def lxml_extractor(html, url):
'''LXML PARSER'''
cleaner = Cleaner()
cleaner.javascript = True # This is True because we want to activate the javascript filter
cleaner.style = True # This is True because we want to activate the styles & stylesheet filter
cleaner.comments = True
cleaner.embedded = True
cleaner.forms= True
cleaner.frames = True
cleaner.annoying_tags = True
cleaner.kill_tags = NEGATIVE_K
cleaner.allow_tag = POSITIVE_K
cleaner.safe_attrs_only = True
#~ oc = document_fromstring(html, parser=parser, base_url=base_url, **kw)
#~ File "/usr/local/lib/python2.7/dist-packages/lxml/html/__init__.py", line 752, in document_fromstring
#~ value = etree.fromstring(html, parser, **kw)
try:
html = lxml.html.fromstring(html, base_url="url")
tree = cleaner.clean_html(html)
#tree.make_links_absolute(url)
doc = lxml.html.tostring(tree)
doc = soup_extractor(doc, url)
except ValueError:
doc = soup_extractor(html, url)
#~ (title, doc, article, text) = read_extractor(html, url)
#~ print title
#~ doc = (self.doc).replace(unichr(160), " ")
#~ doc = re.sub(spaces,"",self.doc)
return doc
def cleanpage(html):
# cleaner setup
cleaner = Cleaner()
cleaner.html = True
cleaner.page_structure = False
cleaner.meta = False
cleaner.safe_attrs_only = False
cleaner.links = False
cleaner.javascript = True # activate the javascript filter
cleaner.style = True # activate the styles & stylesheet filter
cleaner.links = False
cleaner.frames = True
cleaner.embedded = True
cleaner.comments = True
cleaner.annoying_tags = True
cleaner.inline_style = True
cleaner.page_structure = False
# cleaner.remove_tags = ['b','img','h']
cleaner.kill_tags = ['img','script']
#invoke cleaner
try:
content=cleaner.clean_html(html)
except:
#error: ValueError: Unicode strings with encoding declaration are not supported. Please use bytes input or XML fr
content = u""
return content
def clean_html(html_text,
javascript=True,
scripts=True,
style=True,
embedded=True,
links=True,
forms=True,
frames=True,
comments=True,
annoying_tags=True,
meta=True,
safe_attrs_only=True,
remove_unknown_tags=True,
processing_instructions=True):
"""Clean all the javascript and styles from the HTML returning the string with only the html content"""
# True = Remove | False = Keep
cleaner = Cleaner()
cleaner.javascript = javascript # This is True because we want to activate the javascript filter
cleaner.scripts = scripts # This is True because we want to activate the scripts filter
cleaner.style = style
cleaner.embedded = embedded
cleaner.links = links
cleaner.forms = forms
cleaner.frames = frames
cleaner.comments = comments
cleaner.page_structure = False # Keep page structure
cleaner.annoying_tags = annoying_tags
cleaner.meta = meta
cleaner.safe_attrs_only = safe_attrs_only
cleaner.remove_unknown_tags = remove_unknown_tags
cleaner.processing_instructions = processing_instructions
clean_content = cleaner.clean_html(lxml.html.fromstring(html_text))
return lxml.html.tostring(clean_content)
def cleaner_parameters():
reject_list = [
'script', 'noscript', 'style', 'meta', 'semantics', 'img', 'label',
'table', 'li', 'ul', 'ol', 'nav', 'dl', 'dd', 'sub', 'sup', 'math'
]
accept_list = [
'div', 'p', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6'
'span', 'b', 'a', 'u', 'i', 'body'
]
html_cleaner = Cleaner()
html_cleaner.remove_unknown_tags = True
html_cleaner.processing_instructions = True
html_cleaner.style = True
html_cleaner.comments = True
