Bill versions? about congress HOT 5 CLOSED

On 09/04/2012 05:09 PM, Derek Willis wrote:

Do we want to include info about different versions of a bill? Some
bills don't have more than one, but for others, Thomas shows something
like this:

http://thomas.loc.gov/cgi-bin/query/z?c112:S.978:

—
Reply to this email directly or view it on GitHub
https://github.com/congress/thomas/issues/11.

I've only used that for bill text, which could/should be a separate scraper?

• Josh Tauberer

http://razor.occams.info | @JoshData

from congress.

Yeah, I think this is information best gotten from GPO. I have two parsers that work with bill versions and text -

One uses FDSys' crawlable sitemaps (a feature Josh pointed out to me a while back) to download and keep in sync with all of the bill versions that get posted to their bills collection:
https://github.com/sunlightlabs/realtimecongress/blob/master/tasks/bulk_gpo_bills/bulk_gpo_bills.rb

The other reads those in (their text and their associated MODS XML file), pulls out version information and attaches it to any existing bill information I got from GovTrack/THOMAS, and does various kinds of post-processing (like US Code citation extraction):
https://github.com/sunlightlabs/realtimecongress/blob/master/tasks/bill_text_archive/bill_text_archive.rb

It might be good to publish this version information as part of this project's output. I've seen GPO run behind THOMAS in posting bill text, so it'd be best for the THOMAS-collection part to not depend on version info being posted on GPO, and for bill version information to be pulled and published separately from (but alongside) the regular bill data.

If someone feels like porting my Ruby code into a bill_versions.py task that syncs with GPO and publishes metadata and downloaded text for each bill version, that would certainly fit well with the project.

from congress.

Or - it could be its own small repo, since it's not "thomas" data. It's not a scrape job, it won't need unit tests, its sole purpose is to keep a predictable dataset up to date. That could even free it up to be another language. As long as each repo, when deployed, uses the same directory for outputting data, it would be pretty seamless.

from congress.

In fact, for a while I've thought about making a slightly generalized FDSys syncing service, since each collection is structured so similarly and could be crawled the same way. I'm getting ahead of myself, but doing that (this would make it easier to get committee reports and treaties, as well) is possibly a reason to consider making a separate 'gpo' repository.

from congress.

I'm gonna close this one for the time being, feel free to re-open it if you want to talk through it more.

from congress.