|
|
|
@ -3,23 +3,21 @@ import logging
|
|
|
|
|
import re
|
|
|
|
|
import sys
|
|
|
|
|
|
|
|
|
|
from collections import defaultdict
|
|
|
|
|
from lxml.etree import tostring
|
|
|
|
|
from lxml.etree import tounicode
|
|
|
|
|
from lxml.html import document_fromstring
|
|
|
|
|
from lxml.html import fragment_fromstring
|
|
|
|
|
|
|
|
|
|
from cleaners import clean_attributes
|
|
|
|
|
from cleaners import html_cleaner
|
|
|
|
|
from htmls import build_doc
|
|
|
|
|
from htmls import get_body
|
|
|
|
|
from htmls import get_title
|
|
|
|
|
from htmls import shorten_title
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
logging.basicConfig(level=logging.INFO)
|
|
|
|
|
log = logging.getLogger()
|
|
|
|
|
from .cleaners import clean_attributes
|
|
|
|
|
from .cleaners import html_cleaner
|
|
|
|
|
from .htmls import build_doc
|
|
|
|
|
from .htmls import get_body
|
|
|
|
|
from .htmls import get_title
|
|
|
|
|
from .htmls import shorten_title
|
|
|
|
|
from encoding import get_encoding
|
|
|
|
|
from debug import describe, text_content, open_in_browser
|
|
|
|
|
|
|
|
|
|
log = logging.getLogger(__file__)
|
|
|
|
|
|
|
|
|
|
REGEXES = {
|
|
|
|
|
'unlikelyCandidatesRe': re.compile('combx|comment|community|disqus|extra|foot|header|menu|remark|rss|shoutbox|sidebar|sponsor|ad-break|agegate|pagination|pager|popup|tweet|twitter', re.I),
|
|
|
|
@ -41,21 +39,6 @@ class Unparseable(ValueError):
|
|
|
|
|
pass
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def describe(node, depth=1):
|
|
|
|
|
if not hasattr(node, 'tag'):
|
|
|
|
|
return "[%s]" % type(node)
|
|
|
|
|
name = node.tag
|
|
|
|
|
if node.get('id', ''):
|
|
|
|
|
name += '#' + node.get('id')
|
|
|
|
|
if node.get('class', ''):
|
|
|
|
|
name += '.' + node.get('class').replace(' ', '.')
|
|
|
|
|
if name[:4] in ['div#', 'div.']:
|
|
|
|
|
name = name[3:]
|
|
|
|
|
if depth and node.getparent() is not None:
|
|
|
|
|
return name + ' - ' + describe(node.getparent(), depth - 1)
|
|
|
|
|
return name
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def to_int(x):
|
|
|
|
|
if not x:
|
|
|
|
|
return None
|
|
|
|
@ -68,43 +51,50 @@ def to_int(x):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def clean(text):
|
|
|
|
|
text = re.sub('[ \t]+', ' ', text)
|
|
|
|
|
text = re.sub('\s*\n\s*', '\n', text)
|
|
|
|
|
text = re.sub('[ \t]{2,}', ' ', text)
|
|
|
|
|
return text.strip()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def text_length(i):
|
|
|
|
|
return len(clean(i.text_content() or ""))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
regexp_type = type(re.compile('hello, world'))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def compile_pattern(elements):
|
|
|
|
|
if not elements:
|
|
|
|
|
return None
|
|
|
|
|
if isinstance(elements, regexp_type):
|
|
|
|
|
return elements
|
|
|
|
|
if isinstance(elements, basestring):
|
|
|
|
|
|
|
|
|
|
if isinstance(elements, _basestring):
|
|
|
|
|
elements = elements.split(',')
|
|
|
|
|
return re.compile(u'|'.join([re.escape(x.lower()) for x in elements]), re.U)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
class Document:
|
|
|
|
|
"""Class to build a etree document out of html."""
|
|
|
|
|
TEXT_LENGTH_THRESHOLD = 25
|
|
|
|
|
RETRY_LENGTH = 250
|
|
|
|
|
|
|
|
|
|
def __init__(self, input, positive_keywords=None, negative_keywords=None, **options):
|
|
|
|
|
def __init__(self, input, positive_keywords=None, negative_keywords=None,
|
|
|
|
|
**options):
|
|
|
|
|
"""Generate the document
|
|
|
|
|
|
|
|
|
|
:param input: string of the html content.
