def _scrape_unit(self, article_id):
article = HTMLDocument(url = self.article_url.format(**locals()))
article.prepare(self)
article.props.text = article.doc.cssselect("font.artbody")
if len("".join([t.text_content() for t in article.props.text])) < 100:
return
for i, table in enumerate(article.doc.cssselect("table")):
if table.get('class') == "body":
table_after_body = article.doc.cssselect("table")[i + 1]
page_date = re.search(
"Pagina ([0-9]+), ([0-9]{2}\-[0-9]{2}\-[0-9]{4})",
table_after_body.text_content())
article.props.pagenr = page_date.group(1)
article.props.date = readDate(page_date.group(2))
article.props.section = self.current_section
article.props.headline = article.doc.cssselect("td.artheader")[0].text_content().strip()
if article.doc.cssselect(".artsubheader"):
article.props.byline = article.doc.cssselect(".artsubheader")[0]
if article.doc.cssselect("td.artauthor"):
article.props.author = article.doc.cssselect("td.artauthor")[0].text.split(":")[1].strip()
dateline_match = re.search(
"^([A-Z][a-z]+(( |/)[A-Z][a-z]+)?)\n",
"\n".join([n.text_content() for n in article.props.text]).strip())
if dateline_match:
article.props.dateline = dateline_match.group(1)
yield article
def _get_units(self):
for section in self.sections:
page = 1
url = self.page_url.format(**locals())
date = _date.today()
ipage = self.getdoc(url)
while date >= self.options['date']:
if not ipage.cssselect("#main ul.snelnieuws_list li.item"):
print("\nNo articles found as far back as given date\n")
break
for unit in ipage.cssselect(
'#main ul.snelnieuws_list li.item'):
href = unit.cssselect('a')[0].get('href')
article = HTMLDocument(url=href)
article.prepare(self)
try:
date = readDate(
article.doc.cssselect("span.datum")
[0].text).date()
except IndexError:
continue
if date == self.options['date']:
yield article
elif date < self.options['date']:
break
page += 1
nxt_url = self.page_url.format(**locals())
ipage = self.getdoc(nxt_url)
def _get_units(self):
for section in self.sections:
page = 1
url = self.page_url.format(**locals())
date = _date.today()
ipage = self.getdoc(url)
while date >= self.options['date']:
if not ipage.cssselect("#main ul.snelnieuws_list li.item"):
print("\nNo articles found as far back as given date\n")
break
for unit in ipage.cssselect('#main ul.snelnieuws_list li.item'):
href = unit.cssselect('a')[0].get('href')
article = HTMLDocument(url=href)
article.prepare(self)
try:
date = readDate(article.doc.cssselect("span.datum")[0].text).date()
except IndexError:
continue
if date == self.options['date']:
yield article
elif date < self.options['date']:
break
page += 1
nxt_url = self.page_url.format(**locals())
ipage = self.getdoc(nxt_url)
def _scrape_unit(self, li):
a = li.cssselect("li > a")[0]
article = HTMLDocument(url=urljoin(self.index_url, a.get('href')))
article.props.headline = a.text
article.props.kicker = li.cssselect("div.infoboard a.kicker")[0].text
article.props.intro = li.cssselect("p")
article.props.date = readDate(
li.cssselect("div.infoboard span.time")[0].text_content())
article.prepare(self)
articletime = article.doc.cssselect("p.articletime")[0].text_content()
if len(articletime.split("|")) > 2:
article.props.date = readDate(" ".join(
articletime.split("|")[:-1]))
article.props.author = articletime.split("|")[-1]
else:
article.props.author = articletime.strip()
if " Korrespondent" in article.props.author:
article.props.author = article.props.author.split(
"Korrespondent")[1].strip()
for ad in article.doc.cssselect("div.noprint"):
ad.drop_tree()
article.props.text = article.doc.cssselect(
"p.articlelead, #articletext")
article.props.section = article.doc.cssselect(
"div.headtop span.sitetop")[0].text_content()
yield article
def get_comments(self, page):
for li in page.doc.cssselect("ul.commentlist li.comment"):
comment = HTMLDocument()
comment.parent = page
try:
dateauthor = li.cssselect("div.commentsbox")[0].text_content()
except IndexError:
comment.props.author = li.text_content().split(":")[0]
