neurocore_skill_neuroweave.skill
================================

.. py:module:: neurocore_skill_neuroweave.skill

.. autoapi-nested-parse::

   NeuroWeave skill — knowledge graph memory for NeuroCore agents.

   Wraps NeuroWeave's async API as a sync FlowEngine Skill with three modes:

   - **process**: Extract knowledge from a message, update the graph.
     Consumes: ``message`` → Provides: ``neuroweave_result``
   - **query**: Query the knowledge graph for relevant context.
     Consumes: ``query`` → Provides: ``neuroweave_result``
   - **context**: Extract + query in one step (default).
     Consumes: ``message`` → Provides: ``neuroweave_result``, ``neuroweave_context``

   The async NeuroWeave API is bridged to sync via ``asyncio.run()``.
   The NeuroWeave instance is created on ``init(config)`` and started
   lazily on the first ``process()`` call.

   Config keys (passed via neurocore_ai.yaml or blueprint):
       mode: str — "process", "query", or "context" (default: "context")
       llm_provider: str — "mock", "anthropic", or "openai" (default: "mock")
       llm_model: str — LLM model identifier
       llm_api_key: str — API key (or set via ANTHROPIC_API_KEY env)
       enable_visualization: bool — start the graph viz server (default: false)



Classes
-------

.. autoapisummary::

   neurocore_skill_neuroweave.skill.NeuroWeaveSkill


Module Contents
---------------

.. py:class:: NeuroWeaveSkill(name: str | None = None)

   Bases: :py:obj:`neurocore_ai.skills.base.Skill`


   Knowledge graph memory skill powered by NeuroWeave.

   Bridges NeuroWeave's async API into NeuroCore's synchronous
   FlowEngine execution model. Manages a persistent NeuroWeave
   instance across the skill lifecycle.


   .. py:attribute:: skill_meta
      :type:  ClassVar[neurocore_ai.skills.base.SkillMeta]


   .. py:method:: init(config: dict[str, Any]) -> None

      Initialize the skill and create (but don't start) the NeuroWeave instance.



   .. py:method:: setup(context: flowengine.FlowContext) -> None

      Start NeuroWeave on the first run.



   .. py:method:: process(context: flowengine.FlowContext) -> flowengine.FlowContext

      Execute the NeuroWeave operation based on configured mode.

      Args:
          context: FlowContext with input data.

      Returns:
          FlowContext with neuroweave results.



   .. py:method:: teardown(context: flowengine.FlowContext) -> None

      Stop the NeuroWeave instance.



   .. py:method:: health_check() -> bool

      Check if NeuroWeave is operational.