|
|
|
|
|
|
|
|
|
|
kwargs:
|
|
|
|
|
- attributes:
|
|
|
|
|
- debug: output debug messages
|
|
|
|
|
- min_text_length:
|
|
|
|
|
- retry_length:
|
|
|
|
|
- url: will allow adjusting links to be absolute
|
|
|
|
|
- positive_keywords: the list of positive search patterns in classes and ids, for example: ["news-item", "block"]
|
|
|
|
|
- negative_keywords: the list of negative search patterns in classes and ids, for example: ["mysidebar", "related", "ads"]
|
|
|
|
|
- positive_keywords: the list of positive search patterns in
|
|
|
|
|
classes and ids, for example: ["news-item", "block"]
|
|
|
|
|
- negative_keywords: the list of negative
|
|
|
|
|
search patterns in classes
|
|
|
|
|
and ids, for example: ["mysidebar", "related", "ads"]
|
|
|
|
|
Also positive_keywords and negative_keywords could be a regexp.
|
|
|
|
|
"""
|
|
|
|
|
self.input = input
|
|
|
|
@ -127,6 +117,11 @@ class Document:
|
|
|
|
|
doc.make_links_absolute(base_href, resolve_base_href=True)
|
|
|
|
|
else:
|
|
|
|
|
doc.resolve_base_href()
|
|
|
|
|
if self.options.get('xpath'):
|
|
|
|
|
root = doc.getroottree()
|
|
|
|
|
for i in doc.getiterator():
|
|
|
|
|
#print root.getpath(i)
|
|
|
|
|
i.attrib['x'] = root.getpath(i)
|
|
|
|
|
return doc
|
|
|
|
|
|
|
|
|
|
def content(self):
|
|
|
|
@ -139,7 +134,7 @@ class Document:
|
|
|
|
|
return shorten_title(self._html(True))
|
|
|
|
|
|
|
|
|
|
def get_clean_html(self):
|
|
|
|
|
return clean_attributes(tounicode(self.html))
|
|
|
|
|
return clean_attributes(tounicode(self.html))
|
|
|
|
|
|
|
|
|
|
def summary(self, html_partial=False):
|
|
|
|
|
"""Generate the summary of the html docuemnt
|
|
|
|
@ -165,18 +160,18 @@ class Document:
|
|
|
|
|
|
|
|
|
|
if best_candidate:
|
|
|
|
|
article = self.get_article(candidates, best_candidate,
|
|
|
|
|
html_partial=html_partial)
|
|
|
|
|
html_partial=html_partial)
|
|
|
|
|
else:
|
|
|
|
|
if ruthless:
|
|
|
|
|
log.debug("ruthless removal did not work. ")
|
|
|
|
|
log.info("ruthless removal did not work. ")
|
|
|
|
|
ruthless = False
|
|
|
|
|
self.debug(
|
|
|
|
|
log.info(
|
|
|
|
|
("ended up stripping too much - "
|
|
|
|
|
"going for a safer _parse"))
|
|
|
|
|
"going for a safer parse"))
|
|
|
|
|
# try again
|
|
|
|
|
continue
|
|
|
|
|
else:
|
|
|
|
|
log.debug(
|
|
|
|
|
log.info(
|
|
|
|
|
("Ruthless and lenient parsing did not work. "
|
|
|
|
|
"Returning raw html"))
|
|
|
|
|
article = self.html.find('body')
|
|
|
|
@ -194,7 +189,7 @@ class Document:
|
|
|
|
|
continue
|
|
|
|
|
else:
|
|
|
|
|
return cleaned_article
|
|
|
|
|
except StandardError, e:
|
|
|
|
|
except Exception as e:
|
|
|
|
|
log.exception('error getting summary: ')
|
|
|
|
|
raise Unparseable(str(e)), None, sys.exc_info()[2]
|
|
|
|
|
|
|
|
|
@ -219,7 +214,8 @@ class Document:
|
|
|
|
|
append = True
|
|
|
|
|
sibling_key = sibling # HashableElement(sibling)
|
|
|
|
|
if sibling_key in candidates and \
|
|
|
|
|
candidates[sibling_key]['content_score'] >= sibling_score_threshold:
|
|
|
|
|
candidates[sibling_key]['content_score'] >= \
|
|
|
|
|
sibling_score_threshold:
|
|
|
|
|
append = True
|
|
|
|
|
|
|
|
|
|
if sibling.tag == "p":
|
|
|
|
@ -230,8 +226,8 @@ class Document:
|
|
|
|
|
if node_length > 80 and link_density < 0.25:
|
|
|
|
|
append = True
