GTFS Schedule Schema: add filesize and calendar range metadata

[mil]

This PR adds 3 new system generated metadata properties to the bounding_box specification for newly added GTFS schedule sources:

• extracted_filesize: indicates GTFS archive filesize in bytes
• extracted_calendar_start: indicates GTFS archive calendar/calendar_dates min date
• extracted_calendar_end: indicates GTFS archive calendar/calendar_dates max date

These properties function similarly to the existing extracted_on metadata property in the schema.

These 3 new metadata properties (extracted_filesize, extracted_calendar_start, and extracted_calendar_end) would be very helpful for end-consumers for certain applications. For example, with respect to end-users understanding general GTFS filesize consider consuming an archive that is 10MB is very different from consuming an archive that is 500MB in both download & processing time; as such it would be very helpful for end-consumers to know this stat flagged ahead-of-time. (One example of an application that would benefit from this change is my android app, Transito, which consumes GTFS indicated from MDB; and I would greatly appreciate the ability to pass for example filesize metadata along to my end-users). Additionally the calendar start/end range would help in historically understanding when the source was updated/added what the original calendar range was. While this PR only adds the 3 new properties; followup PR(s) could address updating existing sources and all new sources would have this metadata by default once applied.

In addition to the updated tests, if you just want to quickly test to see what the new format will look like for a sample, you can use for example:

add_gtfs_schedule_source(provider="foo", "country_code="bar", direct_download_url='http://data.trilliumtransit.com/gtfs/cedarrapids-ia-us/cedarrapids-ia-us.zip')