Missed TensorFlow Dev Summit? Check out the video playlist. Watch recordings

tfx.orchestration.experimental.interactive.interactive_context.InteractiveContext

View source on GitHub

TFX interactive context for interactive TFX notebook development.

tfx.orchestration.experimental.interactive.interactive_context.InteractiveContext(
    pipeline_name=None, pipeline_root=None, metadata_connection_config=None
)

Used in the notebooks

Used in the tutorials

Args:

  • pipeline_name: Optional name of the pipeline for ML Metadata tracking purposes. If not specified, a name will be generated for you.
  • pipeline_root: Optional path to the root of the pipeline's outputs. If not specified, an ephemeral temporary directory will be created and used.
  • metadata_connection_config: Optional metadata_store_pb2.ConnectionConfig instance used to configure connection to a ML Metadata connection. If not specified, an ephemeral SQLite MLMD connection contained in the pipeline_root directory with file name "metadata.sqlite" will be used.

Methods

export_to_pipeline

View source

export_to_pipeline(
    *args, **kwargs
)

Exports a notebook to a .py file as a runnable pipeline.

Args:

  • notebook_filepath: String path of the notebook file, e.g. '/path/to/notebook.ipynb'.
  • export_filepath: String path for the exported pipeline python file, e.g. '/path/to/exported_pipeline.py'.
  • runner_type: String indicating type of runner, e.g. 'beam', 'airflow'.

run

View source

run(
    *args, **kwargs
)

Run a given TFX component in the interactive context.

Args:

  • component: Component instance to be run.
  • enable_cache: whether caching logic should be enabled in the driver.
  • beam_pipeline_args: Optional Beam pipeline args for beam jobs within executor. Executor will use beam DirectRunner as Default.

Returns:

execution_result.ExecutionResult object.

show

View source

show(
    *args, **kwargs
)

Show the given object in an IPython notebook display.