2.id'), table=None)¶
-
id: int = Column(None, Integer(), table=None, primary_key=True, nullable=False)¶
-
immutable_fields= ['parameters']¶
-
name: Optional[str] = Column(None, String(length=100), table=None)¶
-
parameters: Dict[str, Union[str, bool, float, int, None]] = Column(None, JSONEncodedText(), table=None, nullable=False)¶
-
unique_id= 'name'¶
-
weight: float = Column(None, Float(), table=None, nullable=False, default=ColumnDefault(1.0))¶
-
-
class
ax.storage.sqa_store.sqa_classes.SQAData(*args, **kwargs)[source]¶ Bases:
sqlalchemy.ext.declarative.declarative_base-
data_json: str = Column(None, Text(length=4294967295), table=None, nullable=False)¶
-
description: Optional[str] = Column(None, String(length=255), table=None)¶
-
experiment_id: int = Column(None, Integer(), ForeignKey('experiment_v2.id'), table=None)¶
-
generation_strategy_id: Optional[int] = Column(None, Integer(), ForeignKey('generation_strategy.id'), table=None)¶
-
id: int = Column(None, Integer(), table=None, primary_key=True, nullable=False)¶
-
time_created: int = Column(None, BigInteger(), table=None, nullable=False)¶
-
trial_index: Optional[int] = Column(None, Integer(), table=None)¶
-
unique_id= 'time_created'¶
-
-
class
ax.storage.sqa_store.sqa_classes.SQAExperiment(*args, **kwargs)[source]¶ Bases:
sqlalchemy.ext.declarative.declarative_base-
data: List[ax.storage.sqa_store.sqa_classes.SQAData] = <RelationshipProperty at 0x7f0573c347c8; no key>¶
-
default_trial_type: Optional[str] = Column(None, String(length=100), table=None)¶
-
description: Optional[str] = Column(None, String(length=255), table=None)¶
-
experiment_type: Optional[int] = Column(None, Integer(), table=None)¶
-
generation_strategy: Optional[ax.storage.sqa_store.sqa_classes.SQAGenerationStrategy] = <RelationshipProperty at 0x7f0573c34ac8; no key>¶
-
id: int = Column(None, Integer(), table=None, primary_key=True, nullable=False)¶
-
ignore_during_update_fields= ['time_created']¶
-
immutable_fields= ['name']¶
-
is_test: bool = Column(None, Boolean(), table=None, nullable=False, default=ColumnDefault(False))¶
-
metrics: List[ax.storage.sqa_store.sqa_classes.SQAMetric] = <RelationshipProperty at 0x7f0573c34848; no key>¶
-
name: str = Column(None, String(length=100), table=None, nullable=False)¶
-
parameter_constraints: List[ax.storage.sqa_store.sqa_classes.SQAParameterConstraint] = <RelationshipProperty at 0x7f0573c34948; no key>¶
-
parameters: List[ax.storage.sqa_store.sqa_classes.SQAParameter] = <RelationshipProperty at 0x7f0573c348c8; no key>¶
-
properties: Optional[Dict[str, Any]] = Column(None, JSONEncodedText(), table=None, default=ColumnDefault({}))¶
-
runners: List[ax.storage.sqa_store.sqa_classes.SQARunner] = <RelationshipProperty at 0x7f0573c349c8; no key>¶
-
status_quo_name: Optional[str] = Column(None, String(length=100), table=None)¶
-
status_quo_parameters: Optional[Dict[str, Union[str, bool, float, int, None]]] = Column(None, JSONEncodedText(), table=None)¶
-
time_created: datetime.datetime = Column(None, IntTimestamp(), table=None, nullable=False)¶
-
trials: List[ax.storage.sqa_store.sqa_classes.SQATrial] = <RelationshipProperty at 0x7f0573c34a48; no key>¶
-
-
class
ax.storage.sqa_store.sqa_classes.SQAGenerationStrategy(*args, **kwargs)[source]¶ Bases:
sqlalchemy.ext.declarative.declarative_base-
curr_index: int = Column(None, Integer(), table=None, nullable=False)¶
-
data: ax.storage.sqa_store.sqa_classes.SQAData = <RelationshipProperty at 0x7f0573c23ac8; no key>¶
-
experiment_id: Optional[int] = Column(None, Integer(), ForeignKey('experiment_v2.id'), table=None)¶
-
generator_runs: List[ax.storage.sqa_store.sqa_classes.SQAGeneratorRun] = <RelationshipProperty at 0x7f0573c23a48; no key>¶
-
