# tfma.metrics.Preprocessor.BundleContextParam

<br />

Allows one to use a context manager to manage bundle-scoped parameters.

#### View aliases


**Main aliases**

[`tfma.utils.DoFnWithModels.BundleContextParam`](https://www.tensorflow.org/tfx/model_analysis/api_docs/python/tfma/metrics/Preprocessor/BundleContextParam)

<br />

    tfma.metrics.Preprocessor.BundleContextParam(
        context_manager_constructor, args=(), kwargs=None, *, name=None
    )

The context will be entered at the start of each bundle and exited at the
end, equivalent to the `start_bundle` and `finish_bundle` methods on a DoFn.

The object returned from `__enter__`, if any, will be substituted for this
parameter in invocations. Multiple context manager parameters may be
specified which will all be evaluated (in an unspecified order).

This can be especially useful for setting up shared context in transforms
like `Map`, `FlatMap`, and `Filter` where one does not have start_bundle
and finish_bundle methods.

## Methods

### `create_and_enter`

    create_and_enter()