def scrape_bill(self, session, chamber, bill_type, url):
bill_html = self.get(url).text
bill_page = lxml.html.fromstring(bill_html)
qs = dict(urlparse.parse_qsl(urlparse.urlparse(url).query))
bill_id = "{}{}".format(qs["billtype"], qs["billnumber"])
versions = bill_page.xpath(
"//table[contains(@id, 'GridViewVersions')]")[0]
metainf_table = bill_page.xpath(
'//div[contains(@id, "itemPlaceholder")]//table[1]')[0]
action_table = bill_page.xpath(
'//div[contains(@id, "UpdatePanel1")]//table[1]')[0]
meta = self.parse_bill_metainf_table(metainf_table)
subs = [s.strip() for s in meta["Report Title"].split(";")]
if "" in subs:
subs.remove("")
b = Bill(
bill_id,
session,
meta["Measure Title"],
chamber=chamber,
classification=bill_type,
)
if meta["Description"]:
b.add_abstract(meta["Description"], "description")
for subject in subs:
b.add_subject(subject)
if url:
b.add_source(url)
prior_session = "{} Regular Session".format(str(int(session[:4]) - 1))
companion = meta["Companion"].strip()
if companion:
b.add_related_bill(
identifier=companion.replace(u"\xa0", " "),
legislative_session=prior_session,
relation_type="companion",
)
if bill_page.xpath(
"//table[@id='ContentPlaceHolderCol1_GridViewStatus']/tr/td/font/text()"
):
prior = bill_page.xpath(
"//table[@id='ContentPlaceHolderCol1_GridViewStatus']/tr/td/font/text()"
)[-1]
if "carried over" in prior.lower():
b.add_related_bill(
identifier=bill_id.replace(u"\xa0", " "),
legislative_session=prior_session,
relation_type="companion",
)
for sponsor in meta["Introducer(s)"]:
if "(Introduced by request of another party)" in sponsor:
sponsor = sponsor.replace(
" (Introduced by request of another party)", "")
b.add_sponsorship(sponsor, "primary", "person", True)
self.parse_bill_versions_table(b, versions)
self.parse_testimony(b, bill_page)
self.parse_cmte_reports(b, bill_page)
yield from self.parse_bill_actions_table(b, action_table, bill_id,
session, url, chamber)
yield b
def _scrape_bill(self, session, bill_data):
details = self._parse_bill_details(bill_data)
if details is None:
return
(
senate_url,
assembly_url,
bill_chamber,
bill_type,
bill_id,
title,
(prefix, number, active_version),
) = details
bill = Bill(
bill_id,
legislative_session=session,
chamber=bill_chamber,
title=title or bill_data["summary"],
classification=bill_type,
)
if bill_data["summary"]:
bill.add_abstract(bill_data["summary"], note="")
bill_active_version = None
if active_version != "":
bill_active_version = bill_data["amendments"]["items"][active_version]
else:
self.warning("No active version for {}".format(bill_id))
# Parse sponsors.
if bill_data["sponsor"] is not None:
if bill_data["sponsor"]["rules"] is True:
bill.add_sponsorship(
"Rules Committee",
entity_type="organization",
classification="primary",
primary=True,
)
elif not bill_data["sponsor"]["budget"]:
primary_sponsor = bill_data["sponsor"]["member"]
bill.add_sponsorship(
primary_sponsor["shortName"],
entity_type="person",
classification="primary",
primary=True,
)
if bill_active_version:
# There *shouldn't* be cosponsors if there is no sponsor.
cosponsors = bill_active_version["coSponsors"]["items"]
for cosponsor in cosponsors:
bill.add_sponsorship(
cosponsor["shortName"],
entity_type="person",
classification="cosponsor",
primary=False,
)
if bill_active_version:
# List companion bill.
same_as = bill_active_version.get("sameAs", {})
# Check whether "sameAs" property is populated with at least one bill.
if same_as["items"]:
# Get companion bill ID.
companion_bill_id = same_as["items"][0]["basePrintNo"]
# Build companion bill session.
start_year = same_as["items"][0]["session"]
end_year = start_year + 1
companion_bill_session = "-".join([str(start_year), str(end_year)])
# Attach companion bill data.
bill.add_related_bill(
companion_bill_id, companion_bill_session, relation_type="companion"
)
# Parse actions.
chamber_map = {"senate": "upper", "assembly": "lower"}
for action in bill_data["actions"]["items"]:
chamber = chamber_map[action["chamber"].lower()]
action_datetime = datetime.datetime.strptime(action["date"], "%Y-%m-%d")
action_date = action_datetime.date()
types, _ = NYBillScraper.categorizer.categorize(action["text"])
bill.add_action(
action["text"],
action_date.strftime("%Y-%m-%d"),
chamber=chamber,
classification=types,
)
# Handling of sources follows. Sources serving either chamber
# maintain duplicate data, so we can see certain bill data
# through either chamber's resources. However, we have to refer
# to a specific chamber's resources if we want to grab certain
# specific information such as vote data.
#
# As such, I'm placing all potential sources in the interest of
# thoroughness. - Andy Lo
# List Open Legislation API endpoint as a source.
api_url = self.api_client.root + self.api_client.resources["bill"].format(
session_year=session, bill_id=bill_id, summary="", detail=""
)
bill.add_source(api_url)
bill.add_source(senate_url)
bill.add_source(assembly_url)
# Chamber-specific processing.
for vote_data in bill_data["votes"]["items"]:
yield self._parse_senate_votes(vote_data, bill, api_url)
yield from self.scrape_assembly_votes(session, bill, assembly_url, bill_id)
# A little strange the way it works out, but the Assembly
# provides the HTML version documents and the Senate provides
# the PDF version documents.
amendments = bill_data["amendments"]["items"]
for key, amendment in amendments.items():
version = amendment["printNo"]
html_url = (
"http://assembly.state.ny.us/leg/?sh=printbill&bn="
"{}&term={}&Text=Y".format(bill_id, self.term_start_year)
)
bill.add_version_link(
version, html_url, on_duplicate="ignore", media_type="text/html"
)
pdf_url = "http://legislation.nysenate.gov/pdf/bills/{}/{}".format(
self.term_start_year, version
)
bill.add_version_link(
version, pdf_url, on_duplicate="ignore", media_type="application/pdf"
)
yield bill
def scrape_bill(self, session, bill_url):
page = self.get(bill_url).text
page = lxml.html.fromstring(page)
page.make_links_absolute(bill_url)
try:
bill_id = page.xpath('//span[@id="lblBillNumber"]/a[1]')[0].text
except IndexError:
self.logger.warning("Something is wrong with bill page, skipping.")
return
secondary_bill_id = page.xpath('//span[@id="lblCompNumber"]/a[1]')
# checking if there is a matching bill
if secondary_bill_id:
secondary_bill_id = secondary_bill_id[0].text
# swap ids if * is in secondary_bill_id
if "*" in secondary_bill_id:
bill_id, secondary_bill_id = secondary_bill_id, bill_id
secondary_bill_id = secondary_bill_id.strip()
secondary_bill_id = secondary_bill_id.replace(" ", " ")
bill_id = bill_id.replace("*", "").replace(" ", " ").strip()
if "B" in bill_id:
bill_type = "bill"
elif "JR" in bill_id:
bill_type = "joint resolution"
elif "R" in bill_id:
bill_type = "resolution"
primary_chamber = "lower" if "H" in bill_id else "upper"
# secondary_chamber = 'upper' if primary_chamber == 'lower' else 'lower'
title = page.xpath("//span[@id='lblAbstract']")[0].text
if title is None:
msg = "%s detail page was missing title info."