comment.props.date = readDate(":".join(li.text_content().split(":")[1:2]))
try:
comment.props.text = li.cssselect("div.comment-text-reply")[0]
except UnicodeDecodeError:
continue
else:
comment.props.author = dateauthor.split("Geplaatst door")[1].split(" op ")[0]
try:
li.cssselect("div.commentsbox a")[0].drop_tree()
except:
pass
comment.props.date = readDate(dateauthor.split(" op ")[1])
try:
comment.props.text = li.cssselect("div.comment-text")[0]
except UnicodeDecodeError:
continue
yield comment
def _scrape_unit(self, url):
article = HTMLDocument(url=url)
article.prepare(self)
content = article.doc.cssselect("#content-column")[0]
article.props.date = readDate(content.cssselect("p.article-meta")[0].text.split("|")[1])
article.props.headline = content.cssselect("h1")[0].text
for x in [
content.cssselect("h1")[0],
content.cssselect("p.article-meta")[0],
content.cssselect("p.sharing")[0]
]:
x.drop_tree()
article.props.text = content.text_content()
for block in article.doc.cssselect("#aside-column div.block"):
title = block.cssselect("h2")[0].text
if "Verantwoordelijk" in title and "ministerie" in title:
article.props.author = "; ".join([a.text for a in block.cssselect("ul.list-common li a")])
break
try:
if len(article.props.author) > 100:
article.props.author = article.props.author[:100]
except AttributeError:
pass
yield article
def _scrape_unit(self, url):
page = HTMLDocument(url=url)
page.prepare(self)
for comment in self.get_comments(page):
yield comment
comment.is_comment = True
yield self.get_article(page)
def _scrape_unit(self, article_id):
article = HTMLDocument(url=self.article_url.format(**locals()))
article.prepare(self)
article.props.text = article.doc.cssselect("font.artbody")
if len("".join([t.text_content() for t in article.props.text])) < 100:
return
for i, table in enumerate(article.doc.cssselect("table")):
if table.get('class') == "body":
table_after_body = article.doc.cssselect("table")[i + 1]
page_date = re.search(
"Pagina ([0-9]+), ([0-9]{2}\-[0-9]{2}\-[0-9]{4})",
table_after_body.text_content())
article.props.pagenr = page_date.group(1)
article.props.date = readDate(page_date.group(2))
article.props.section = self.current_section
article.props.headline = article.doc.cssselect(
"td.artheader")[0].text_content().strip()
if article.doc.cssselect(".artsubheader"):
article.props.byline = article.doc.cssselect(".artsubheader")[0]
if article.doc.cssselect("td.artauthor"):
article.props.author = article.doc.cssselect(
"td.artauthor")[0].text.split(":")[1].strip()
dateline_match = re.search(
"^([A-Z][a-z]+(( |/)[A-Z][a-z]+)?)\n",
"\n".join([n.text_content() for n in article.props.text]).strip())
if dateline_match:
article.props.dateline = dateline_match.group(1)
yield article
def _scrape_unit(self, url):
page = HTMLDocument(url=url)
page.prepare(self)
for comment in self.get_comments(page):
yield comment
comment.is_comment = True
yield self.get_article(page)
def _scrape_unit(self, pagenr_url):
pagenumber, url = pagenr_url
page = HTMLDocument(date=self.options.get("date"), url=url, pagenumber=pagenumber)
page.prepare(self)
article = self._get_article(page)
if article:
yield article
def get_comments(self, page):
for li in page.doc.cssselect("ul.commentlist li.comment"):
comment = HTMLDocument()
comment.parent = page
try:
dateauthor = li.cssselect("div.commentsbox")[0].text_content()
except IndexError:
comment.props.author = li.text_content().split(":")[0]
comment.props.date = readDate(":".join(
li.text_content().split(":")[1:2]))
try:
comment.props.text = li.cssselect(
"div.comment-text-reply")[0]
except UnicodeDecodeError:
continue
else:
comment.props.author = dateauthor.split(
"Geplaatst door")[1].split(" op ")[0]
try:
li.cssselect("div.commentsbox a")[0].drop_tree()
except:
pass
comment.props.date = readDate(dateauthor.split(" op ")[1])
try:
comment.props.text = li.cssselect("div.comment-text")[0]
except UnicodeDecodeError:
continue
yield comment
def scrape_media(self,doc,_type):
scrn = HTMLDocument()
