Add concept of "Data Pipelines"

[acouch]

Would like to have an API to apply data to documents. Examples include GA stats, #19, #20 and so on.

Would like to align this with https://github.com/frictionlessdata/datapackage-pipelines if possible though that appears to be a "per file" spec and not one that is something that is applied to a large number of files. Will try and ask them about that.

Might want to have a separate pipelines folder where different pipelines can be declared and then have a pipelines.yml file or something similar where they can be declared with arguments like the collection and frequency.

This maybe should be called "tasks" or "cron" or something but going with "Data Pipelines" for now.

Issue Type

• Bug (https://github.com/interra/catalog-generate/blob/master/CONTRIBUTING.md#bug-reports)
• Feature (https://github.com/interra/catalog-generate/blob/master/CONTRIBUTING.md#feature-requests)