reagent.workflow package

Submodules

reagent.workflow.cli module

reagent.workflow.cli.select_relevant_params(config_dict, ConfigClass)

reagent.workflow.data_fetcher module

reagent.workflow.env module

reagent.workflow.env.get_workflow_id() → int

reagent.workflow.gym_batch_rl module

reagent.workflow.identify_types_flow module

reagent.workflow.result_registries module

class reagent.workflow.result_registries.PublishingResult(success: bool)

Bases: object

REGISTRY = {'no_publishing_results': <class 'reagent.workflow.result_types.NoPublishingResults'>}
REGISTRY_FROZEN = False
REGISTRY_NAME = 'PublishingResult'
class reagent.workflow.result_registries.TrainingReport

Bases: object

REGISTRY = {'actor_critic_report': <class 'reagent.workflow.training_reports.ActorCriticTrainingReport'>, 'dqn_report': <class 'reagent.workflow.training_reports.DQNTrainingReport'>, 'parametric_dqn_report': <class 'reagent.workflow.training_reports.ParametricDQNTrainingReport'>}
REGISTRY_FROZEN = False
REGISTRY_NAME = 'TrainingReport'
class reagent.workflow.result_registries.ValidationResult(should_publish: bool)

Bases: object

REGISTRY = {'no_validation_results': <class 'reagent.workflow.result_types.NoValidationResults'>}
REGISTRY_FROZEN = False
REGISTRY_NAME = 'ValidationResult'

reagent.workflow.result_types module

class reagent.workflow.result_types.NoPublishingResults(success: bool)

Bases: reagent.workflow.result_registries.PublishingResult

class reagent.workflow.result_types.NoValidationResults(should_publish: bool)

Bases: reagent.workflow.result_registries.ValidationResult

reagent.workflow.spark_utils module

reagent.workflow.spark_utils.SPARK_JAR_FROM_ROOT_DIR = 'preprocessing/target/rl-preprocessing-1.1.jar'

SPARK_JAR is abspath to the above jar file.

Assume file structure ReAgent/

preprocessing/… reagent/…

reagent.workflow.spark_utils.call_spark_class(spark, class_name: str, args: str)
reagent.workflow.spark_utils.get_spark_session(config: Optional[Dict[str, str]] = {'spark.app.name': 'ReAgent', 'spark.driver.extraClassPath': '/Users/jjg/github/ReAgent/reagent/../preprocessing/target/rl-preprocessing-1.1.jar', 'spark.driver.host': '127.0.0.1', 'spark.master': 'local[*]', 'spark.sql.catalogImplementation': 'hive', 'spark.sql.execution.arrow.enabled': 'true', 'spark.sql.session.timeZone': 'UTC', 'spark.sql.shuffle.partitions': '12', 'spark.sql.warehouse.dir': '/Users/jjg/github/ReAgent/docs/spark-warehouse'})
reagent.workflow.spark_utils.get_table_url(table_name: str) → str

reagent.workflow.tagged_union module

reagent.workflow.training module

reagent.workflow.training_reports module

class reagent.workflow.training_reports.ActorCriticTrainingReport

Bases: reagent.workflow.result_registries.TrainingReport

class reagent.workflow.training_reports.DQNTrainingReport(td_loss: Union[float, NoneType] = None, mc_loss: Union[float, NoneType] = None, reward_ips: Union[reagent.evaluation.cpe.CpeEstimate, NoneType] = None, reward_dm: Union[reagent.evaluation.cpe.CpeEstimate, NoneType] = None, reward_dr: Union[reagent.evaluation.cpe.CpeEstimate, NoneType] = None, value_sequential_dr: Union[reagent.evaluation.cpe.CpeEstimate, NoneType] = None, value_weighted_dr: Union[reagent.evaluation.cpe.CpeEstimate, NoneType] = None, value_magic_dr: Union[reagent.evaluation.cpe.CpeEstimate, NoneType] = None)

Bases: reagent.workflow.result_registries.TrainingReport

mc_loss = None
reward_dm = None
reward_dr = None
reward_ips = None
td_loss = None
value_magic_dr = None
value_sequential_dr = None
value_weighted_dr = None
class reagent.workflow.training_reports.ParametricDQNTrainingReport

Bases: reagent.workflow.result_registries.TrainingReport

reagent.workflow.types module

reagent.workflow.utils module

Module contents