scrn.doc = doc
try:
scrn.props.text = scrn.doc.cssselect("div.mediaDescription")[0]
except IndexError:
scrn.props.text = "none"
try:
scrn.props.headline = "{} {}".format(scrn.doc.cssselect("div.screenshotAppName")[0].text,_type)
except IndexError:
scrn.props.headline = "unknown"
author_url = "/".join(scrn.doc.cssselect("div.linkAuthor a")[0].get('href').split("/")[:-2])
scrn = self.get_author_props(scrn, author_url)
for obj in scrn.doc.cssselect("div.rightDetailsBlock div.detailsStatRight"):
try:
scrn.props.date = readDate(obj.text)
except ValueError:
continue
else:
break
if not scrn.doc.cssselect("div.commentthread_paging"):
yield scrn;return
if not scrn.doc.cssselect("div.commentthread_header div.commentthread_paging span")[1].text_content():
for comment in self.scrape_comments(scrn):
yield comment
else:
raise NotImplementedError
yield scrn
def get_comment(self, page, header,table):
comment = HTMLDocument()
comment.parent = page
comment.props.date = readDate(header.cssselect("span.kmsgdate")[0].get('title'))
comment.props.headline = header.cssselect("h2 span")[0].text_content()
comment.props.author = table.cssselect("li.kpost-username")[0].text_content()
comment.props.text = table.cssselect("div.kmsgtext")[0]
return comment
def get_comment(self, page, header, table):
comment = HTMLDocument()
comment.parent = page
comment.props.date = readDate(header.cssselect("span.kmsgdate")[0].get("title"))
comment.props.headline = header.cssselect("h2 span")[0].text_content()
comment.props.author = table.cssselect("li.kpost-username")[0].text_content()
comment.props.text = table.cssselect("div.kmsgtext")[0]
return comment
def get_comments(self,page):
for div in page.doc.cssselect("div.comment"):
comment = HTMLDocument()
comment.props.text = div.cssselect("div.reactie")[0]
comment.props.author = div.cssselect("li.naam")[0].text_content()
comment.props.date = readDate(div.cssselect("li.date")[0].text_content())
comment.parent = page
yield comment
def get_comments(self, page):
for li in page.doc.cssselect("#detail_reactions #reaction ul.clear li"):
comment = HTMLDocument()
comment.props.author = li.cssselect("cite")[0].text.strip()
comment.props.text = li.cssselect("blockquote")[0]
comment.props.date = readDate(li.cssselect("span.time")[0].text)
comment.parent = page
yield comment
def get_comments(self, page):
for li in page.doc.cssselect("#detail_reactions #reaction ul.clear li"):
comment = HTMLDocument()
comment.props.author = li.cssselect("cite")[0].text.strip()
comment.props.text = li.cssselect("blockquote")[0]
comment.props.date = readDate(li.cssselect("span.time")[0].text)
comment.parent = page
yield comment
def _get_units(self):
for url, doc in self.getdocs():
date = readDate(doc.cssselect("section.headlinedivider p.lfloat")[0].text_content().strip().split("am")[1])
print(date)
if date.date() != self.options['date']:
continue
article = HTMLDocument(url = url, date = date)
article.doc = doc
yield article
def get_comments(self, page):
for div in page.doc.cssselect("div.comment"):
comment = HTMLDocument()
comment.props.text = div.cssselect("div.reactie")[0]
comment.props.author = div.cssselect("li.naam")[0].text_content()
comment.props.date = readDate(
div.cssselect("li.date")[0].text_content())
comment.parent = page
yield comment
def _scrape_unit(self, url):
article = HTMLDocument(url = url)
article.prepare(self)
article.props.date = self.options['date']
article.props.section = " > ".join([
a.text for a in article.doc.cssselect("#SubHeaderBreadcrumbs a")])
article.props.headline = article.doc.cssselect("title")[0].text.split("—")[0].strip()
article.props.text = article.doc.cssselect(
"#parent-fieldname-description") + article.doc.cssselect("#parent-fieldname-text")
yield article
def _scrape_unit(self, ipage):
page = ipage
ipage = HTMLDocument(ipage)
ipage.doc = self.open(page['url'])
text = wegenertools.clean(ipage.doc.read())
err_text = "Uw account is niet geregistreerd voor de door u gekozen uitgave."
