The Open States project collects and makes available data about state legislative activities, including bill summaries, votes, sponsorships and state legislator information. This data is gathered directly from the states and made available in a common format for interested developers, through a JSON API and data dumps.
We use Docker to provide a reproducible development environment. Make sure you have Docker installed.
Inside your Open States repository directory, run a scrape for a specific state by running:
docker-compose run --rm scrape <state postal code>
You can also choose to scrape only specific data types for a state. The data types available vary from state to state; look at the scrapers
listed in the state's __init__.py
for a list. For example, Tennessee (with a state postal code of tn
) has:
scrapers = {
'bills': TNBillScraper,
'committees': TNCommitteeScraper,
'events': TNEventScraper,
'people': TNPersonScraper,
}
So you can limit a Tennessee scrape to only committees and legislators using:
docker-compose run --rm scrape tn committees people
After retrieving everything from the state, scrape
imports the data into a PostgreSQL database. If you want to skip this step, include a --scrape
flag at the end of your command, like so:
docker-compose run --rm scrape tn committees people --scrape
If you do want to import data into Postgres, start a Postgres service using Docker Compose:
docker-compose up postgres
Then run database migrations and import jurisdictions, thus initializing the database contents:
docker-compose run --rm dbinit
Now you can run the scrape service without the --scrape
flag, and data will be imported into Postgres. You can connect to the database and inspect data using psql
(credentials are set in docker-compose.yml
):
psql postgres://postgres:secret@localhost:5432/openstates
After you run scrape
(with or without the Postgres import), it will leave one JSON file in the _data
subdirectory for each entity that was scraped. These JSON files contain the transformed, scraped data, and are very useful for debugging.
Check out the writing scrapers guide to understand more about how the scrapers work, and how you can contribute.
Our scraping framework, Pupa, has a strong test harness, and requires well-structured data when ingesting. Furthermore, Open States scrapers should be written to fail when they encounter unexpected data, rather than guessing at its format and possibly ingesting bad data. Together, this means that there aren't many benefits to writing unit tests for particular Open States scrapers, versus high upkeep costs.
Occasionally, states will have unit tests, though, for specific structural cases. To run all tests:
docker-compose run --rm --entrypoint=nosetests scrape /srv/openstates-web/openstates
A few states require credentials to access their APIs. If you want to run code for these states, get credentials for yourself using these steps:
- NY
- Get credentials at: https://legislation.nysenate.gov/#signup
- Set in environment prior to running scrape:
NEW_YORK_API_KEY
- IN
- Get credentials at: http://docs.api.iga.in.gov/introduction.html#security-and-authentication
- Set in environment prior to running scrape:
INDIANA_API_KEY
- As a side note, Indiana also requires a user-agent string, so set that in your environment as well, prior to running scrape:
USER_AGENT