caikit.interfaces.ts.data_model.toolkit.optional_dependencies

This module encapsulates the core optional dependencies using lazy imports.

Attributes

`log`
`pd`
`pyspark`
`HAVE_NUMPY`
`HAVE_PANDAS`
`HAVE_PYSPARK`

Classes

LazyModule

A LazyModule is a module subclass that wraps another module but imports

Functions

have_module(→ bool)

This method can be used to check whether a given optional dependency is

Module Contents

caikit.interfaces.ts.data_model.toolkit.optional_dependencies.log[source]

class caikit.interfaces.ts.data_model.toolkit.optional_dependencies.LazyModule(name: str, package: str | None = None)[source]

Bases: types.ModuleType

A LazyModule is a module subclass that wraps another module but imports it lazily and then aliases __getattr__ to the lazily imported module.

__name

__package = None

__wrapped_module = None

__getattr__(name: str) → any[source]: When asked for an attribute, make sure the wrapped module is imported and then delegate

caikit.interfaces.ts.data_model.toolkit.optional_dependencies.have_module(name: str, package: str | None = None) → bool[source]

This method can be used to check whether a given optional dependency is available and should primarily be used for assertions when coding defensively.

NOTE: Nested modules WILL force the import of parent modules

TODO: Move this to import_tracker

Args:

name: str: The name of the module
package: Optional[str]: The qualifying package for the module under investigation

Returns:

have_module: bool: True if the module can be imported, False otherwise

caikit.interfaces.ts.data_model.toolkit.optional_dependencies.pd

caikit.interfaces.ts.data_model.toolkit.optional_dependencies.pyspark

caikit.interfaces.ts.data_model.toolkit.optional_dependencies.HAVE_NUMPY

caikit.interfaces.ts.data_model.toolkit.optional_dependencies.HAVE_PANDAS

caikit.interfaces.ts.data_model.toolkit.optional_dependencies.HAVE_PYSPARK