if err_text in text:
raise Exception(err_text)
for article_ids in wegenertools.get_article_ids(text):
body, headline, byline = wegenertools.get_article(
text, article_ids)
if len(
body
) >= 300: #filtering non-articles, image links and other html crap
artpage = HTMLDocument()
stop = False
for part in body.split("\n\n"):
if part.isupper():
pass
else:
if "\n" in part:
#when title has a linebreak it's probably not an article
stop = True
break
else:
artpage.props.headline = part
break
if stop:
break
else:
p = re.compile("[\\\]udc[\w\w]")
artpage.props.text = literal_eval(p.sub("", repr(body)))
artpage.props.edition = page['edition']
artpage.props.byline = byline
artpage.props.section = page['section']
if re.match("[A-Z][0-9]+", page['page_str']):
artpage.props.section += " - section " + page[
'page_str'][0]
artpage.props.pagenr = int(page['page_str'][1:])
else:
artpage.props.pagenr = int(page['page_str'])
dateline_pattern = re.compile(
"(^[^\n]+\n\n([A-Z]+( [A-Z]+)?) -\n)|(([A-Z]+( [A-Z]+)?)\n\n)"
)
match = dateline_pattern.search(artpage.props.text)
if match:
#dateline and theme have the same syntax and are therefore undistinguishable
artpage.props.dateline_or_theme = match.group(
2) or match.group(5)
artpage.props.url = page['url']
yield artpage
def _scrape_unit(self, url):
article = HTMLDocument(url = url)
article.prepare(self.scraper)
article.props.headline = article.doc.cssselect("#articleDetailTitle")[0].text_content()
time_post = article.doc.cssselect("div.time_post")[0]
if time_post.cssselect("span.author"):
article.props.author = time_post.cssselect("span.author")[0].text_content().lstrip("Dor:")
time_post.cssselect("span.author")[0].drop_tree()
article.props.date = readDate(time_post.text_content())
article.props.text = article.doc.cssselect("#art_box2")[0]
yield article
def _scrape_unit(self, url):
article = HTMLDocument(url = url)
article.prepare(self.scraper)
article.props.date = readDate(article.doc.cssselect("div.dateplace-data")[0].text)
article.props.headline = article.doc.cssselect("h1")[0].text_content().strip()
[s.drop_tree() for s in article.doc.cssselect("script")]
article.props.text = article.doc.cssselect("#leadarticle div.content")[0]
author = article.doc.cssselect("#leadarticle span.smallprint")
if author:
article.props.author = author[0].text.strip("| ")
yield article
def _scrape_unit(self, url):
article = HTMLDocument(url = url)
article.prepare(self.scraper)
article.props.date = readDate(article.doc.cssselect("#artikel span.datum,#artikel span.datum-premium-content")[0].text_content())
article.props.headline = article.doc.cssselect("#artikel h1")[0].text_content()
author = article.doc.cssselect("#artikel span.auteur")
if author:
article.props.author = author[0].text_content()
[s.drop_tree() for s in article.doc.cssselect("#artikelKolom script")]
article.props.text = article.doc.cssselect("#artikelKolom,#artikel div.zak-txt-premium-content")[0]
yield article
def _scrape_unit(self, url):
doc = self.getdoc(url)
for a in doc.cssselect('#Articles a'):
page = HTMLDocument(date=self.options.get('date'))
page.coords = stoolkit.parse_coords(doc.cssselect('div.%s' % a.get('class')))
page.props.url = urljoin(url, '%s_text.html' % a.get('class'))
page.prepare(self)
article = self.get_article(page)
if article:
yield article
def get_article(self, url, datetime):
page = HTMLDocument(url = url)
page.prepare(self)
page.props.headline = page.doc.cssselect("div.article h1")[0]
page.props.text = page.doc.cssselect("#broodtekst")[0]
page.props.date = datetime
if page.doc.cssselect("div.auteursinfo"):
page.props.author = page.doc.cssselect("div.auteursinfo h2")[0].text_content()
page.props.section = url.split("/")[3]
page.props.html = html.tostring(page.doc)
return page
def _scrape_unit(self, url):
article = HTMLDocument(url = url)
article.prepare(self)
article.props.date = readDate(article.doc.cssselect("#artikel span.datum")[0].text_content())
article.props.headline = article.doc.cssselect("#artikel h1")[0].text_content()
author = article.doc.cssselect("#artikel span.auteur")
if author:
article.props.author = author[0].text_content()
[s.drop_tree() for s in article.doc.cssselect("#artikelKolom script")]
article.props.text = article.doc.cssselect("#artikelKolom")[0]
yield article
def _scrape_unit(self, data):
headline, article_date, pagenr, section, url = data
art = HTMLDocument(
headline = headline, date = article_date,
pagenr = pagenr, section = section, url = url)
art.doc = self.open(url).read()
text = self.pdf_to_text(art.doc).decode('utf-8')
art.props.text = self.fix_text(text)
art.props.source = "dekrantvantoen.nl"
yield art
def _scrape_unit(self, url):
article = HTMLDocument(url = url)
article.prepare(self)
article.props.headline = article.doc.cssselect("#articleDetailTitle")[0].text_content()
time_post = article.doc.cssselect("div.time_post")[0]
if time_post.cssselect("span.author"):
article.props.author = time_post.cssselect("span.author")[0].text_content().lstrip("Dor:")
