Allow for dynamic module to handle scrape output.

[doubleswirve]

Howdy,

This PR allows users to specify a module to handle scrape objects as an alternative to writing them to files. This user module name can be specified via the environment variable SCRAPE_OUTPUT_HANDLER. Once imported, the module expects the following class structure to be available:

class Handler:

    def __init__(self, scraper):
        # Custom code...
        pass

    def handle(self, obj):
        # Custom code...
        pass

This is related to opencivicdata/pupa#315 but is much lighter and does not impose additional dependencies.

Thanks!

[jamesturk]

👍