def __init__(self, session):
self.session = session
self.current_url = "http://www.pmg.org.za/hansard"
self.current_page = scrapertools.URLFetcher(self.current_url,
self.session).html
self.current_hansard = {}
self.stats = {"total_hansards": 0, "new_hansards": 0, "errors": []}
def run_scraper(self):
committees = Agent.query.filter(Agent.type == "committee").filter(
Agent.url != None).all()
shuffle(
committees) # randomize the order, just to keep things interesting
for i, committee in enumerate(committees):
self.current_committee = committee
self.current_url = committee.url
try:
self.current_page = scrapertools.URLFetcher(
self.current_url, self.session).html
logger.debug("Committee: " + str(committee.name))
self.scrape_committee()
# give some progress feedback
logger.info(
str(i + 1) + " out of " + str(len(committees)) +
" committees' reports have been scraped.")
logger.info(json.dumps(self.stats, indent=4))
# commit entries to database, once per committee
logger.debug("SAVING TO DATABASE")
db.session.commit()
except Exception as e:
msg = "Error scraping committee's reports."
self.stats["errors"].append(msg)
logger.error(msg)
logger.exception(str(e))
return
def run_scraper(self):
"""
Iterate through bill pages, and run the state machine for each page.
"""
pager = Pager()
# iterate through bill pages
for url in pager.next_page:
logger.info(url)
# initiate parser for this page
self.state_fn = self.start_state
html = scrapertools.URLFetcher(url, self.session).html
soup = BeautifulSoup(html,
convertEntities=BeautifulSoup.HTML_ENTITIES)
table = soup.find("tbody")
rows = table.findAll("tr")
# feed rows into state machine
for row in rows:
while not self.state_fn(row):
pass
# commit to database after each page
# db.session.commit()
return
def run_scraper(self):
# define agents
na = Agent.query.filter(Agent.name == "National Assembly").first()
ncop = Agent.query.filter(
Agent.name == "National Council Of Provinces").first()
while True:
for (j, (date, title,
href_hansard)) in enumerate(self.next_hansard):
logger.debug("\t\t" + str(date) + " - " + title)
tmp_url = href_hansard
html = scrapertools.URLFetcher(tmp_url, self.session).html
soup = BeautifulSoup(
html, convertEntities=BeautifulSoup.HTML_ENTITIES)
content = soup.find(id="content")
# find bills that are mentioned in the text
bills = scrapertools.find_bills(str(content))
# only save hansards that are related to bills
if bills:
# infer location from title, where appropriate
location = None
agent = None
if title.startswith("NA:"):
location = 1
agent = na
elif title.startswith("NCOP:"):
location = 2
agent = ncop
self.current_hansard = {
"bills": bills,
"url": tmp_url,
"date": date,
"title": title,
"location": location
}
if agent:
self.current_hansard['agent'] = agent
try:
self.add_or_update()
except Exception:
msg = "Could not add hansard to database: "
if self.current_hansard.get("title"):
msg += self.current_hansard["title"]
self.stats["errors"].append(msg)
logger.error(msg)
self.current_hansard = {}
# save hansards to database, once per scraped page
db.session.commit()
logger.debug(json.dumps(self.stats, indent=4))
# test whether we have reached the end
if not self.next_page:
break
return
def next_page(self):
soup = BeautifulSoup(self.current_page)
next_link = soup.find("li", {"class": "pager-next"})
if next_link:
href = "http://www.pmg.org.za" + next_link.find('a').attrs[0][1]
self.current_url = href
self.current_page = scrapertools.URLFetcher(
self.current_url, self.session).html
return True
return False
def next_page(self):
"""
Extract the 'next' link, if there is one.
"""
soup = BeautifulSoup(self.current_page)
reports_tab = soup.find(id="quicktabs_tabpage_committees_tabs_1")
next_link = reports_tab.find("li", {"class": "pager-next"})
if next_link:
href = "http://www.pmg.org.za" + next_link.find('a').attrs[0][1]
self.current_url = href
self.current_page = scrapertools.URLFetcher(
self.current_url, self.session).html
return True
return False
def next_committee(self):
html = scrapertools.URLFetcher("http://www.pmg.org.za/committees", self.session).html
soup = BeautifulSoup(html, convertEntities=BeautifulSoup.HTML_ENTITIES)
container = soup.find(id="committees-all")
committee_lists = container.findAll("div", {"class": "item-list"})
for committee_list in committee_lists:
list_name = committee_list.find('h3').contents[0]
logger.debug("\n" + list_name + ":")
committees = committee_list.findAll('li')
for committee in committees:
href = "http://www.pmg.org.za" + committee.find('a').attrs[0][1]
name = committee.find('a').contents[0]
logger.debug("\t" + name)
yield list_name, href, name
def scrape_committee(self):
"""
Scrape all meeting reports for a particular committee.
"""
for (j, (date, title, href_report)) in enumerate(self.next_report):
logger.debug("\t\t" + str(date) + " - " +
(title[0:45]) if len(title) > 45 else title)
tmp_url = href_report
html = scrapertools.URLFetcher(tmp_url, self.session).html
soup = BeautifulSoup(html)
content = soup.find(id="content")
bills = scrapertools.find_bills(str(content))
# only save report entries that can be tagged to bills
if bills:
self.current_report = {
"entry_type": "committee-meeting",
"bills": bills,
"url": tmp_url,
"date": date,
"title": title,
"agent": self.current_committee,
}
# report URL may have changed after editing on pmg.org.za, check for this
possible_duplicates = Entry.query.filter(Entry.agent == self.current_committee)\
.filter(Entry.url != None)\
.filter(Entry.url != tmp_url)\
.filter(Entry.type == "committee-meeting")\
.filter(Entry.is_deleted == False)\
.filter(Entry.date == date)\
.order_by(Entry.entry_id).all()
deletion_flag = False
if possible_duplicates:
logger.debug(
str(len(possible_duplicates)) +
" possible duplicates found")
for possible_duplicate in possible_duplicates:
redirect_url = scrapertools.URLFetcher(
possible_duplicate.url,
self.session).follow_redirect()
if possible_duplicate.url != redirect_url:
logger.debug('redirect encountered')
if redirect_url == tmp_url:
logger.info("Updating entry URL")
# update the existing record's URL
possible_duplicate.url = tmp_url
# # delete all but one entry, if there are multiple duplicates
# if deletion_flag:
# logger.info('duplicate entry deleted')
# possible_duplicate.is_deleted = True
db.session.add(possible_duplicate)
db.session.commit()
deletion_flag = True
if self.current_committee.location:
self.current_report[
"location"] = self.current_committee.location
try:
self.add_or_update()
except Exception, e:
msg = "Could not add committee report to database: "
if self.current_report.get("title"):
msg += self.current_report["title"]
self.stats["errors"].append(msg)
logger.error(msg)
logger.exception(str(e))
self.current_report = {}
else:
logger.debug('no bills found in committee meeting report')