BaseSampler

BaseSampler is a generic class designed for sampling environments and agents in a parallel simulation framework. It provides a method to sample an environment and a list of agents based on specified parameters.

Class Definition

class BaseSampler(Generic[ObsType, ActType]):
    def __init__(
        self,
        env_candidates: Sequence[EnvironmentProfile | str] | None = None,
        agent_candidates: Sequence[AgentProfile | str] | None = None,
    ) -> None:
        self.env_candidates = env_candidates
        self.agent_candidates = agent_candidates

Parameters

• env_candidates (Sequence[EnvironmentProfile | str] | None, optional): A sequence of environment profiles or strings. Defaults to None.
• agent_candidates (Sequence[AgentProfile | str] | None, optional): A sequence of agent profiles or strings. Defaults to None.

Methods

sample

def sample(
    self,
    agent_classes: Type[BaseAgent[ObsType, ActType]]
    | list[Type[BaseAgent[ObsType, ActType]]],
    n_agent: int = 2,
    replacement: bool = True,
    size: int = 1,
    env_params: dict[str, Any] = {},
    agents_params: list[dict[str, Any]] = [{}, {}],
) -> Generator[EnvAgentCombo[ObsType, ActType], None, None]:

Description

Sample an environment and a list of agents.

Parameters

• agent_classes (Type[BaseAgent[ObsType, ActType]] | list[Type[BaseAgent[ObsType, ActType]]]): A single agent class for all sampled agents or a list of agent classes.
• n_agent (int, optional): Number of agents. Defaults to 2.
• replacement (bool, optional): Whether to sample with replacement. Defaults to True.
• size (int, optional): Number of samples. Defaults to 1.
• env_params (dict[str, Any], optional): Parameters for the environment. Defaults to {}.
• agents_params (list[dict[str, Any]], optional): Parameters for the agents. Defaults to [{}, {}].

Returns

• Generator[EnvAgentCombo[ObsType, ActType], None, None]: A generator yielding tuples containing an environment and a list of agents.

Usage Example

from sotopia.agents.base_agent import BaseAgent
from sotopia.database.persistent_profile import AgentProfile, EnvironmentProfile
from sotopia.envs.parallel import ParallelSotopiaEnv
 
# Define a custom agent class inheriting from BaseAgent
class CustomAgent(BaseAgent):
    pass
 
# Initialize the BaseSampler
sampler = BaseSampler()
 
# Sample an environment and agents
samples = sampler.sample(agent_classes=[CustomAgent], n_agent=3, size=5)
 
# Iterate over the generated samples
for env, agents in samples:
    print(f"Environment: {env}")
    print(f"Agents: {agents}")

Note: The sample method raises NotImplementedError and must be implemented in a subclass to function correctly.