time_post.cssselect("span.author")[0].drop_tree()
article.props.date = readDate(time_post.text_content())
article.props.text = article.doc.cssselect("#art_box2")[0]
yield article
def _scrape_unit(self, url):
article = HTMLDocument(url = url)
article.prepare(self)
article.props.date = readDate(article.doc.cssselect("div.dateplace-data")[0].text)
article.props.headline = article.doc.cssselect("h1")[0].text_content().strip()
[s.drop_tree() for s in article.doc.cssselect("script")]
article.props.text = article.doc.cssselect("#leadarticle div.content")[0]
author = article.doc.cssselect("#leadarticle span.smallprint")
if author:
article.props.author = author[0].text.strip("| ")
yield article
def _scrape_unit(self, url):
page = HTMLDocument(url = url)
page.prepare(self)
page.props.headline = page.doc.cssselect("#art_box2 h1")[0].text_content()
for h1 in page.doc.cssselect("h1"):
h1.drop_tree()
page.props.author = self.getauthor(page.doc)
page.props.text = page.doc.cssselect("#art_box2 p")
page.props.date = readDate(page.doc.cssselect("div.time_post")[0].text.split("Bron:")[0])
page.props.section = re.search("parool/nl/[0-9]+/([\w\d\-]+)/article", page.props.url).group(1).capitalize()
page.props.html = html.tostring(page.doc)
yield page
def get_article(self, url):
url = "{}_body.html".format(url[:-5])
pagenum = url.split("/")[7][0:5]
article = HTMLDocument(url = url, pagenr = int(pagenum))
article.doc = self.getdoc(url)
article.props.headline = article.doc.cssselect("td.artheader")[0].text_content().strip()
article.props.text = article.doc.cssselect("table.body")[0]
if article.doc.cssselect("td.artauthor"):
article.props.author = article.doc.cssselect("td.artauthor")[0].text_content().lstrip("dor")
article.props.date = self.options['date']
article.props.section = self.section
return article
def get_article(self, url, datetime):
page = HTMLDocument(url=url)
page.prepare(self)
page.props.headline = page.doc.cssselect("div.article h1")[0]
page.props.text = page.doc.cssselect("#broodtekst")[0]
page.props.date = datetime
if page.doc.cssselect("div.auteursinfo"):
page.props.author = page.doc.cssselect(
"div.auteursinfo h2")[0].text_content()
page.props.section = url.split("/")[3]
return page
def _scrape_unit(self, tr):
"""gets articles from a page"""
url = urljoin("http://forum.fok.nl",tr.cssselect("td.tTitel a")[0].get('href'))
topic = HTMLDocument(url=url,
section=tr.cssselect("td.tFolder")[0].text_content())
topic.prepare(self)
content = topic.doc.text_content()
if any([(s in content) for s in SEARCHTERMS]):
for comment in self.get_comments(topic):
comment.is_comment = True
yield comment
yield self.get_article(topic)
def _scrape_unit(self, data):
headline, article_date, pagenr, section, url = data
art = HTMLDocument(headline=headline,
date=article_date,
pagenr=pagenr,
section=section,
url=url)
art.doc = self.open(url).read()
text = self.pdf_to_text(art.doc).decode('utf-8')
art.props.text = self.fix_text(text)
art.props.source = "dekrantvantoen.nl"
yield art
def _scrape_unit(self, url):
article = HTMLDocument(url = url)
article.prepare(self)
firstitem = article.doc.cssselect("#content div.left_div div.item")[0]
article.props.text = firstitem.cssselect("p, div")
article.props.date = readDate(firstitem.cssselect("span.date")[0].text)
article.props.section = "nieuws"
article.props.headline = firstitem.cssselect("h3")[0].text
article.props.externalid = url.split("/")[-1]
em = firstitem.cssselect("em")
if em:
article.props.author = "".join(em[0].text.split("Door ")[1:])
yield article
def get_comments(self, article):
for li in article.doc.cssselect("li.comment"):
comment = HTMLDocument()
comment.props.text = li.cssselect("div.comment-text")[0]
pattern = re.compile("Geplaatst door ([\w ]+) op ([\w :]+)")
result = pattern.search(
li.cssselect("div.commentsbox span")[0].text_content()
)
comment.props.author = result.group(1)
comment.props.date = readDate(result.group(2))
comment.parent = article
yield comment
def _scrape_unit(self, urldoc):
article = HTMLDocument(url = urldoc[0])
article.doc = urldoc[1]
_date = [
int(urldoc[0].split("/")[6]),
int(urldoc[0].split("/")[7].split("_")[0]),
int(urldoc[0].split("/")[8])]
article.props.date = date(*_date)
article.props.section = urldoc[0].split("/")[9]
article.props.author = article.doc.cssselect("div.fullarticle_tagline")[0].text.split("|")[0]
article.props.headline = article.doc.cssselect("h1.title")[0].text
article.props.text = article.doc.cssselect("article")[0]
yield article
def _scrape_unit(self, url):
doc = self.getdoc(url)
for li in doc.cssselect("div#article ul.news-list li"):
url = li.cssselect("a")[0].get('href')
url = urljoin(INDEX_URL, url)
page = HTMLDocument(date=self.options['date'], url=url)
page.prepare(self)
page.doc = self.getdoc(page.props.url)
try:
yield self.get_article(page)
except IndexError:
pass
def scrape_comments(self,page):
p = page.props.url+"?page={}"
