rlang package

rlang.core module

This module contains simple user-facing functions for parsing RLang into an RLangKnowledge object, which can be provided to an RLang-enabled agent.

rlang.core.parse_file(rlang_fname, prior_knowledge=None)[source]

Parses an .rlang file into an RLangKnowledge object.

Parameters

  • rlang_fname (str) – filename

  • prior_knowledge (Optional[RLangKnowledge]) – prior knowledge that should be retained after parsing

Return type

RLangKnowledge

rlang.core.parse(rlang, prior_knowledge=None)[source]

Parses an rlang string into an RLangKnowledge object.

Parameters

  • rlang (str) – string containing rlang

  • prior_knowledge (Optional[RLangKnowledge]) – prior knowledge that should be retained after parsing

Return type

RLangKnowledge

rlang.knowledge module

Every RLang program (including any vocabulary files) grounds to an RLangKnowledge object.

class rlang.knowledge.RLangKnowledge[source]

Bases: MutableMapping

Provides an interface for accessing stored RLang information. Behaves similarly to a Python dictionary.

Note

In typical usage, an RLangKnowledge object is not instantiated by the user but is instead returned from a call to parse_file() or parse().

Examples:

base = RLangKnowledge()
base['x_location'] = Factor([1])
__init__()[source]
policy

A Policy object

reward_function

A RewardFunction object

transition_function

A TransitionFunction object

predictions(*args, **kwargs)[source]

Get a dictionary of Grounding objects whose value for the next state can be predicted using the keyword arguments provided.

Parameters

  • state (Optional[State]) – a given current state

  • action (Optional[Action]) – a given action

Return type

Dict[Grounding, Any]

Subpackages