openva / rs-api Goto Github PK

Caching obstructions should match reality. Prior years’ legislation should have long expire times, but this year’s active legislation should have times set relative to when the bulk data is updated.

• issue a query for a list of all bills for a year
• test that there are a large number of results
• test that the right fields are present in the results

e.g., https://api.richmondsunlight.com/1.1/legislator/erfiller-corn.json fails, despite that this is a valid URL structure for Eileen Filler-Corn. It's gotta be in htdocs/1.1/.htaccess. Loosen up the regex to support all valid legislator URL slugs.

This is a problem, of course. Fix it.

Just to list all comments for a given bill.

• capture the JSON returned for a query for a bill, save that as a static file in the repo
• create a test that will issue a query for that bill
• deep-compare the two objects
• if they don't match, error out, highlighting precisely how they don't match

There's no need to have every request hitting the server. Set up a CloudFront distribution for these requests.

• Review the work done in #1, figure out what went wrong
• Set up proper cache headers in API responses (#2)
• Enable the CloudFront distribution

The Content-type header is text/html. I can't grasp why, because Header add Content-type "application/json" is in .htaccess. Figure out what's up.

This is clearly not the way to handle this.

AWS Certificate Manager can issue the certificate, as long as the site is behind Cloudfront. (Or other services, but that's the viable one here.) You can't load it onto an EC2 instance. Figure out how to do that, to make the site portable, off its current server.

• list every legislator who we have a record of
• include sufficient metadata for the requester to determine programmatically which ones are in office now
• provide a filter to restrict the query to current legislators
• document it

Per https://twitter.com/mjgoldsmith/status/963598295661776896

• create an endpoint for a single vote
• modify the bill endpoint to include vote IDs, and possibly the URL for the API call for that vote information
• perhaps create an endpoint to list all votes for a given year (that might be better as bulk data)

• capture the JSON returned for a query for a retired legislator, save that as a static file in the repo
• create a test that will issue a query for that legislator
• deep-compare the two objects
• if they don't match, error out, highlighting precisely how they don't match

• index page lists all districts, with the chamber, number, internal ID, and place names for each
• each district endpoint should list all data that we have about that endpoint

Include:

• all legislators
• all committees
• all legislators-committees
• several bills
• all cosponsors for those bills

• capture the JSON returned for a query for a portfolio, save that as a static file in the repo
• create a test that will issue a query for that portfolio
• deep-compare the two objects
• if they don't match, error out, highlighting precisely how they don't match