The pipeline created in the pipeline check was actually running on the Dummy data source.  I've created a NoData datasource that never
actually runs the pipeline.

There are now two crate features "python" which enables the python module interface, and "python-pipe", which enables the custom pipelines
written in python.