if not page.doc.cssselect("ul.pager"):
return
total = int(page.doc.cssselect("ul.pager li.pager-last a")[0].get('href').split("page=")[-1].split("&")[0]) + 1
docs = [self.getdoc(p.format(x)) for x in range(total)]
for doc in docs:
for div in doc.cssselect("#comments div.comment"):
comment = HTMLDocument()
comment.props.text = div.cssselect("div.content")[0]
comment.props.author = div.cssselect("span.submitted-username")[0].text_content()
comment.props.date = readDate(div.cssselect("div.submitted div.floatr")[0].text_content())
comment.parent = page
yield comment
def _scrape_unit(self, bits):
date, url = bits
article = HTMLDocument(date = date, url = url)
article.prepare(self)
content = article.doc.cssselect("#content")[0]
article.props.section = content.cssselect("div.info-block p.meta a.label")[0].text
article.props.headline = content.cssselect("div.title h1")[0].text
article.props.externalid = url.split("-")[-1].strip("W/")
article.props.text = content.cssselect("div.article")
article.props.author = content.cssselect("p.meta span.user a.label")[0].text.strip()
article.props.tags = set([a.text for a in content.cssselect("ul.taglist li a")])
article.props.view_count = int(content.cssselect("div.info-block span.view-count")[0].text)
yield article
self.clearcookies()
def get_article(self, url):
url = "{}_body.html".format(url[:-5])
pagenum = url.split("/")[7][0:5]
article = HTMLDocument(url=url, pagenr=int(pagenum))
article.doc = self.getdoc(url)
article.props.headline = article.doc.cssselect(
"td.artheader")[0].text_content().strip()
article.props.text = article.doc.cssselect("table.body")[0]
if article.doc.cssselect("td.artauthor"):
article.props.author = article.doc.cssselect(
"td.artauthor")[0].text_content().lstrip("dor")
article.props.date = self.options['date']
article.props.section = self.section
return article
def _scrape_unit(self, props):
props["date"] = self.options["date"]
if props["type"] == "short":
yield Document(**props)
elif props["type"] == "full":
article = HTMLDocument(**props)
article.prepare(self)
article.props.section = " > ".join(
[li.text_content().strip("|") for li in article.doc.cssselect("#a-breadcrumb-component li.a-nav-item")]
)
article.props.headline = article.doc.cssselect("div.column_main_wrapper h1.blue")[0].text
article.props.text = article.doc.cssselect("div.column_main_wrapper div")[0]
yield article
def _scrape_unit(self, url):
doc = self.getdoc(url)
for li in doc.cssselect("div#article ul.news-list li"):
url = li.cssselect("a")[0].get('href')
url = urljoin(INDEX_URL,url)
page = HTMLDocument(date = self.options['date'],url=url)
page.prepare(self)
page.doc = self.getdoc(page.props.url)
try:
yield self.get_article(page)
except IndexError:
pass
def get_comments(self, topic):
first = True
for page in self.get_pages(topic.doc):
if first == True:
comments = page.cssselect("div.post")[1:]
first = False
else:
comments = page.cssselect("div.post")
for div in comments:
comment = HTMLDocument()
comment.parent = topic
comment.props.author = div.cssselect("div.postholder_top a.username")[0]
comment.props.date = readDate(div.cssselect("div.postholder_top span.post_time")[0].text_content())
comment.props.text = div.cssselect("div.postholder_bot div.contents")[0]
yield comment
def _scrape_unit(self, xmlitem):
html_content = html.fromstring(xmlitem.cssselect("description")[0].text)
url = xmlitem.cssselect("link")[0].tail.split("&url=")[-1]
article = HTMLDocument(url = url)
article.props.headline = " - ".join(xmlitem.cssselect("title")[0].text.split(" - ")[:-1])
article.props.source = xmlitem.cssselect("title")[0].text.split(" - ")[-1]
article.props.section = xmlitem.cssselect("category")[0].text
article.props.date = readDate(xmlitem.cssselect("pubdate")[0].text)
article.props.snippet = html_content.cssselect("div.lh font")[1].text
try:
article.prepare(self)
except Exception:
yield article;return
article.props.html = html.tostring(article.doc)
yield article
def _scrape_unit(self, urldoc):
url, doc = urldoc
article = HTMLDocument(url = url)
article.doc = doc
article.props.images = [self.open(img.get('src')).read() for img in article.doc.cssselect("div.broodMediaBox div.image img")]
article.props.headline = "Poll:" + doc.cssselect("#artikel h1")[0].text_content().split(":")[1]
article.props.byline = doc.cssselect("#artikel span.auteur")[0].text
article.props.date = readDate(doc.cssselect("#artikel span.datum")[0].text)
article.props.externalid = article.props.url.split("/")[-2]
article.props.text = doc.cssselect("#artikelKolom div.zaktxt,p")
article.props.dateline = doc.cssselect("#artikelKolom span.location")[0]
for comment in self.get_comments(article):
comment.is_comment = True
yield comment
yield article
def _scrape_unit(self, ipage):
page = ipage
ipage = HTMLDocument(ipage)
ipage.doc = self.open(page['url'])
text = wegenertools.clean(ipage.doc.read())
err_text = "Uw account is niet geregistreerd voor de door u gekozen uitgave."
