Add ability to execute multiple importers serially

We can have a configuration file that describes arguments and order of importers to execute. That'd would make it easier to separate importer functionality more freely, specially in cases of DB where frequent datasource access is not a problem ( unlike file datasources ).

The issue names no files, tests, entry points, or configuration format. Start by locating the importer execution entry point and existing configuration handling; done should be demonstrated by tests showing configured importers run serially in order with their arguments, including the stated database and file-datasource cases.

Written by the indexing model from the issue text.