:class:`FluidsLeafMCP`
======================


.. py:class:: ansys.fluent.mcp.common.base.FluidsLeafMCP(*, backends: dict[str, ansys.fluent.mcp.common.backend.Backend], expose_tools: Iterable[str] = ALL_TOOLS, name: Optional[str] = None, **fastmcp_kwargs: Any)

   Bases: :py:obj:`ansys.common.mcp.server.PyAnsysBaseMCP`


   
   Base server class for a single Fluids MCP leaf.
















   ..
       !! processed by numpydoc !!


.. py:currentmodule:: FluidsLeafMCP

Overview
--------

.. tab-set::



   .. tab-item:: Methods

      .. list-table::
          :header-rows: 0
          :widths: auto

          * - :py:attr:`~run`
            - Start the MCP server via the parent implementation when available.
          * - :py:attr:`~product_startup`
            - No persistent product session at startup. Backends connect lazily.
          * - :py:attr:`~product_cleanup`
            - Disconnect any active backend at shutdown.
          * - :py:attr:`~build_toolsets`
            - Build toolset definitions filtered to the tools this leaf exposes.
          * - :py:attr:`~register_run_code_observer`
            - Register a callable invoked after every ``run_code`` tool call.


   .. tab-item:: Properties

      .. list-table::
          :header-rows: 0
          :widths: auto

          * - :py:attr:`~backend`
            - Return the active backend instance.


   .. tab-item:: Attributes

      .. list-table::
          :header-rows: 0
          :widths: auto

          * - :py:attr:`~leaf_name`
            - 
          * - :py:attr:`~default_backend_kind`
            - 
          * - :py:attr:`~cache_ttl_seconds`
            - 
          * - :py:attr:`~component_label`
            - 






Import detail
-------------

.. code-block:: python

    from ansys.fluent.mcp.common.base import FluidsLeafMCP

Property detail
---------------

.. py:property:: backend
   :type: ansys.fluent.mcp.common.backend.Backend


   
   Return the active backend instance.





   :Returns:

       :obj:`Backend`
           Backend produced by the operation.











   ..
       !! processed by numpydoc !!



Attribute detail
----------------

.. py:attribute:: leaf_name
   :type:  str
   :value: 'fluids'


.. py:attribute:: default_backend_kind
   :type:  Optional[str]
   :value: None


.. py:attribute:: cache_ttl_seconds
   :type:  float
   :value: 30.0


.. py:attribute:: component_label
   :type:  str
   :value: ''




Method detail
-------------

.. py:method:: run(transport: str = 'stdio', host: str | None = None, port: int | None = None) -> None

   
   Start the MCP server via the parent implementation when available.


   :Parameters:

       **transport** : :class:`python:str`
           Transport name used to run the MCP server.

       **host** : :class:`python:str` | :data:`python:None`
           Host interface used by the server transport.

       **port** : :class:`python:int` | :data:`python:None`
           Port used by the selected server transport.



   :Returns:

       :data:`python:None`
           The function completes through its side effects.











   ..
       !! processed by numpydoc !!

.. py:method:: product_startup() -> None

   
   No persistent product session at startup. Backends connect lazily.





   :Returns:

       :data:`python:None`
           The function completes through its side effects.











   ..
       !! processed by numpydoc !!

.. py:method:: product_cleanup() -> None

   
   Disconnect any active backend at shutdown.





   :Returns:

       :data:`python:None`
           The function completes through its side effects.











   ..
       !! processed by numpydoc !!

.. py:method:: build_toolsets() -> list[dict[str, Any]]

   
   Build toolset definitions filtered to the tools this leaf exposes.

   Returns a list of toolset dicts conforming to the
   PyAnsysMCPService conductor contract.




   :Returns:

       :class:`python:list`\[:class:`python:dict`\[:class:`python:str`, :obj:`Any`]]
           Mapping containing the operation result.











   ..
       !! processed by numpydoc !!

.. py:method:: register_run_code_observer(observer: Callable[Ellipsis, Awaitable[None] | None]) -> None

   
   Register a callable invoked after every ``run_code`` tool call.

   The observer is called as ``observer(code=..., result=...,
   error=...)``. ``result`` is the backend's :class:`RunCodeResult`
   on success (and ``None`` on exception); ``error`` is the raised
   exception on failure (and ``None`` on success). Observers may be
   sync or async. Exceptions raised by an observer are logged and
   swallowed so the original tool result is preserved.

   :Parameters:

       **observer** : :obj:`Callable`\[..., :obj:`Awaitable`\[:data:`python:None`] | :data:`python:None`]
           Observer callback invoked after run-code execution.



   :Returns:

       :data:`python:None`
           The function completes through its side effects.











   ..
       !! processed by numpydoc !!