if err_text in text:
raise Exception(err_text)
for article_ids in wegenertools.get_article_ids(text):
body,headline,byline = wegenertools.get_article(text,article_ids)
if len(body) >= 300: #filtering non-articles, image links and other html crap
artpage = HTMLDocument()
stop = False
for part in body.split("\n\n"):
if part.isupper():
pass
else:
if "\n" in part:
#when title has a linebreak it's probably not an article
stop=True
break
else:
artpage.props.headline = part
break
if stop:
break
else:
p = re.compile("[\\\]udc[\w\w]")
artpage.props.text = literal_eval(p.sub("",repr(body)))
artpage.props.edition = page['edition']
artpage.props.byline = byline
artpage.props.section = page['section']
if re.match("[A-Z][0-9]+", page['page_str']):
artpage.props.section += " - section " + page['page_str'][0]
artpage.props.pagenr = int(page['page_str'][1:])
else:
artpage.props.pagenr = int(page['page_str'])
dateline_pattern = re.compile("(^[^\n]+\n\n([A-Z]+( [A-Z]+)?) -\n)|(([A-Z]+( [A-Z]+)?)\n\n)")
match = dateline_pattern.search(artpage.props.text)
if match:
#dateline and theme have the same syntax and are therefore undistinguishable
artpage.props.dateline_or_theme = match.group(2) or match.group(5)
artpage.props.url = page['url']
yield artpage
def scrape_file(self, _html, t):
if "werkmap" in t:
divs = _html.cssselect("#articleTable div")
elif "intranet/rss" in t:
divs = [
div for div in _html.cssselect("#sort div")
if "sort_" in div.get('id')
]
for div in divs:
article = HTMLDocument()
article.props.html = div
article.props.headline = div.cssselect(
"#articleTitle")[0].text_content()
article.props.text = div.cssselect("#articleIntro")[0]
articlepage = div.cssselect("#articlePage")
if articlepage:
article.props.pagenr, article.props.section = self.get_pagenum(
articlepage[0].text)
if not div.cssselect("#sourceTitle")[0].text:
article.props.medium = Medium.get_or_create("unknown medium")
else:
article.props.medium = Medium.get_or_create(
div.cssselect("#sourceTitle")[0].text)
date_str = div.cssselect("#articleDate")[0].text
try:
article.props.date = readDate(date_str)
except ValueError:
log.error(
"parsing date \"{date_str}\" failed".format(**locals()))
else:
yield article
def _get_units(self):
for x in range(3):
try:
self._cookie()
except:
print(
'Error 503 at _cookie function, trying again in a minute...'
