API Reference

This section documents PyDSMC's public API. Everything below is importable directly from the top-level pydsmc package:

from pydsmc import (
    Evaluator,
    Property,
    create_predefined_property,
    create_custom_property,
    get_predefined_properties,
    create_eval_envs,
    jsons_to_df,
)

Overview

Symbol	Purpose
Evaluator	Runs Monte-Carlo rollouts of a policy and estimates each registered property until convergence or a resource limit.
Property	A trajectory-based quantity to estimate, together with its statistical method.
create_predefined_property	Build a ready-made property (e.g. return, episode_length).
create_custom_property	Build a property from your own check_fn.
get_predefined_properties	List the available predefined property IDs.
create_eval_envs	Helper to construct vectorized evaluation environments.
jsons_to_df	Aggregate JSON-Lines run logs into a pandas DataFrame.

See Usage for a conceptual walkthrough.