Adds ids mapping to GPO's Pictorial Member Guide
pictorialapi.gpo.gov
This PR includes ids for all members of 118th Congress. It may also work
for historical files -- at least going back to 110 -- but I have not
included that work here, as it requires manual fixes and I imagine most
interest is in current membership.
Closes#942
This can happen when new legislators are added before a bioguide ID is published by the House, especially when staging election results before the next Congress has begun.
* Add some basic checks.
* If a TWITTER_API_BEARER_TOKEN environment variable is set, query the Twitter API to check that Twitter usernames and IDs match, and if usernames are in canonical case.
* Fix Rep. Coons's twitter_id to match the twitter handle. The existing ID corresponds to the Twitter handle SenCoonsOffice which also appears to be a correct account, but his website links to the ChrisCoons account.
* Update various Twitter handles to TitleCase if the account itself uses TitleCase
This data came from my original import from GovTrack's legislator database, and the religion field probably came from my original import of data from the MIT Media Lab's Government Information Awareness project in 2003. The field was never maintained.
Data from https://www.senate.gov/senators/AppointedSenators.htm. Where the appointment date was after our term start date, the term start date is updated. (In other cases I presume our term start date to correctly reflect the swearing-in date or the start of the session.) Where there was ambiguity on the Senate page about appointment date versus effective date, I used the effective date.
see #41
* merged generate_json.py and alternate_bulk_formats.py and generate pretty JSON
* have them write to ../ rather than ../alternate_formats
* delete the old bulk data files since they'll be in gh-pages
* add CircleCI file to manage this
* add scripts/update_gh_pages.sh which updates the gh-pages branch with the latest bulk data files in multiple formats
* remove test/test_json_matches.py because it's no longer needed
* add links to downloadable files in README
All id types are white-listed so it has to be added, but the ids aren't unique assigned so that test is skipped for google_entity_ids for now, and the uniqueness test is improved to better report the duplicates.