html_cleaner.scripts = True
html_cleaner.javascript = True
html_cleaner.meta = True
html_cleaner.links = True
html_cleaner.embedded = True
html_cleaner.annoying_tags = True
html_cleaner.frames = True
html_cleaner.forms = True
html_cleaner.remove_tags = accept_list
html_cleaner.kill_tags = reject_list
return html_cleaner
def get_cleaner():
cleaner = Cleaner()
cleaner.embedded = True
cleaner.frames = True
cleaner.style = True
cleaner.remove_unknown_tags = True
cleaner.processing_instructions = True
cleaner.annoying_tags = True
cleaner.remove_tags = ['h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'b', 'a', 'u', 'i', 'body', 'div', 'span', 'p']
cleaner.kill_tags = ['table', 'img', 'semantics', 'script', 'noscript', 'style', 'meta', 'label', 'li', 'ul',
'ol', 'sup', 'math', 'nav', 'dl', 'dd', 'sub']
return cleaner
def clean_text(data):
cleaner = Cleaner()
cleaner.javascript = True
cleaner.style = True
cleaner.scripts = True
cleaner.comments = True
cleaner.meta = True
cleaner.annoying_tags = True
stuff = lxml.html.tostring(cleaner.clean_html(data))
soup = BeautifulSoup(stuff.decode('utf-8', 'ignore'))
all_text = ' '.join(filter(lambda val: val, \
map(lambda x: x.strip(), soup.findAll(text=True))))
return all_text
def clean_text(data):
cleaner = Cleaner()
cleaner.javascript = True
cleaner.style = True
cleaner.scripts = True
cleaner.comments = True
cleaner.meta = True
cleaner.annoying_tags = True
doc = UnicodeDammit(data, is_html=True)
parser = html.HTMLParser(encoding=doc.original_encoding)
root = html.document_fromstring(data, parser=parser)
stuff = lxml.html.tostring(cleaner.clean_html(root))
soup = BeautifulSoup(stuff.decode('utf-8', 'ignore'))
all_text = ' '.join(filter(lambda val: val, \
map(lambda x: x.strip(), soup.findAll(text=True))))
return all_text.encode('ascii', 'ignore')
import re from lxml import etree from lxml.html.clean import Cleaner from .filters import duplicate_test, textfilter from .settings import CUT_EMPTY_ELEMS, DEFAULT_CONFIG, MANUALLY_CLEANED, MANUALLY_STRIPPED from .utils import trim LOGGER = logging.getLogger(__name__) # HTML_CLEANER config # http://lxml.de/api/lxml.html.clean.Cleaner-class.html # https://lxml.de/apidoc/lxml.html.clean.html HTML_CLEANER = Cleaner() HTML_CLEANER.annoying_tags = False # True HTML_CLEANER.comments = True HTML_CLEANER.embedded = False # True HTML_CLEANER.forms = False # True HTML_CLEANER.frames = False # True HTML_CLEANER.javascript = False HTML_CLEANER.links = False HTML_CLEANER.meta = False HTML_CLEANER.page_structure = False HTML_CLEANER.processing_instructions = True HTML_CLEANER.remove_unknown_tags = False HTML_CLEANER.safe_attrs_only = False HTML_CLEANER.scripts = False HTML_CLEANER.style = False #HTML_CLEANER.remove_tags = MANUALLY_STRIPPED #HTML_CLEANER.kill_tags = MANUALLY_CLEANED
import urllib.request
import sys
import testApp.processing as process
import re
from lxml.html.clean import Cleaner
cleaner = Cleaner()
cleaner.javascript = True # This is True because we want to activate the javascript filter
cleaner.style = True
cleaner.scripts = True
cleaner.links = True
cleaner.meta = True
cleaner.page_structure = True
cleaner.frames = True
cleaner.forms = True
cleaner.annoying_tags = True
def get_url_content(url):
try:
with urllib.request.urlopen(url) as page:
text = page.read()
except Exception as e:
return "Couldn't load url"
return text
def index(request):
return HttpResponse("Hello, world. You're at the homepage.")