id: int = Column(None, Integer(), table=None, primary_key=True, nullable=False)¶
-
name: str = Column(None, String(length=100), table=None, nullable=False)¶
-
steps: List[Dict[str, Any]] = Column(None, JSONEncodedObject(length=4096), table=None, nullable=False)¶
-
-
class
ax.storage.sqa_store.sqa_classes.SQAGeneratorRun(*args, **kwargs)[source]¶ Bases:
sqlalchemy.ext.declarative.declarative_base-
arms: List[ax.storage.sqa_store.sqa_classes.SQAArm] = <RelationshipProperty at 0x7f0573c232c8; no key>¶
-
best_arm_name: Optional[str] = Column(None, String(length=100), table=None)¶
-
best_arm_parameters: Optional[Dict[str, Union[str, bool, float, int, None]]] = Column(None, JSONEncodedText(), table=None)¶
-
best_arm_predictions: Optional[Tuple[Dict[str, float], Optional[Dict[str, Dict[str, float]]]]] = Column(None, JSONEncodedObject(length=4096), table=None)¶
-
bridge_kwargs: Optional[Dict[str, Any]] = Column(None, JSONEncodedText(), table=None)¶
-
candidate_metadata_by_arm_signature: Optional[Dict[str, Any]] = Column(None, JSONEncodedText(), table=None)¶
-
fit_time: Optional[float] = Column(None, Float(), table=None)¶
-
gen_metadata: Optional[Dict[str, Any]] = Column(None, JSONEncodedText(), table=None)¶
-
gen_time: Optional[float] = Column(None, Float(), table=None)¶
-
generation_step_index: Optional[int] = Column(None, Integer(), table=None)¶
-
generation_strategy_id: Optional[int] = Column(None, Integer(), ForeignKey('generation_strategy.id'), table=None)¶
-
generator_run_type: Optional[int] = Column(None, Integer(), table=None)¶
-
id: int = Column(None, Integer(), table=None, primary_key=True, nullable=False)¶
-
ignore_during_update_fields= ['time_created']¶
-
index: Optional[int] = Column(None, Integer(), table=None)¶
-
metrics: List[ax.storage.sqa_store.sqa_classes.SQAMetric] = <RelationshipProperty at 0x7f0573c23348; no key>¶
-
model_key: Optional[str] = Column(None, String(length=100), table=None)¶
-
model_kwargs: Optional[Dict[str, Any]] = Column(None, JSONEncodedText(), table=None)¶
-
model_predictions: Optional[Tuple[Dict[str, List[float]], Dict[str, Dict[str, List[float]]]]] = Column(None, JSONEncodedObject(length=4096), table=None)¶
-
model_state_after_gen: Optional[Dict[str, Any]] = Column(None, JSONEncodedText(), table=None)¶
-
parameter_constraints: List[ax.storage.sqa_store.sqa_classes.SQAParameterConstraint] = <RelationshipProperty at 0x7f0573c23448; no key>¶
-
parameters: List[ax.storage.sqa_store.sqa_classes.SQAParameter] = <RelationshipProperty at 0x7f0573c233c8; no key>¶
-
time_created: datetime.datetime = Column(None, IntTimestamp(), table=None, nullable=False, default=ColumnDefault(<function datetime.now>))¶
-
trial_id: Optional[int] = Column(None, Integer(), ForeignKey('trial_v2.id'), table=None)¶
-
unique_id= 'index'¶
-
weight: Optional[float] = Column(None, Float(), table=None)¶
-
-
class
ax.storage.sqa_store.sqa_classes.SQAMetric(*args, **kwargs)[source]¶ Bases:
sqlalchemy.ext.declarative.declarative_base-
bound: Optional[float] = Column(None, Float(), table=None)¶
-
canonical_name: Optional[str] = Column(None, String(length=100), table=None)¶
-
experiment_id: Optional[int] = Column(None, Integer(), ForeignKey('experiment_v2.id'), table=None)¶
-
generator_run_id: Optional[int] = Column(None, Integer(), ForeignKey('generator_run_v2.id'), table=None)¶
-
id: int = Column(None, Integer(), table=None, primary_key=True, nullable=False)¶
-
immutable_fields= ['name']¶
-
intent: ax.storage.utils.MetricIntent = Column(None, StringEnum(length=100), table=None, nullable=False)¶
-
lower_is_better: Optional[bool] = Column(None, Boolean(), table=None)¶
-
metric_type: int = Column(None, Integer(), table=None, nullable=False)¶
-
minimize: Optional[bool] = Column(None, Boolean(), table=None)¶
-
