caikit.runtime.utils.import_util
================================

.. py:module:: caikit.runtime.utils.import_util

.. autoapi-nested-parse::

   A generic module to help import dynamic modules



Attributes
----------

.. autoapisummary::

   caikit.runtime.utils.import_util.log


Classes
-------

.. autoapisummary::

   caikit.runtime.utils.import_util.UnifiedDataModel


Functions
---------

.. autoapisummary::

   caikit.runtime.utils.import_util.get_data_model
   caikit.runtime.utils.import_util._get_cdm_from_lib
   caikit.runtime.utils.import_util.get_dynamic_module
   caikit.runtime.utils.import_util.clean_lib_names


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

.. py:data:: log

.. py:class:: UnifiedDataModel

   This class acts as an aggregator between the data models of multiple
   Caikit libraries.


   .. py:attribute:: _libraries


   .. py:method:: add_library(lib_name: str, lib: types.ModuleType)

      Add the given library to the unified data model



   .. py:method:: __getattr__(name: str) -> Any

      Fetch an attribute by name aliasing into child libraries



.. py:function:: get_data_model(config: aconfig.Config = None) -> UnifiedDataModel

   Get the data model from the Caikit library of interest. This is accomplished
   via dynamic import on the caikit_library's environment variable.

   NOTE: This function also has the side-effect of importing each of the
       caikit_library libraries for the first time, causing their modules to
       be registered with the caikit.core module registry. It is a critical
       step in initializing the set of modules that can be loaded by this
       running server instance.

   Args:
       config(aconfig.Config): caikit configuration

   Returns:
       (module): Handle to the module after dynamic wild import


.. py:function:: _get_cdm_from_lib(lib_name: str, cdm: UnifiedDataModel)

   Function to get caikit core CDM from library name

   Args:
       lib_name (str): Caikit core library name
       cdm (UnifiedDataModel): Caikit core CDM
   Returns:
       cdm: UnifiedDataModel


.. py:function:: get_dynamic_module(module_name: str, module_dir: str = None) -> types.ModuleType

   Get the dynamic module of interest.

   Args:
       module_name(str): Name of the module to be dynamically imported
       (Optional) module_dir(str): Name of the directory from where the module is
               to be dynamically imported

   Returns:
       (module): Handle to the module after dynamic import


.. py:function:: clean_lib_names(caikit_library: str) -> List[str]