Package to scrape and process UK Parliamentary bills, votes and MP contact information.
Written as part of University of Bath Integrated Project module.
pipenv install ParlPy~=2.6.0
A list of bills to collect information about is generated using either the update_all_bills_in_session
or the
get_changed_bills_in_session
methods of a BillsOverview object from parlpy.bills.bill_list_fetcher. The former
simply gets a list of all bills and places them in member DataFrame bills_overview_data
. The latter has persistence,
so that its first run behaves like update_all_bills_in_session
, but after this places only bills updated after it
was last run in the DataFrame.
The BillsOverview object is then passed to iterator parlpy.bills.bill_details_iterator.get_bill_details(), which yields an object containing information on the bill, including divisions data
import parlpy.bills.bill_list_fetcher as blf
import parlpy.bills.bill_details_iterator as bdi
test_fetcher = blf.BillsOverview()
test_fetcher.get_changed_bills_in_session()
for s in bdi.get_bill_details(test_fetcher):
print(f"summary {s.summary}")
for d in s.divisions_list:
print(f"division name {d.division_name}")
print(f"division stage {d.division_stage}")
print(f"ayes {d.ayes}")
print(f"noes {d.noes}")
# ...
# repeat, call get_changed_bills_in_session() again (on the same object or new, the pickle works either way), then
# pass the object to the iterator to get new results
The above gets summary and a list of DivisionDetail objects for each bill. BillsOverview has persistence, storing the
last time get_changed_bills_in_session
was called using a pickle variable. So eg the script can be run, the values
stored in a DB, then run again later to get info on bills that the parliamentary website says have updated.
parlpy.bills.bill_details_iterator for fetching bill data
Iterator that yields BillDetails object
Instance variables
- self.title_stripped: str
- self.title_postfix: str
- self.sessions: List[str]
- self.summary: str
- self.divisions_list: List[parlpy.bills.bill_votes_fetcher.DivisionInformation]
- self.url: str
- self.last_updated: datetime.datetime
Not intended for public API use, BillsOverview object is instead passed to parlpy.bills.bill_details_iterator.get_bill_details()
Constructs an object for collecting basic information on all bills in the current Parliamentary session. Its purpose is to
- get a list of existing bills - so that we know what to scrape
- get path to further details for each bill - so that we can assemble the url to scrape further info from
- be passed to parlpy.bills.bill_details_iterator.get_bill_details()
Instance variables:
-
self.bills_overview_data -> pd.DataFrame
DF containing bill titles, their last updated time and the path to further details on the bill. Determines which bills parlpy.bills.bill_details_iterator.get_bill_details() collects data on
Public instance methods:
-
get_changed_bills_in_session(session_name="2019-21", fetch_delay=0) -> None
-
reset_datetime_last_scraped() -> None
-
update_all_bills_in_session(session_name="2019-21", fetch_delay=0) -> None
Method to called to update self.bills_overview_data, fetching pages at maximum rate when fetch_delay=0
A list of extant bills and acts are scraped from https://bills.parliament.uk. Scraping was the only choice for collecting this list as outlined in the bullet points below
-
There is an official RSS feed, however this only lists the 50 most recently updated bills, so we do not use this source as we would not have been able to collect older bills.
-
The API at https://explore.data.parliament.uk/?endpoint=bills does not contain bills more recent than 2018, so we do not utilise this as it does not present new bills.
-
The recent API https://developer.parliament.uk/ does not provide bills.
The package however does make use of the official UK Parliamentary API for collecting division data and for collecting data on MP's.
BillsOverview.get_changed_bills_in_session()
and BillsOverview.update_all_bills_in_session()
Both scrape data from https://bills.parliament.uk
get_summary()
Scrapes data from https://bills.parliament.uk
- add originating house data to overview and details
- correct demo program in README
- add 2021-22 parliamentary session codes to scraper
- updates to account for recent changes to parliamentary website
- include optional app engine pickling code
- include method in BillsOverview to mock a datetime last scraped pickle
- fix: bills/act titles not conforming to standard layout breaking the bill list fetcher, they are now not added
- mp data includes if they are current and their email
- mp data includes constituency
- iterator return object now includes url and last updated datetime
- add functionality to get constituency from postcode
- iterator now returns list of BillDetails objects rather than tuple (API incompat change, hence change of major)
- iterator BillDetails object also now contains sessions for the bill
- improve division categorisation "second reading" vs "third reading" vs "amendments"
- make 2019 onwards session tag have consistent value of 2019-21 (session is actually ongoing at time of writing, but Parliament API for divisions uses 2019-21 as the value)
- make scraper session keys match those in utils.dates
- add function to get parties and their corresponding ids
- improve accuracy of associating divisions with bills
- add iterator functionality to get details on bills provided by a BillsOverview object
- adds method to get a list only of bills updated since method was last run, uses pickled datetime so that script can be stopped and run as required
- gets a DataFrame containing bill titles, their last updated times and page paths (page paths to be used in future versions)
- page request delay configurable
- collecting emails using the Parliament API significantly increases the time to run mp_fetcher methods, so by default
only gets emails for current MPs.
- 33 seconds to run
test_get_active_mps
without email data - 272 seconds to run
test_get_active_mps
with email data
- 33 seconds to run