def __init__(self):
self.base_url = "http://recherche.liberation.fr/recherche/"
self.constant_args = "&period=custom&editorial_source=&paper_channel=&sort=-publication_date_time"
self.fetcher = Fetcher()
# Computed
self.authors = {} # name associated with a number of articles
self.links = [] # links toward found articles
self.txt = [] # articles texts
# Regexps
self.re_article = '<div class="article">.*?<div class="object-content">(.*?)</div>'
self.re_author = 'Par <strong>(.*?)</strong>'
self.re_link = '<a href="(.*?)">'
#self.re_nnext = '<a href="(.*?)" class="next">'
self.re_nnext = '<a class="next" href="(.*?)">'
self.re_txt = '<div class="articleContent">(.*?)</div>'
class LiberationScraper:
def __init__(self):
self.base_url = "http://recherche.liberation.fr/recherche/"
self.constant_args = "&period=custom&editorial_source=&paper_channel=&sort=-publication_date_time"
self.fetcher = Fetcher()
# Computed
self.authors = {} # name associated with a number of articles
self.links = [] # links toward found articles
self.txt = [] # articles texts
# Regexps
self.re_article = '<div class="article">.*?<div class="object-content">(.*?)</div>'
self.re_author = 'Par <strong>(.*?)</strong>'
self.re_link = '<a href="(.*?)">'
#self.re_nnext = '<a href="(.*?)" class="next">'
self.re_nnext = '<a class="next" href="(.*?)">'
self.re_txt = '<div class="articleContent">(.*?)</div>'
def scrape_period(self, expression, bday, bmonth, byear, eday, emonth, eyear):
"""
Scrapes Liberation for a query among a period of time.
"""
period = "&period_start_day=%d&period_start_month=%d&period_start_year=%d&period_end_day=%d&period_end_month=%d&period_end_year=%d" % (bday, bmonth, byear, eday, emonth, eyear)
query = "%s?q=%s%s%s" % (self.base_url, expression, period, self.constant_args)
next = True
count = 1
while next is True:
print("\t[%d]" % count)
count += 1
# fetching page
page = None
page = self.fetcher.fetch(query)
if page is None:
# Let's try again
time.sleep(3)
page = self.fetcher.fetch(query)
if page is None:
print( "Impossible to fetch [%s]" % query )
break
page = html5wrapper.clean_html( page )
# is there a next link
m = re.search(self.re_nnext, page, re.U)
if m is not None:
url = m.group(1).strip()
query = "%s%s" % (self.base_url, url)
query = query.replace("&amp=", "")
else:
print("\t> Last page reached.")
next = False
links = self._extract_articles(page)
self.links.extend( links )
for l in links:
# Let's now extract text from each article, with some courtoisy
time.sleep( random.uniform(1, 3) )
p = self.fetcher.fetch( l )
if p is not None:
p = html5wrapper.clean_html( p )
m = re.search(self.re_txt, p, re.U|re.M|re.S)
if m is not None:
txt = m.group(1).strip()
txt = "[%s]\n%s" % (l, txt)
self.txt.append( txt )
else:
print("\n\t[%s] impossible to extract article" % l)
#FIXME: FAKE break for testing purpose
#break
# Building final result
res = {}
res["links"] = self.links
res["authors"] = self.authors
res["txt"] = self.txt
return res
def _extract_articles(self, page):
"""
Extracts links toward articles and authors names
from a results page.
"""
links = []
articles = re.findall(self.re_article, page, re.U|re.M|re.S)
for article in articles:
# Looking for author
m = re.search(self.re_author, article, re.U|re.M|re.S)
if m is not None:
author = m.group(1).strip()
try:
self.authors[author] += 1
except:
self.authors[author] = 1
# Looking for article link
m = re.search(self.re_link, article, re.U|re.M|re.S)
if m is not None:
link = m.group(1).strip()
links.append( link )
return links