name: str = Column(None, String(length=255), table=None, nullable=False)¶
-
op: Optional[ax.core.types.ComparisonOp] = Column(None, IntEnum(), table=None)¶
-
properties: Optional[Dict[str, Any]] = Column(None, JSONEncodedText(), table=None, default=ColumnDefault({}))¶
-
relative: Optional[bool] = Column(None, Boolean(), table=None)¶
-
scalarized_objective_children_metrics= <RelationshipProperty at 0x7f0573c0bb48; no key>¶
-
scalarized_objective_id= Column(None, Integer(), ForeignKey('metric_v2.id'), table=None)¶
-
scalarized_objective_weight: Optional[float] = Column(None, Float(), table=None)¶
-
trial_type: Optional[str] = Column(None, String(length=100), table=None)¶
-
unique_id= 'name'¶
-
-
class
ax.storage.sqa_store.sqa_classes.SQAParameter(*args, **kwargs)[source]¶ Bases:
sqlalchemy.ext.declarative.declarative_base-
choice_values: Optional[List[Union[str, bool, float, int, None]]] = Column(None, JSONEncodedObject(length=4096), table=None)¶
-
digits: Optional[int] = Column(None, Integer(), table=None)¶
-
domain_type: ax.storage.utils.DomainType = Column(None, IntEnum(), table=None, nullable=False)¶
-
experiment_id: Optional[int] = Column(None, Integer(), ForeignKey('experiment_v2.id'), table=None)¶
-
fixed_value: Union[str, bool, float, int, None] = Column(None, JSONEncodedObject(length=4096), table=None)¶
-
generator_run_id: Optional[int] = Column(None, Integer(), ForeignKey('generator_run_v2.id'), table=None)¶
-
id: int = Column(None, Integer(), table=None, primary_key=True, nullable=False)¶
-
immutable_fields= ['name']¶
-
is_fidelity: Optional[bool] = Column(None, Boolean(), table=None)¶
-
is_ordered: Optional[bool] = Column(None, Boolean(), table=None)¶
-
is_task: Optional[bool] = Column(None, Boolean(), table=None)¶
-
log_scale: Optional[bool] = Column(None, Boolean(), table=None)¶
-
lower: Optional[float] = Column(None, Float(), table=None)¶
-
name: str = Column(None, String(length=100), table=None, nullable=False)¶
-
parameter_type: ax.core.parameter.ParameterType = Column(None, IntEnum(), table=None, nullable=False)¶
-
target_value: Union[str, bool, float, int, None] = Column(None, JSONEncodedObject(length=4096), table=None)¶
-
unique_id= 'name'¶
-
upper: Optional[float] = Column(None, Float(), table=None)¶
-
-
class
ax.storage.sqa_store.sqa_classes.SQAParameterConstraint(*args, **kwargs)[source]¶ Bases:
sqlalchemy.ext.declarative.declarative_base-
bound: float = Column(None, Float(), table=None, nullable=False)¶
-
constraint_dict: Dict[str, float] = Column(None, JSONEncodedObject(length=4096), table=None, nullable=False)¶
-
experiment_id: Optional[int] = Column(None, Integer(), ForeignKey('experiment_v2.id'), table=None)¶
-
generator_run_id: Optional[int] = Column(None, Integer(), ForeignKey('generator_run_v2.id'), table=None)¶
-
id: int = Column(None, Integer(), table=None, primary_key=True, nullable=False)¶
-
immutable_fields= ['type', 'constraint_dict', 'bound']¶
-
type: ax.storage.sqa_store.sqa_enum.IntEnum = Column(None, IntEnum(), table=None, nullable=False)¶
-
-
class
ax.storage.sqa_store.sqa_classes.SQARunner(*args, **kwargs)[source]¶ Bases:
sqlalchemy.ext.declarative.declarative_base-
experiment_id: Optional[int] = Column(None, Integer(), ForeignKey('experiment_v2.id'), table=None)¶
-
id: int = Column(None, Integer(), table=None, primary_key=True, nullable=False)¶
-
properties: Optional[Dict[str, Any]] = Column(None, JSONEncodedText(), table=None, default=ColumnDefault({}))¶
-
runner_type: int = Column(None, Integer(), table=None, nullable=False)¶
-
trial_id: Optional[int] = Column(None, Integer(), ForeignKey('trial_v2.id'), table=None)¶
-
trial_type: Optional[str] = Column(None, String(length=100), table=None)¶
-
-
class
