def parse_item(self, item):
#item: a list of html tags
article = Article(metastring={})
article.text = self._parse_text(item)
for tag in item:
if tag.tag == "h2":
if tag.text:
article.headline = tag.text
else:
article.headline = tag.cssselect("span")[0].text_content()
elif tag.tag == "i" or (tag.tag == "p" and tag.cssselect("i")):
article = self.parse_dateline(tag.text_content(), article)
if not article.headline:
raise Exception("Article has no headline")
return article
def parse_item(self, item):
#item: a list of html tags
article = Article(metastring={})
article.text = self._parse_text(item)
for tag in item:
if tag.tag == "h2":
if tag.text:
article.headline = tag.text
else:
article.headline = tag.cssselect("span")[0].text_content()
elif tag.tag == "i" or (tag.tag == "p" and tag.cssselect("i")):
article = self.parse_dateline(tag.text_content(), article)
if not article.headline:
raise Exception("Article has no headline")
return article
def scrape_1(self, _html, t):
"""format of mostly 2013"""
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 = Article(metastring={})
article.metastring['html'] = div
article.headline = div.cssselect("#articleTitle")[0].text_content()
article.text = div.cssselect("#articleIntro")[0]
articlepage = div.cssselect("#articlePage")
if articlepage:
article.pagenr, article.section = self.get_pagenum(
articlepage[0].text)
article.medium = self.get_medium(
div.cssselect("#sourceTitle")[0].text)
date_str = div.cssselect("#articleDate")[0].text
try:
article.date = readDate(date_str)
except ValueError:
log.error(
"parsing date \"{date_str}\" failed".format(**locals()))
else:
yield article
def scrape_1(self, _html, t):
"""format of mostly 2013"""
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')]
else:
raise ValueError("Neither 'werkmap' nor 'intranet/rss' in html.")
for div in divs:
article = Article(metastring=div.text_content())
article.headline = div.cssselect("#articleTitle")[0].text_content()
article.text = div.cssselect("#articleIntro")[0].text_content()
articlepage = div.cssselect("#articlePage")
if articlepage:
article.pagenr, article.section = self.get_pagenum(articlepage[0].text_content())
article.medium = self.get_medium(div.cssselect("#sourceTitle")[0].text_content())
date_str = div.cssselect("#articleDate")[0].text_content()
try:
article.date = readDate(date_str)
except ValueError:
log.error("parsing date \"{date_str}\" failed".format(**locals()))
else:
yield article
def parse_item(self, item):
#item: a list of html tags
article = Article()
article.text = self._parse_text(item)
headline_found = False
dateline_found = False
for tag in item:
if tag.tag == "h2" and not headline_found:
if tag.text:
article.headline = tag.text
else:
article.headline = tag.cssselect("span")[0].text_content()
headline_found = True
elif tag.tag == "i" or (tag.tag == "p" and tag.cssselect("i")) and not dateline_found:
article = self.parse_dateline(tag.text_content(), article)
dateline_found = True
if not article.headline:
raise Exception("Article has no headline")
return article
def parse_item(self, item):
#item: a list of html tags
article = Article(metastring={})
article.text = self._parse_text(item)
for tag in item:
if tag.tag == "h2":
article.headline = tag.text
elif tag.tag == "i":
article = self.parse_dateline(tag.text_content(), article)
return article
def parse_item(self, item):
#item: a list of html tags
article = Article(metastring = {})
for tag in item:
if tag.tag in ("p","div"):
if not (hasattr(article,'text') or article.text):
article.text.append(tag)
else:
article.text = [tag]
elif tag.tag == "h2":
article.headline = tag.text
elif tag.tag == "i":
article = self.parse_dateline(tag.text, article)
#process html
article.text = "\n".join([html2text(html.tostring(bit)) for bit in article.text])
return article
def parse_item(self, item):
#item: a list of html tags
article = Article(metastring={})
for tag in item:
if tag.tag in ("p", "div"):
if not (hasattr(article, 'text') or article.text):
article.text.append(tag)
else:
article.text = [tag]
elif tag.tag == "h2":
article.headline = tag.text
elif tag.tag == "i":
article = self.parse_dateline(tag.text_content(), article)
#process html
article.text = "\n".join(
[html2text(html.tostring(bit)) for bit in article.text])
return article
def _scrape_unit(self, _file):
"""unit: a pdf document"""
res = ""
parser = PDFParser()
doc = parser.load_document(_file, self.options['pdf_password'])
for page in parser.process_document(doc):
page_txt = ""
for line in parser.get_textlines(page):
page_txt += line.get_text() + "\n"
res += page_txt + "\n\n"
article = Article(text = res)
article.headline = self.getheadline(_file)
article.medium = self.options['medium']
article.section = self.options['section']
if self.options['date']:
article.date = self.options['date']
else:
article.date = date.today()
yield article
def _scrape_unit(self, _file):
"""unit: a pdf document"""
res = ""
parser = PDFParser()
doc = parser.load_document(_file, self.options['pdf_password'])
for page in parser.process_document(doc):
page_txt = ""
for line in parser.get_textlines(page):
page_txt += line.get_text() + "\n"
res += page_txt + "\n\n"
article = Article(text=res)
article.headline = self.getheadline(_file)
article.medium = self.options['medium']
article.section = self.options['section']
if self.options['date']:
article.date = self.options['date']
else:
article.date = date.today()