)
time.sleep(60)
else:
break
index_dict = {
'y': self.options['date'].year,
'm': self.options['date'].month,
'd': self.options['date'].day
}
url = INDEX_URL.format(**index_dict)
for x in range(3):
try:
index = self.getdoc(url)
except Exception:
time.sleep(5)
else:
break
articles = index.cssselect('.title')
for article_unit in articles:
href = article_unit.cssselect('a')[0].get('href')
yield HTMLDocument(url=href, date=self.options['date'])
def _get_units(self):
"""
PhpBB forum scraper
"""
index = self.getdoc(self.index_url)
for cat_title, cat_doc in self.get_categories(index):
for page in self.get_pages(cat_doc):
for fbg in page.cssselect('.forumbg'):
for li in fbg.cssselect('.topics > li'):
url = urljoin(
self.index_url,
li.cssselect("a.topictitle")[0].get('href'))
_date = etree.tostring(
li.cssselect("dd.lastpost")[0]).split("br />")[1]
date = toolkit.readDate(_date)
yield {
'date':
date,
'object':
HTMLDocument(
headline=li.cssselect("a.topictitle")[0].text,
url=url,
category=cat_title)
}
def _scrape_unit(self, url):
"""gets articles from a page"""
_json = self.open(str(url)).read()
data = json.loads(_json)
done = False
while data['has_more_items'] and done == False:
doc = html.fromstring(data['items_html'])
for div in doc.cssselect("div.tweet"):
tweet = HTMLDocument()
tweet.props.author = div.cssselect(
"strong.fullname")[0].text_content()
tweet.props.date = datetime.fromtimestamp(
float(
div.cssselect("a.tweet-timestamp ._timestamp")[0].get(
'data-time')))
tweet.props.text = div.cssselect("p.js-tweet-text")[0]
maxid = div.get('data-tweet-id')
if tweet.props.date.date() < self.options['date']:
done = True
break
elif tweet.props.date.date() == self.options['date']:
yield tweet
if done == False:
nexturl = url + "&max_id={}".format(maxid)
data = json.loads(self.open(str(nexturl)).read())
def _get_units(self):
back = (date.today() - self.options['date']).days
index_text = self.open(
self.index_url.format(**locals())).read().decode('utf-8')
#A character in the header makes the html library fail to parse the page correctly (it silently returns half the page without warning -.-)
#The character is located in the class attribute of each article tag that is to be scraped, so we take the article tag's inner wrapper and parse that instead.
article = 0
arts = []
for part in index_text.split("<article"):
article = part.split("</article>")[0]
arts.append(article)
for art in set(arts):
item = html.fromstring(art)
try:
_time = time(
*map(int,
item.cssselect("div.time")[0].text.split(":")))
except IndexError:
continue
article = HTMLDocument(
date=datetime.combine(self.options['date'], _time),
headline=item.cssselect("h2.title")[0].text,
url=urljoin(
self.index_url.format(**locals()),
item.cssselect("h2.title")[0].getparent().get('href')),
)
yield article
def _get_units(self):
self.open("http://www.powned.tv")
self.open("http://cookies.publiekeomroep.nl/accept/")
d = self.options['date']
docs = []
for x in range(d.day - 7, d.day + 7):
archive_url = ARCHIVE_URL.format(**locals())
try:
doc = self.getdoc(archive_url)
except HTTPError:
pass
else:
docs.append(doc)
entries = set([])
for doc in docs:
for li in doc.cssselect("ul.articlelist li"):
_date = readDate(
" ".join(li.cssselect("span.t")[0].text.split()[:2]) +
" " + str(self.options['date'].year)).date()
url = urljoin(archive_url, li.cssselect("a")[0].get('href'))
entries.add((_date, url))
for _date, url in entries:
if _date == self.options['date']:
article = HTMLDocument(date=_date, url=url)
yield article
def extract_articles(self, url):
try:
doc = self.getdoc(url)
except HTTPError:
return
for tag in doc.cssselect("#main article.news"):
if 'poll' in tag.get('class'):
continue
_date = datetime.fromtimestamp(int(tag.get('created')))
article = HTMLDocument(date = _date)
if tag.cssselect("div.tweet"):
article.props.type = "tweet"
article.props.text = tag.cssselect("p")[0]
article.props.author = article.props.text.cssselect("b a")[0].get('title')
article.props.url = url.split("?")[0]
elif tag.cssselect("div.quoteBody"):
article.props.type = "quote"
a = tag.cssselect("div.quoteBody a")[0]
article.props.text = a.text_content()
article.props.url = urljoin(url, a.get('href'))
article.props.author = tag.cssselect("span.author")[0].text.strip()
elif tag.cssselect("div.videoContainer") or 'promo' in tag.get('class'):
continue
elif tag.cssselect("div.tagline h4"):
self.stories.add(urljoin(url, tag.cssselect("h4 a")[0].get('href')))
continue
else:
h = tag.cssselect("div.body h3")[0]
article.props.type = "article"
article.props.headline = h.text_content().strip()
if h.cssselect("a"):
article.props.url = urljoin(url, h.cssselect("a")[0].get('href'))
else:
article.props.url = url
yield article
def _scrape_unit(self, url):
article = HTMLDocument(url = url)
article.prepare(self)
article.props.text = article.doc.cssselect("#article p:not(#article-info):not(#metadata)")
info = article.doc.cssselect("#article-info a")
article.props.date = readDate(info[0].text)
article.props.section = info[1].text
article.props.page_str = info[2].text
article.props.headline = article.doc.cssselect("#article h1")[0].text
if article.doc.cssselect("#metadata"):
metadata = article.doc.cssselect("#metadata")[0].text_content().split("|")
for m in metadata:
if m.strip().startswith("Trefwoord"):
article.props.tags = [t.strip() for t in m.strip().split(";")]
yield article