|
|
|
|
|
elif node_length <= 80 \
|
|
|
|
|
and link_density == 0 \
|
|
|
|
|
and re.search('\.( |$)', node_content):
|
|
|
|
|
and link_density == 0 \
|
|
|
|
|
and re.search('\.( |$)', node_content):
|
|
|
|
|
append = True
|
|
|
|
|
|
|
|
|
|
if append:
|
|
|
|
@ -241,21 +237,26 @@ class Document:
|
|
|
|
|
output.append(sibling)
|
|
|
|
|
else:
|
|
|
|
|
output.getchildren()[0].getchildren()[0].append(sibling)
|
|
|
|
|
#if output is not None:
|
|
|
|
|
# output.append(best_elem)
|
|
|
|
|
# if output is not None:
|
|
|
|
|
# output.append(best_elem)
|
|
|
|
|
return output
|
|
|
|
|
|
|
|
|
|
def select_best_candidate(self, candidates):
|
|
|
|
|
sorted_candidates = sorted(candidates.values(), key=lambda x: x['content_score'], reverse=True)
|
|
|
|
|
if not candidates:
|
|
|
|
|
return None
|
|
|
|
|
|
|
|
|
|
sorted_candidates = sorted(
|
|
|
|
|
candidates.values(),
|
|
|
|
|
key=lambda x: x['content_score'],
|
|
|
|
|
reverse=True
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
for candidate in sorted_candidates[:5]:
|
|
|
|
|
elem = candidate['elem']
|
|
|
|
|
self.debug("Top 5 : %6.3f %s" % (
|
|
|
|
|
log.info("Top 5 : %6.3f %s" % (
|
|
|
|
|
candidate['content_score'],
|
|
|
|
|
describe(elem)))
|
|
|
|
|
|
|
|
|
|
if len(sorted_candidates) == 0:
|
|
|
|
|
return None
|
|
|
|
|
|
|
|
|
|
best_candidate = sorted_candidates[0]
|
|
|
|
|
return best_candidate
|
|
|
|
|
|
|
|
|
@ -292,7 +293,8 @@ class Document:
|
|
|
|
|
candidates[parent_node] = self.score_node(parent_node)
|
|
|
|
|
ordered.append(parent_node)
|
|
|
|
|
|
|
|
|
|
if grand_parent_node is not None and grand_parent_node not in candidates:
|
|
|
|
|
if grand_parent_node is not None and \
|
|
|
|
|
grand_parent_node not in candidates:
|
|
|
|
|
candidates[grand_parent_node] = self.score_node(
|
|
|
|
|
grand_parent_node)
|
|
|
|
|
ordered.append(grand_parent_node)
|
|
|
|
@ -315,11 +317,8 @@ class Document:
|
|
|
|
|
candidate = candidates[elem]
|
|
|
|
|
ld = self.get_link_density(elem)
|
|
|
|
|
score = candidate['content_score']
|
|
|
|
|
self.debug("Candid: %6.3f %s link density %.3f -> %6.3f" % (
|
|
|
|
|
score,
|
|
|
|
|
describe(elem),
|
|
|
|
|
ld,
|
|
|
|
|
score * (1 - ld)))
|
|
|
|
|
log.debug("Branch %6.3f %s link density %.3f -> %6.3f" % (
|
|
|
|
|
score, describe(elem), ld, score * (1 - ld)))
|
|
|
|
|
candidate['content_score'] *= (1 - ld)
|
|
|
|
|
|
|
|
|
|
return candidates
|
|
|
|
@ -334,16 +333,20 @@ class Document:
|
|
|
|
|
if REGEXES['positiveRe'].search(feature):
|
|
|
|
|
weight += 25
|
|
|
|
|
|
|
|
|
|
if self.positive_keywords and self.positive_keywords.search(feature):
|
|
|
|
|
if self.positive_keywords and self.positive_keywords.search(
|
|
|
|
|
feature):
|
|
|
|
|
weight += 25
|
|
|
|
|
|
|
|
|
|
if self.negative_keywords and self.negative_keywords.search(feature):
|
|
|
|
|
if self.negative_keywords and self.negative_keywords.search(
|
|
|
|
|
feature):
|
|
|
|
|
weight -= 25
|
|
|
|
|
|
|
|
|
|
if self.positive_keywords and self.positive_keywords.match('tag-'+e.tag):
|
|
|
|
|
if self.positive_keywords and self.positive_keywords.match(
|
|
|
|
|
'tag-' + e.tag):
|
|
|
|
|
weight += 25
|
|
|
|
|
|
|
|
|
|
if self.negative_keywords and self.negative_keywords.match('tag-'+e.tag):
|
|
|
|
|