yield article
def scrape_3(self, _html):
"""Some ugly MS Word format, as of 2014-03-03"""
# Partition articles
part = []
articles = []
for tag in _html.cssselect("body > div > *"):
if tag.cssselect("hr"):
articles.append(part)
part = []
else:
part.append(tag)
for tags in articles[1:]:
article = Article()
dateline = tags[1].text_content().strip()
article = self.parse_dateline(dateline, article)
article.headline = tags[1].text_content().strip()
html_str = "".join([html.tostring(t) for t in tags[2:]])
article.text = html2text(html_str)
article.metastring = {'html': html_str}
yield article
def scrape_3(self, _html):
"""Some ugly MS Word format, as of 2014-03-03"""
# Partition articles
part = []
articles = []
for tag in _html.cssselect("body > div > *"):
if tag.cssselect("hr"):
articles.append(part)
part = []
else:
part.append(tag)
for tags in articles[1:]:
article = Article()
dateline = tags[1].text_content().strip()
article = self.parse_dateline(dateline, article)
article.headline = tags[1].text_content().strip()
html_str = "".join([html.tostring(t) for t in tags[2:]])
article.text = html2text(html_str)
article.metastring = {'html': html_str}
yield article
def _scrape_unit(self, _file):
readlines = _file.readlines()
file_date_line = [l for l in readlines if l.startswith("Date:")][0]
file_date = readDate(file_date_line.split("Date:")[1])
lines = []
mail_header = []
for line in readlines:
if lines:
lines.append(line.rstrip("\r\n"))
else:
mail_header.append(line)
if line.startswith("1red"): #actual content starts
lines.append("")
article = Article(metastring={'mail_header': "".join(mail_header)})
while True: #loop through lines up to and including headline
line = lines.pop(0)
if line.isupper(): #headline
article.headline = line
break
elif line: #first non-empty line, contains metadata
data = line.split(", ")
datestr = data[0]
if "'" in datestr:
split = datestr.split("'")
datestr = split[0] + "20" + split[1]
if "=" in datestr: # if this is true, the year is not parsable
# we take the year the mail was sent, might fail around december
datestr = datestr.split("=")[0] + str(file_date.year)
article.date = readDate(datestr)
if (
article.date - file_date).days > 200: #likely a misparse, with the mail being sent the next year
article.date -= timedelta(years=1)
else:
article.date = readDate(datestr)
if data[2] in BZK_ALIASES.keys():
medium_str = BZK_ALIASES[data[1]]
else:
medium_str = data[2]
article.medium = Medium.get_or_create(medium_str)
article.section = data[1]
paragraphs = []
paragraph = ""
while True:
line = lines.pop(0).rstrip("=")
if not line:
paragraphs.append(paragraph)
paragraph = ""
elif line.isupper(): #subheader
paragraph += line + "\n"
else:
paragraph += line
if not lines:
break
paragraphs = [p for p in paragraphs if p]
article.text = ""
for p in paragraphs:
article.text += p + "\n\n"
if p.startswith("(") and len(p.split(",")) > 1: #laatste regel van normale content
break
# Add non-ascii characters
# Takes the '=AB' occurrences and turns them into latin-1 characters.
def character(match):
code = match.group()[1:]
char = r"\x{}".format(code).decode('string-escape').decode('latin-1')
if code == "92":
return "'"
elif code == "85":
return "..."
return char
article.text = re.sub(
"=[A-Z0-9]{2}",
character,
article.text)
yield article
def _scrape_unit(self, _file):
readlines = _file.readlines()
file_date_line = [l for l in readlines if l.startswith("Date:")][0]
file_date = readDate(file_date_line.split("Date:")[1])
lines = []
mail_header = []
for line in readlines:
if lines:
lines.append(line.rstrip("\r\n"))
else:
mail_header.append(line)
if line.startswith("1red"): #actual content starts
lines.append("")
article = Article(metastring={'mail_header': "".join(mail_header)})
while True: #loop through lines up to and including headline
line = lines.pop(0)
if line.isupper(): #headline
article.headline = line
break
elif line: #first non-empty line, contains metadata
data = line.split(", ")
datestr = data[0]
if "'" in datestr:
split = datestr.split("'")
datestr = split[0] + "20" + split[1]
if "=" in datestr: # if this is true, the year is not parsable
# we take the year the mail was sent, might fail around december
datestr = datestr.split("=")[0] + str(file_date.year)
article.date = readDate(datestr)
if (
article.date - file_date
).days > 200: #likely a misparse, with the mail being sent the next year
article.date -= timedelta(years=1)
else:
article.date = readDate(datestr)
if data[2] in BZK_ALIASES.keys():
medium_str = BZK_ALIASES[data[1]]
else:
medium_str = data[2]
article.medium = Medium.get_or_create(medium_str)
article.section = data[1]
paragraphs = []
paragraph = ""
while True:
line = lines.pop(0).rstrip("=")
if not line:
paragraphs.append(paragraph)
paragraph = ""
elif line.isupper(): #subheader
paragraph += line + "\n"
else:
paragraph += line
if not lines:
break
paragraphs = [p for p in paragraphs if p]
article.text = ""
for p in paragraphs:
article.text += p + "\n\n"
if p.startswith("(") and len(
p.split(",")) > 1: #laatste regel van normale content
break
# Add non-ascii characters
# Takes the '=AB' occurrences and turns them into latin-1 characters.
def character(match):
code = match.group()[1:]
char = r"\x{}".format(code).decode('string-escape').decode(
'latin-1')
if code == "92": return "'"
elif code == "85": return "..."
return char
article.text = re.sub("=[A-Z0-9]{2}", character, article.text)
yield article