ax.storage.sqa_store.sqa_classes.SQATrial(*args, **kwargs)[source]¶ Bases:
sqlalchemy.ext.declarative.declarative_base-
abandoned_arms: List[ax.storage.sqa_store.sqa_classes.SQAAbandonedArm] = <RelationshipProperty at 0x7f0573c343c8; no key>¶
-
abandoned_reason: Optional[str] = Column(None, String(length=100), table=None)¶
-
deployed_name: Optional[str] = Column(None, String(length=100), table=None)¶
-
experiment_id: int = Column(None, Integer(), ForeignKey('experiment_v2.id'), table=None)¶
-
generation_step_index: Optional[int] = Column(None, Integer(), table=None)¶
-
generator_runs: List[ax.storage.sqa_store.sqa_classes.SQAGeneratorRun] = <RelationshipProperty at 0x7f0573c34448; no key>¶
-
id: int = Column(None, Integer(), table=None, primary_key=True, nullable=False)¶
-
ignore_during_update_fields= ['time_created']¶
-
immutable_fields= ['is_batch']¶
-
index: int = Column(None, Integer(), table=None, nullable=False)¶
-
is_batch: bool = Column('is_batched', Boolean(), table=None, nullable=False, default=ColumnDefault(True))¶
-
num_arms_created: int = Column(None, Integer(), table=None, nullable=False, default=ColumnDefault(0))¶
-
optimize_for_power: Optional[bool] = Column(None, Boolean(), table=None)¶
-
properties: Optional[Dict[str, Any]] = Column(None, JSONEncodedText(), table=None, default=ColumnDefault({}))¶
-
run_metadata: Optional[Dict[str, Any]] = Column(None, JSONEncodedText(), table=None)¶
-
runner: ax.storage.sqa_store.sqa_classes.SQARunner = <RelationshipProperty at 0x7f0573c344c8; no key>¶
-
status: ax.core.base_trial.TrialStatus = Column(None, IntEnum(), table=None, nullable=False, default=ColumnDefault(<TrialStatus.CANDIDATE: 0>))¶
-
status_quo_name: Optional[str] = Column(None, String(length=100), table=None)¶
-
time_completed: Optional[datetime.datetime] = Column(None, IntTimestamp(), table=None)¶
-
time_created: datetime.datetime = Column(None, IntTimestamp(), table=None, nullable=False)¶
-
time_run_started: Optional[datetime.datetime] = Column(None, IntTimestamp(), table=None)¶
-
time_staged: Optional[datetime.datetime] = Column(None, IntTimestamp(), table=None)¶
-
trial_type: Optional[str] = Column(None, String(length=100), table=None)¶
-
ttl_seconds: Optional[int] = Column(None, Integer(), table=None)¶
-
unique_id= 'index'¶
-
ax.storage.sqa_store.sqa_config module¶
-
class
ax.storage.sqa_store.sqa_config.SQAConfig[source]¶ Bases:
tupleMetadata needed to save and load an experiment to SQLAlchemy.
-
class_to_sqa_class¶ Mapping of user-facing class to SQLAlchemy class that it will be encoded to. This allows overwriting of the default classes to provide custom save functionality.
-
experiment_type_enum¶ Enum containing valid Experiment types.
-
generator_run_type_enum¶ Enum containing valid Generator Run types.
-
property
class_to_sqa_class Alias for field number 0
-
property
experiment_type_enum Alias for field number 1
-
property
generator_run_type_enum Alias for field number 2
-
ax.storage.sqa_store.sqa_enum module¶
-
class
ax.storage.sqa_store.sqa_enum.BaseNullableEnum(enum, *arg, **kw)[source]¶ Bases:
sqlalchemy.sql.type_api.TypeDecorator-
process_bind_param(value, dialect)[source]¶ Receive a bound parameter value to be converted.
Subclasses override this method to return the value that should be passed along to the underlying
TypeEngineobject, and from there to the DBAPIexecute()method.The operation could be anything desired to perform custom behavior, such as transforming or serializing data. This could also be used as a hook for validating logic.
This operation should be designed with the reverse operation in mind, which would be the process_result_value method of this class.
-
process_result_value(value, dialect)[source]¶ Receive a result-row column value to be converted.