self.logger.warning(msg % bill_id)
return
# bill subject
subject_pos = title.find("-")
subjects = [s.strip() for s in title[:subject_pos - 1].split(",")]
subjects = filter(None, subjects)
bill = Bill(
bill_id,
legislative_session=session,
chamber=primary_chamber,
title=title,
classification=bill_type,
)
for subject in subjects:
bill.add_subject(subject)
if secondary_bill_id:
bill.add_identifier(secondary_bill_id)
if page.xpath('//span[@id="lblCompNumber"]/a'):
companion_id = (page.xpath('//span[@id="lblCompNumber"]/a')
[0].text_content().strip())
bill.add_related_bill(
identifier=companion_id,
legislative_session=session,
relation_type="companion",
)
bill.add_source(bill_url)
# Primary Sponsor
sponsor = (page.xpath("//span[@id='lblBillPrimeSponsor']")
[0].text_content().split("by")[-1])
sponsor = sponsor.replace("*", "").strip()
if sponsor:
bill.add_sponsorship(sponsor,
classification="primary",
entity_type="person",
primary=True)
# bill text
btext = page.xpath("//span[@id='lblBillNumber']/a")[0]
bill.add_version_link("Current Version",
btext.get("href"),
media_type="application/pdf")
# documents
summary = page.xpath('//a[contains(@href, "BillSummaryArchive")]')
if summary:
bill.add_document_link("Summary", summary[0].get("href"))
fiscal = page.xpath('//span[@id="lblFiscalNote"]//a')
if fiscal:
bill.add_document_link("Fiscal Note", fiscal[0].get("href"))
amendments = page.xpath('//a[contains(@href, "/Amend/")]')
for amendment in amendments:
bill.add_version_link(
"Amendment " + amendment.text,
amendment.get("href"),
media_type="application/pdf",
)
# amendment notes in image with alt text describing doc inside <a>
amend_fns = page.xpath('//img[contains(@alt, "Fiscal Memo")]')
for afn in amend_fns:
bill.add_document_link(afn.get("alt"),
afn.getparent().get("href"),
on_duplicate="ignore")
# actions
atable = page.xpath("//table[@id='gvBillActionHistory']")[0]
actions_from_table(bill, atable)
# if there is a matching bill
if secondary_bill_id:
# secondary sponsor
secondary_sponsor = (
page.xpath("//span[@id='lblCompPrimeSponsor']")
[0].text_content().split("by")[-1])
secondary_sponsor = (secondary_sponsor.replace("*", "").replace(
")", "").strip())
# Skip black-name sponsors.
if secondary_sponsor:
bill.add_sponsorship(
secondary_sponsor,
classification="primary",
entity_type="person",
primary=True,
)
# secondary actions
if page.xpath("//table[@id='gvCoActionHistory']"):
cotable = page.xpath("//table[@id='gvCoActionHistory']")[0]
actions_from_table(bill, cotable)
# votes
yield from self.scrape_vote_events(bill, page, bill_url)
bill.actions.sort(key=lambda a: a["date"])
yield bill
def scrape_bills(self, session, year_abr):
# Main Bill information
main_bill_csv = self.to_csv("MAINBILL.TXT")
# keep a dictionary of bills (mapping bill_id to Bill obj)
bill_dict = {}
for rec in main_bill_csv:
bill_type = rec["BillType"].strip()
bill_number = int(rec["BillNumber"])
bill_id = bill_type + str(bill_number)
title = rec["Synopsis"]
if bill_type[0] == "A":
chamber = "lower"
else:
chamber = "upper"
# some bills have a blank title.. just skip it
if not title:
continue
bill = Bill(