'.//*[(self::div or self::section or self::ul)][starts-with(@class, "author-") or starts-with(@id, "shar") or starts-with(@class, "shar") or contains(@class, "share-") or contains(@id, "social") or contains(@class, "social") or starts-with(@id, "jp-") or starts-with(@id, "dpsp-content")]', \
'.//*[(self::div or self::section)][contains(@id, "author") or contains(@class, "author")]', \
# './/aside', \ # conflicts with text extraction
]
COMMENTS_DISCARD_XPATH = ['.//*[(self::div or self::section)][starts-with(@id, "respond")]', \
'.//cite', \
'.//quote', \
'.//*[starts-with(@id, "reply-") or starts-with(@class, "reply-title")]', \
'.//*[contains(@id, "akismet") or contains(@class, "akismet")]', \
]
# HTML_CLEANER config # http://lxml.de/api/lxml.html.clean.Cleaner-class.html
HTML_CLEANER = Cleaner()
HTML_CLEANER.annoying_tags = True
HTML_CLEANER.comments = True
HTML_CLEANER.embedded = True
HTML_CLEANER.forms = True
HTML_CLEANER.frames = True
HTML_CLEANER.javascript = True
HTML_CLEANER.links = False
HTML_CLEANER.meta = False
HTML_CLEANER.page_structure = False
HTML_CLEANER.processing_instructions = True
HTML_CLEANER.remove_unknown_tags = False
HTML_CLEANER.safe_attrs_only = False
HTML_CLEANER.scripts = True
HTML_CLEANER.style = False
HTML_CLEANER.remove_tags = [
'a', 'abbr', 'acronym', 'address', 'big', 'cite', 'font', 'ins', 'meta',
def scrape(lineHashDB, html, encoding):
# cleaner setup
cleaner = Cleaner(allow_tags=['div', 'p'], remove_unknown_tags=False)
cleaner.javascript = True # activate the javascript filter
cleaner.style = True # activate the styles & stylesheet filter
cleaner.comments = True
cleaner.annoying_tags = True
cleaner.inline_style = True
cleaner.page_structure = False
cleaner.remove_tags = ['b', 'a', 'h']
cleaner.kill_tags = ['script']
#invoke cleaner
try:
page = cleaner.clean_html(html)
except:
#error: ValueError: Unicode strings with encoding declaration are not supported. Please use bytes input or XML fr
content = u""
return content
page8 = page
page8 = re.sub(u'\n', ' ', page8) # remove NL
# page8 = re.sub(u'\s','',page8,re.UNICODE) # blanks -> space
page8 = re.sub(u' ', ' ', page8) # remove CR
page8 = re.sub(u'<!--.*?-->', ' ', page8) # remove comments
page8 = re.sub(u' class=".*?"', ' ', page8) # remove attributes
page8 = re.sub(u' id=".*?"', ' ', page8)
page8 = re.sub(u' rel=".*?"', ' ', page8)
page8 = re.sub(u'\[an error occurred while processing this directive\]',
' ', page8)
page8 = re.sub(u'>\s*?<', '><', page8) # remove blanks between tags
# cycle to remove spurious divs
for count in range(1, 20):
page8 = re.sub(u'>.{0,10}<', '><',
page8) # remove words under 10 chars between tags
page8 = re.sub(u'<div></div>', ' ', page8)
page8 = re.sub(u'<p></p>', ' ', page8)
page8 = re.sub(u'<span></span>', ' ', page8)
page8 = re.sub(u'\s+', ' ', page8) # remove repeated blanks
#XPATHs
xpath = '//*[((p) or (a) or (b) or (div) or (span)) ]/node()[(string-length() > 300)]/text()'
xpath = '//*[((p) or (div))]/node()[(string-length() > 100)]/text()'
sel = Selector(text=page8, type="html")
text = sel.xpath(xpath).extract()
content = u""
if text:
for s in text:
# squash duplicate whitespaces
' '.join(s.split())
# remove short lines
# on empirical analysis, no unfrequent sentence under 40 chars is a relevant part of the article text, excluding repetition of title, authors, dates, etc.
if len(s) < 40:
next
# remove leading whitespace
#if s.endswith(" "): s = s[:-1]
if s.startswith(" "): s = s[1:]
content += s
content += "\n"
return content
