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 getarticle(self, headline, lines):
article = Article(headline=headline)
text = ""
for line in lines[2:]:
if len(line) > 2:
text += "\n" + line
text = text.replace("-\n", "")
text = text.replace(" ", " ")
text = text.replace("\n", " ")
article.text = text
date_pattern = re.compile("([0-9]{2,2})\-([0-9]{2,2})\-([0-9]{4,4})")
result = date_pattern.search(lines[1])
article.date = date(int(result.group(3)), int(result.group(2)),
int(result.group(1)))
pagenum_pattern = re.compile("\(p.([0-9]+)([0-9\-]+)?\)")
result = pagenum_pattern.search(lines[1])
if result:
article.pagenr = int(result.group(1))
for h, medium in self.index:
if article.headline.lower().strip() in h.lower().strip():
article.set_property("medium", self.get_medium(medium))
return article
def getarticle(self, headline, lines):
article = Article(headline = headline)
text = ""
for line in lines[2:]:
if len(line) > 2:
text += "\n" + line
text = text.replace("-\n","")
text = text.replace(" "," ")
text = text.replace("\n"," ")
article.text = text
date_pattern = re.compile("([0-9]{2,2})\-([0-9]{2,2})\-([0-9]{4,4})")
result = date_pattern.search(lines[1])
article.date = date(
int(result.group(3)),
int(result.group(2)),
int(result.group(1)))
pagenum_pattern = re.compile("\(p.([0-9]+)([0-9\-]+)?\)")
result = pagenum_pattern.search(lines[1])
if result:
article.pagenr = int(result.group(1))
for h, medium in self.index:
if article.headline.lower().strip() in h.lower().strip():
article.set_property("medium", self.get_medium(medium))
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')]
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 _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_unit(self, _file):
readlines = _file.readlines()
file_date_line = [l for l in readlines if l.startswith("Date:")][0]
file_date = read_date(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():
article.title = 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 = read_date(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 = read_date(datestr)
if data[2] in BZK_ALIASES.keys():
medium_str = BZK_ALIASES[data[1]]
else:
medium_str = data[2]
article.set_property("medium", medium_str)
article.set_property("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 = read_date(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():
article.title = 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 = read_date(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 = read_date(datestr)
if data[2] in BZK_ALIASES.keys():
medium_str = BZK_ALIASES[data[1]]
else:
medium_str = data[2]
article.set_property("medium", medium_str)
article.set_property("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
yield article