bill_id,
title=title,
chamber=chamber,
legislative_session=session,
classification=self._bill_types[bill_type[1:]],
)
if rec["IdenticalBillNumber"].strip():
bill.add_related_bill(
rec["IdenticalBillNumber"].split()[0],
legislative_session=session,
relation_type="companion",
)
# TODO: last session info is in there too
bill_dict[bill_id] = bill
# Sponsors
bill_sponsors_csv = self.to_csv("BILLSPON.TXT")
for rec in bill_sponsors_csv:
bill_type = rec["BillType"].strip()
bill_number = int(rec["BillNumber"])
bill_id = bill_type + str(bill_number)
if bill_id not in bill_dict:
self.warning("unknown bill %s in sponsor database" % bill_id)
continue
bill = bill_dict[bill_id]
name = rec["Sponsor"]
sponsor_type = rec["Type"]
if sponsor_type == "P":
sponsor_type = "primary"
else:
sponsor_type = "cosponsor"
bill.add_sponsorship(
name,
classification=sponsor_type,
entity_type="person",
primary=sponsor_type == "primary",
)
# Documents
bill_document_csv = self.to_csv("BILLWP.TXT")
for rec in bill_document_csv:
bill_type = rec["BillType"].strip()
bill_number = int(rec["BillNumber"])
bill_id = bill_type + str(bill_number)
if bill_id not in bill_dict:
self.warning("unknown bill %s in document database" % bill_id)
continue
bill = bill_dict[bill_id]
document = rec["Document"]
document = document.split("\\")
document = document[-2] + "/" + document[-1]
htm_url = "https://www.njleg.state.nj.us/Bills/{}/{}".format(
year_abr, document.replace(".DOC", ".HTM"))
pdf_url = "https://www.njleg.state.nj.us/Bills/{}/{}".format(
year_abr, document.replace(".DOC", ".PDF"))
# name document based _doctype
try:
doc_name = self._doctypes[rec["DocType"]]
except KeyError:
raise Exception("unknown doctype %s on %s" %
(rec["DocType"], bill_id))
if rec["Comment"]:
doc_name += " " + rec["Comment"]
# Clean links.
if htm_url.endswith("HTMX"):
htm_url = re.sub("X$", "", htm_url)
if pdf_url.endswith("PDFX"):
pdf_url = re.sub("X$", "", pdf_url)
if rec["DocType"] in self._version_types:
if htm_url.lower().endswith("htm"):
mimetype = "text/html"
elif htm_url.lower().endswith("wpd"):
mimetype = "application/vnd.wordperfect"
try:
bill.add_version_link(doc_name,
htm_url,
media_type=mimetype)
bill.add_version_link(doc_name,
pdf_url,
media_type="application/pdf")
except ValueError:
self.warning(
"Couldn't find a document for bill {}".format(bill_id))
pass
else:
bill.add_document_link(doc_name, htm_url)
# Votes
next_year = int(year_abr) + 1
vote_info_list = [
"A%s" % year_abr,
"A%s" % next_year,
"S%s" % year_abr,
"S%s" % next_year,
"CA%s-%s" % (year_abr, next_year),
"CS%s-%s" % (year_abr, next_year),
]
# keep votes clean globally, a few votes show up in multiple files
votes = {}
for filename in vote_info_list:
s_vote_url = f"https://www.njleg.state.nj.us/votes/{filename}.zip"
try:
s_vote_zip, resp = self.urlretrieve(s_vote_url)
except scrapelib.HTTPError:
self.warning("could not find %s" % s_vote_url)
continue
zippedfile = zipfile.ZipFile(s_vote_zip)
for vfile in ["%s.txt" % (filename), "%sEnd.txt" % (filename)]:
try:
vote_file = io.TextIOWrapper(zippedfile.open(vfile, "r"),
encoding="latin-1")
except KeyError:
#
# Right, so, 2011 we have an "End" file with more
# vote data than was in the original dump.
#
self.warning("No such file: %s" % (vfile))
continue
vdict_file = csv.DictReader(vote_file)
if filename.startswith("A") or filename.startswith("CA"):
chamber = "lower"
else:
chamber = "upper"
if filename.startswith("C"):
vote_file_type = "committee"
else:
vote_file_type = "chamber"
for rec in vdict_file:
if vote_file_type == "chamber":
bill_id = rec["Bill"].strip()
leg = rec["Full_Name"]
date = rec["Session_Date"]
action = rec["Action"]
leg_vote = rec["Legislator_Vote"]
vote_parts = (bill_id, chamber, action)
else:
bill_id = "%s%s" % (rec["Bill_Type"],
rec["Bill_Number"])
leg = rec["Name"]
# drop time portion
date = rec["Agenda_Date"].split()[0]
# make motion readable
action = self._com_vote_motions[rec["BillAction"]]
# first char (Y/N) use [0:1] to ignore ''
leg_vote = rec["LegislatorVote"][0:1]
committee = rec["Committee_House"]
vote_parts = (bill_id, chamber, action, committee)
date = datetime.strptime(date, "%m/%d/%Y")
vote_id = "_".join(vote_parts).replace(" ", "_")
if vote_id not in votes:
votes[vote_id] = VoteEvent(
start_date=TIMEZONE.localize(date),
chamber=chamber,
motion_text=action,
classification="passage",
result=None,
bill=bill_dict[bill_id],
)
votes[vote_id].dedupe_key = vote_id
if leg_vote == "Y":
votes[vote_id].vote("yes", leg)
elif leg_vote == "N":
votes[vote_id].vote("no", leg)
else:
votes[vote_id].vote("other", leg)
# remove temp file
os.remove(s_vote_zip)
# Counts yes/no/other votes and saves overall vote
for vote in votes.values():
counts = collections.defaultdict(int)
for count in vote.votes:
counts[count["option"]] += 1
vote.set_count("yes", counts["yes"])
vote.set_count("no", counts["no"])
vote.set_count("other", counts["other"])
# Veto override.
if vote.motion_text == "OVERRIDE":
# Per the NJ leg's glossary, a veto override requires
# 2/3ds of each chamber. 27 in the senate, 54 in the house.
# http://www.njleg.state.nj.us/legislativepub/glossary.asp
if "lower" in vote.bill:
vote.result = "pass" if counts["yes"] >= 54 else "fail"
elif "upper" in vote.bill:
vote.result = "pass" if counts["yes"] >= 27 else "fail"
else:
# Regular vote.
vote.result = "pass" if counts["yes"] > counts[
"no"] else "fail"
vote.add_source("http://www.njleg.state.nj.us/downloads.asp")
yield vote
# Actions
bill_action_csv = self.to_csv("BILLHIST.TXT")
actor_map = {"A": "lower", "G": "executive", "S": "upper"}
for rec in bill_action_csv:
bill_type = rec["BillType"].strip()
bill_number = int(rec["BillNumber"])
bill_id = bill_type + str(bill_number)
if bill_id not in bill_dict:
self.warning("unknown bill %s in action database" % bill_id)
continue
bill = bill_dict[bill_id]
action = rec["Action"]
date = rec["DateAction"]
date = dateutil.parser.parse(date)
actor = actor_map[rec["House"]]
comment = rec["Comment"]
action, atype = self.categorize_action(action, bill_id)
if comment:
action += " " + comment
bill.add_action(
action,
date=TIMEZONE.localize(date),
classification=atype,
chamber=actor,
)
# Subjects
subject_csv = self.to_csv("BILLSUBJ.TXT")
for rec in subject_csv:
bill_id = rec["BillType"].strip() + str(int(rec["BillNumber"]))
if bill_id not in bill_dict:
self.warning("unknown bill %s in subject database" % bill_id)
continue
bill = bill_dict.get(bill_id)
if bill:
bill.subject.append(rec["SubjectKey"])
else:
self.warning("invalid bill id in BillSubj: %s" % bill_id)
phony_bill_count = 0
# save all bills at the end
for bill in bill_dict.values():
# add sources
if not bill.actions and not bill.versions:
self.warning("probable phony bill detected %s",
bill.identifier)
phony_bill_count += 1
else:
bill.add_source("http://www.njleg.state.nj.us/downloads.asp")
yield bill
if phony_bill_count:
self.warning("%s total phony bills detected", phony_bill_count)
def test_full_bill():
create_jurisdiction()
person = Person.objects.create(name="Adam Smith")
lower = Organization.objects.create(jurisdiction_id="jid",
name="House",
classification="lower")
Membership.objects.create(person_id=person.id, organization_id=lower.id)
Organization.objects.create(
jurisdiction_id="jid",
name="Arbitrary Committee",
classification="committee",
parent=lower,
)
oldbill = ScrapeBill(
"HB 99",
"1899",
"Axe & Tack Tax Act",
classification="tax bill",
chamber="lower",
)
bill = ScrapeBill("HB 1",
"1900",
"Axe & Tack Tax Act",
classification="tax bill",
chamber="lower")
bill.subject = ["taxes", "axes"]
bill.add_identifier("SB 9")
bill.add_title("Tack & Axe Tax Act")
bill.add_action("introduced in house", "1900-04-01", chamber="lower")
act = bill.add_action("sent to arbitrary committee",
"1900-04-04",
chamber="lower")
act.add_related_entity(
"arbitrary committee",
"organization",
_make_pseudo_id(name="Arbitrary Committee"),
)
bill.add_related_bill("HB 99",
legislative_session="1899",
relation_type="prior-session")
bill.add_sponsorship(
"Adam Smith",
classification="extra sponsor",
entity_type="person",
primary=False,
entity_id=_make_pseudo_id(name="Adam Smith"),
)
bill.add_sponsorship("Jane Smith",
classification="lead sponsor",
entity_type="person",
primary=True)
bill.add_abstract(
"This is an act about axes and taxes and tacks.",
note="official",
date="1969-10-20",
)
bill.add_document_link("Fiscal Note",
"http://example.com/fn.pdf",
media_type="application/pdf")
bill.add_document_link("Fiscal Note",
"http://example.com/fn.html",
media_type="text/html")
bill.add_version_link("Fiscal Note",
"http://example.com/v/1",
media_type="text/html")
bill.add_source("http://example.com/source")
# import bill
BillImporter("jid").import_data([oldbill.as_dict(), bill.as_dict()])
# get bill from db and assert it imported correctly
b = Bill.objects.get(identifier="HB 1")
assert b.from_organization.classification == "lower"
assert b.identifier == bill.identifier
assert b.title == bill.title
assert b.classification == bill.classification
assert b.subject == ["taxes", "axes"]
assert b.abstracts.get().note == "official"
assert b.abstracts.get().date == "1969-10-20"
# other_title, other_identifier added
assert b.other_titles.get().title == "Tack & Axe Tax Act"
assert b.other_identifiers.get().identifier == "SB 9"
# actions
actions = list(b.actions.all())
assert len(actions) == 2
# ensure order was preserved (if this breaks it'll be intermittent)
assert actions[0].organization == Organization.objects.get(
classification="lower")
assert actions[0].description == "introduced in house"
assert actions[1].description == "sent to arbitrary committee"
assert actions[1].related_entities.get(
).organization == Organization.objects.get(classification="committee")
# action computed fields
assert b.first_action_date == "1900-04-01"
assert b.latest_action_date == "1900-04-04"
assert b.latest_action_description == "sent to arbitrary committee"
# related_bills were added
rb = b.related_bills.get()
assert rb.identifier == "HB 99"
# and bill got resolved
assert rb.related_bill.identifier == "HB 99"
# sponsors added, linked & unlinked
sponsorships = b.sponsorships.all()
assert len(sponsorships) == 2
person = Person.objects.get(name="Adam Smith")
for ss in sponsorships:
if ss.primary:
assert ss.person is None
assert ss.organization is None
else:
assert ss.person == person
# versions & documents with their links
versions = b.versions.all()
assert len(versions) == 1
assert versions[0].links.count() == 1
documents = b.documents.all()
assert len(documents) == 1
assert documents[0].links.count() == 2
# sources
assert b.sources.count() == 1