Subclasses should implement this method to operate on data fetched from the database.
Subclasses override this method to return the value that should be passed back to the application, given a value that is already processed by the underlying
TypeEngineobject, originally from the DBAPI cursor methodfetchone()or similar.The operation could be anything desired to perform custom behavior, such as transforming or serializing data. This could also be used as a hook for validating logic.
- Parameters
This operation should be designed to be reversible by the “process_bind_param” method of this class.
- Return type
-
-
class
ax.storage.sqa_store.sqa_enum.IntEnum(enum, *arg, **kw)[source]¶ Bases:
ax.storage.sqa_store.sqa_enum.BaseNullableEnum-
impl¶ alias of
sqlalchemy.sql.sqltypes.SmallInteger
-
-
class
ax.storage.sqa_store.sqa_enum.StringEnum(enum, *arg, **kw)[source]¶ Bases:
ax.storage.sqa_store.sqa_enum.BaseNullableEnum-
impl= VARCHAR(length=100)¶
-
ax.storage.sqa_store.timestamp module¶
-
class
ax.storage.sqa_store.timestamp.IntTimestamp(*args, **kwargs)[source]¶ Bases:
sqlalchemy.sql.type_api.TypeDecorator-
impl¶ alias of
sqlalchemy.sql.sqltypes.Integer
-
process_bind_param(value, dialect)[source]¶ Receive a bound parameter value to be converted.
Subclasses override this method to return the value that should be passed along to the underlying
TypeEngineobject, and from there to the DBAPIexecute()method.The operation could be anything desired to perform custom behavior, such as transforming or serializing data. This could also be used as a hook for validating logic.
This operation should be designed with the reverse operation in mind, which would be the process_result_value method of this class.
-
process_result_value(value, dialect)[source]¶ Receive a result-row column value to be converted.
Subclasses should implement this method to operate on data fetched from the database.
Subclasses override this method to return the value that should be passed back to the application, given a value that is already processed by the underlying
TypeEngineobject, originally from the DBAPI cursor methodfetchone()or similar.The operation could be anything desired to perform custom behavior, such as transforming or serializing data. This could also be used as a hook for validating logic.
- Parameters
This operation should be designed to be reversible by the “process_bind_param” method of this class.
-
ax.storage.sqa_store.utils module¶
ax.storage.sqa_store.validation module¶
-
ax.storage.sqa_store.validation.consistency_exactly_one(instance, exactly_one_fields)[source]¶ Ensure that exactly one of exactly_one_fields has a value set.
- Return type
-
ax.storage.sqa_store.validation.validate_metric(mapper, connection, target)[source]¶ - Return type
None
-
ax.storage.sqa_store.validation.validate_parameter(mapper, connection, target)[source]¶ - Return type
None
Registries¶
-
ax.storage.botorch_modular_registry.ACQUISITION_FUNCTION_REGISTRY: Dict[Type[botorch.acquisition.acquisition.AcquisitionFunction], int] = {<class 'botorch.acquisition.monte_carlo.qExpectedImprovement'>: 0, <class 'botorch.acquisition.monte_carlo.qNoisyExpectedImprovement'>: 1, <class 'botorch.acquisition.knowledge_gradient.qKnowledgeGradient'>: 2, <class 'botorch.acquisition.knowledge_gradient.qMultiFidelityKnowledgeGradient'>: 3, <class 'botorch.acquisition.max_value_entropy_search.qMaxValueEntropy'>: 4, <class 'botorch.acquisition.max_value_entropy_search.qMultiFidelityMaxValueEntropy'>: 5}¶ Mapping of BoTorch MarginalLogLikelihood classes to ints.
-
ax.storage.botorch_modular_registry.ACQUISITION_REGISTRY: Dict[Type[ax.models.torch.botorch_modular.acquisition.Acquisition], int] = {<class 'ax.models.torch.botorch_modular.acquisition.Acquisition'>: 0, <class 'ax.models.torch.botorch_modular.kg.KnowledgeGradient'>: 1, <class 'ax.models.torch.botorch_modular.kg.MultiFidelityKnowledgeGradient'>: 2, <class 'ax.models.torch.botorch_modular.mes.MaxValueEntropySearch'>: 3, <class 'ax.models.torch.botorch_modular.mes.MultiFidelityMaxValueEntropySearch'>: 4}¶ Mapping of BoTorch Model classes to ints.
