Filter tables in gtfs_schedule dataset to be only those in current agencies.yml

[machow]

Related to (but not blocking) #233.

Currently, the gtfs_schedule contains the most recent data for all GTFS schedule feeds in the warehouse. This includes feeds that have been deleted from agencies.yml. We should...

• use gtfs_schedule_history.calitp_status to mark in the gtfs_schedule_history.calitp_feed_latest when a feed is in the latest status (which is a tabled form of agencies.yml). E.g. create a is_in_agencies column.
• create a gtfs_schedule.calitp_feeds table, which contains calitp_itp_id, calitp_url_number, agency_name, status, calitp_extracted_at for the latest calit_status (e.g. basically the same as views.gtfs_agency_names)
• Filter on is_in_agencies in gtfs_schedule_loader DAG task gtfs_schedule_load_tables to only load current feeds.