-
ax.storage.botorch_modular_registry.CLASS_TO_REGISTRY: Dict[Any, Dict[Type[Any], int]] = {<class 'ax.models.torch.botorch_modular.acquisition.Acquisition'>: {<class 'ax.models.torch.botorch_modular.acquisition.Acquisition'>: 0, <class 'ax.models.torch.botorch_modular.kg.KnowledgeGradient'>: 1, <class 'ax.models.torch.botorch_modular.kg.MultiFidelityKnowledgeGradient'>: 2, <class 'ax.models.torch.botorch_modular.mes.MaxValueEntropySearch'>: 3, <class 'ax.models.torch.botorch_modular.mes.MultiFidelityMaxValueEntropySearch'>: 4}, <class 'botorch.acquisition.acquisition.AcquisitionFunction'>: {<class 'botorch.acquisition.monte_carlo.qExpectedImprovement'>: 0, <class 'botorch.acquisition.monte_carlo.qNoisyExpectedImprovement'>: 1, <class 'botorch.acquisition.knowledge_gradient.qKnowledgeGradient'>: 2, <class 'botorch.acquisition.knowledge_gradient.qMultiFidelityKnowledgeGradient'>: 3, <class 'botorch.acquisition.max_value_entropy_search.qMaxValueEntropy'>: 4, <class 'botorch.acquisition.max_value_entropy_search.qMultiFidelityMaxValueEntropy'>: 5}, <class 'gpytorch.mlls.marginal_log_likelihood.MarginalLogLikelihood'>: {<class 'gpytorch.mlls.exact_marginal_log_likelihood.ExactMarginalLogLikelihood'>: 0}, <class 'botorch.models.model.Model'>: {<class 'botorch.models.gp_regression.FixedNoiseGP'>: 0, <class 'botorch.models.gp_regression.SingleTaskGP'>: 1, <class 'botorch.models.gp_regression_fidelity.FixedNoiseMultiFidelityGP'>: 2, <class 'botorch.models.gp_regression_fidelity.SingleTaskMultiFidelityGP'>: 3}}¶ Reverse registries for decoding.
-
ax.storage.botorch_modular_registry.MLL_REGISTRY: Dict[Type[gpytorch.mlls.marginal_log_likelihood.MarginalLogLikelihood], int] = {<class 'gpytorch.mlls.exact_marginal_log_likelihood.ExactMarginalLogLikelihood'>: 0}¶ Overarching mapping from encoded classes to registry map.
-
ax.storage.botorch_modular_registry.MODEL_REGISTRY: Dict[Type[botorch.models.model.Model], int] = {<class 'botorch.models.gp_regression.FixedNoiseGP'>: 0, <class 'botorch.models.gp_regression.SingleTaskGP'>: 1, <class 'botorch.models.gp_regression_fidelity.FixedNoiseMultiFidelityGP'>: 2, <class 'botorch.models.gp_regression_fidelity.SingleTaskMultiFidelityGP'>: 3}¶ Mapping of Botorch AcquisitionFunction classes to ints.
-
ax.storage.botorch_modular_registry.REVERSE_MLL_REGISTRY: Dict[int, Type[gpytorch.mlls.marginal_log_likelihood.MarginalLogLikelihood]] = {0: <class 'gpytorch.mlls.exact_marginal_log_likelihood.ExactMarginalLogLikelihood'>}¶ Overarching mapping from encoded classes to reverse registry map.
Utilities¶
-
class
ax.storage.utils.DomainType[source]¶ Bases:
enum.EnumClass for enumerating domain types.
-
CHOICE: int = 2¶
-
FIXED: int = 0¶
-
RANGE: int = 1¶
-
-
class
ax.storage.utils.EncodeDecodeFieldsMap(python_only, encoded_only, python_to_encoded)[source]¶ Bases:
tuple-
property
encoded_only¶ Alias for field number 1
-
property
python_only¶ Alias for field number 0
-
property
python_to_encoded¶ Alias for field number 2
-
property
-
class
ax.storage.utils.MetricIntent[source]¶ Bases:
enum.EnumClass for enumerating metric use types.
-
ADDITIONAL_OBJECTIVE: str = 'additional_objective'¶
-
MULTI_OBJECTIVE: str = 'multi_objective'¶
-
OBJECTIVE: str = 'objective'¶
-
OUTCOME_CONSTRAINT: str = 'outcome_constraint'¶
-
SCALARIZED_OBJECTIVE: str = 'scalarized_objective'¶
-
TRACKING: str = 'tracking'¶
-