if self.negative_keywords and self.negative_keywords.match(
|
|
|
|
|
'tag-' + e.tag):
|
|
|
|
|
weight -= 25
|
|
|
|
|
|
|
|
|
|
return weight
|
|
|
|
@ -365,8 +368,7 @@ class Document:
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
def debug(self, *a):
|
|
|
|
|
if self.options.get('debug', False):
|
|
|
|
|
log.debug(*a)
|
|
|
|
|
log.warn("debug: " + a[0], *a[1:])
|
|
|
|
|
|
|
|
|
|
def remove_unlikely_candidates(self):
|
|
|
|
|
for elem in self.html.iter():
|
|
|
|
@ -375,22 +377,22 @@ class Document:
|
|
|
|
|
continue
|
|
|
|
|
#self.debug(s)
|
|
|
|
|
if REGEXES['unlikelyCandidatesRe'].search(s) and (not REGEXES['okMaybeItsACandidateRe'].search(s)) and elem.tag not in ['html', 'body']:
|
|
|
|
|
self.debug("Removing unlikely candidate - %s" % describe(elem))
|
|
|
|
|
log.debug("Removing unlikely candidate - %s" % describe(elem))
|
|
|
|
|
elem.drop_tree()
|
|
|
|
|
|
|
|
|
|
def transform_misused_divs_into_paragraphs(self):
|
|
|
|
|
for elem in self.tags(self.html, 'div'):
|
|
|
|
|
# transform <div>s that do not contain other block elements into
|
|
|
|
|
# <p>s
|
|
|
|
|
#FIXME: The current implementation ignores all descendants that
|
|
|
|
|
# FIXME: The current implementation ignores all descendants that
|
|
|
|
|
# are not direct children of elem
|
|
|
|
|
# This results in incorrect results in case there is an <img>
|
|
|
|
|
# buried within an <a> for example
|
|
|
|
|
if not REGEXES['divToPElementsRe'].search(
|
|
|
|
|
unicode(''.join(map(tostring, list(elem))))):
|
|
|
|
|
#self.debug("Altering %s to p" % (describe(elem)))
|
|
|
|
|
# self.debug("Altering %s to p" % describe(elem))
|
|
|
|
|
elem.tag = "p"
|
|
|
|
|
#print "Fixed element "+describe(elem)
|
|
|
|
|
# self.debug("Fixed element "+describe(elem))
|
|
|
|
|
|
|
|
|
|
for elem in self.tags(self.html, 'div'):
|
|
|
|
|
if elem.text and elem.text.strip():
|
|
|
|
@ -398,7 +400,7 @@ class Document:
|
|
|
|
|
p.text = elem.text
|
|
|
|
|
elem.text = None
|
|
|
|
|
elem.insert(0, p)
|
|
|
|
|
#print "Appended "+tounicode(p)+" to "+describe(elem)
|
|
|
|
|
# print "Appended "+tounicode(p)+" to "+describe(elem)
|
|
|
|
|
|
|
|
|
|
for pos, child in reversed(list(enumerate(elem))):
|
|
|
|
|
if child.tail and child.tail.strip():
|
|
|
|
@ -406,9 +408,9 @@ class Document:
|
|
|
|
|
p.text = child.tail
|
|
|
|
|
child.tail = None
|
|
|
|
|
elem.insert(pos + 1, p)
|
|
|
|
|
#print "Inserted "+tounicode(p)+" to "+describe(elem)
|
|
|
|
|
# print "Inserted "+tounicode(p)+" to "+describe(elem)
|
|
|
|
|
if child.tag == 'br':
|
|
|
|
|
#print 'Dropped <br> at '+describe(elem)
|
|
|
|
|
# print 'Dropped <br> at '+describe(elem)
|
|
|
|
|
child.drop_tree()
|
|
|
|
|
|
|
|
|
|
def tags(self, node, *tag_names):
|
|
|
|
@ -422,14 +424,15 @@ class Document:
|
|
|
|
|
yield e
|
|
|
|
|
|
|
|
|
|
def sanitize(self, node, candidates):
|
|
|
|
|
MIN_LEN = self.options.get('min_text_length',
|
|
|
|
|
self.TEXT_LENGTH_THRESHOLD)
|
|
|
|
|
MIN_LEN = self.options.get('min_text_length', self.TEXT_LENGTH_THRESHOLD)
|
|
|
|
|
for header in self.tags(node, "h1", "h2", "h3", "h4", "h5", "h6"):
|
|
|
|
|
if self.class_weight(header) < 0 or self.get_link_density(header) > 0.33:
|
|
|
|
|
if self.class_weight(header) < 0 or \
|
|
|
|
|
self.get_link_density(header) > 0.33:
|
|
|
|
|
header.drop_tree()
|
|
|
|
|
|
|
|
|
|
for elem in self.tags(node, "form", "iframe", "textarea"):
|
|
|
|
|
elem.drop_tree()
|
|
|
|
|
|
|
|
|
|
allowed = {}
|
|
|
|
|
# Conditionally clean <table>s, <ul>s, and <div>s
|
|
|
|
|
for el in self.reverse_tags(node, "table", "ul", "div"):
|
|
|
|
@ -438,13 +441,13 @@ class Document:
|
|
|
|
|
weight = self.class_weight(el)
|
|
|
|
|
if el in candidates:
|
|
|
|
|
content_score = candidates[el]['content_score']
|
|
|
|
|
#print '!',el, '-> %6.3f' % content_score
|
|
|
|
|
# print '!',el, '-> %6.3f' % content_score
|
|
|
|
|
else:
|
|
|
|
|
content_score = 0
|
|
|
|
|
tag = el.tag
|
|
|
|
|
|
|
|
|
|
if weight + content_score < 0:
|
|
|
|
|
self.debug("Cleaned %s with score %6.3f and weight %-3s" %
|
|
|
|
|
log.info("Removed %s with score %6.3f and weight %-3s" %
|
|
|
|
|
(describe(el), content_score, weight, ))
|
|
|
|
|
el.drop_tree()
|
|
|
|
|
elif el.text_content().count(",") < 10:
|
|
|
|
@ -452,6 +455,7 @@ class Document:
|
|
|
|
|
for kind in ['p', 'img', 'li', 'a', 'embed', 'input']:
|
|
|
|
|
counts[kind] = len(el.findall('.//%s' % kind))
|
|
|
|
|
counts["li"] -= 100
|
|
|
|
|
counts["input"] -= len(el.findall('.//input[@type="hidden"]'))
|
|
|
|
|
|
|
|
|
|
# Count the text length excluding any surrounding whitespace
|
|
|
|
|
content_length = text_length(el)
|
|
|
|
@ -459,31 +463,36 @@ class Document:
|
|
|
|
|
parent_node = el.getparent()
|
|
|
|
|
if parent_node is not None:
|
|
|
|
|
if parent_node in candidates:
|
|
|
|
|
content_score = candidates[parent_node]['content_score']
|
|
|
|
|
content_score = candidates[
|
|
|
|
|
parent_node]['content_score']
|
|
|
|
|
else:
|
|
|
|
|
content_score = 0
|
|
|
|
|
#if parent_node is not None:
|
|
|
|
|
#pweight = self.class_weight(parent_node) + content_score
|
|
|
|
|
#pname = describe(parent_node)
|
|
|
|
|
#else:
|
|
|
|
|
#pweight = 0
|
|
|
|
|
#pname = "no parent"
|
|
|
|
|
# if parent_node is not None:
|
|
|
|
|
# pweight = self.class_weight(parent_node) + content_score
|
|
|
|
|
# pname = describe(parent_node)
|
|
|
|
|
# else:
|
|
|
|
|
# pweight = 0
|
|
|
|
|
# pname = "no parent"
|
|
|
|
|
to_remove = False
|
|
|
|
|
reason = ""
|
|
|
|
|
|
|
|
|
|
#if el.tag == 'div' and counts["img"] >= 1:
|
|
|
|
|
# continue
|
|
|
|
|
if counts["p"] and counts["img"] > counts["p"]:
|
|
|
|
|
reason = "too many images (%s)" % counts["img"]
|
|
|
|
|
# if el.tag == 'div' and counts["img"] >= 1:
|
|
|
|
|
# continue
|
|
|
|
|
if content_length and counts["img"] * 100 >= content_length:
|
|
|
|
|
reason = "too many images (%s) for text " % counts["img"]
|
|
|
|
|
to_remove = True
|
|
|
|
|
elif counts["li"] > counts["p"] and tag != "ul" and tag != "ol":
|
|
|
|
|
elif counts["li"] > counts["p"] \
|
|
|
|
|
and tag != "ul" and tag != "ol":
|
|
|
|
|
reason = "more <li>s than <p>s"
|
|
|
|
|
to_remove = True
|
|
|
|
|
elif counts["input"] > (counts["p"] / 3):
|
|
|
|
|
reason = "less than 3x <p>s than <input>s"
|
|
|
|
|
to_remove = True
|
|
|
|
|
elif content_length < (MIN_LEN) and (counts["img"] == 0 or counts["img"] > 2):
|
|
|
|
|
reason = "too short content length %s without a single image" % content_length
|
|
|
|
|
elif content_length < MIN_LEN and not counts["img"]:
|
|
|
|
|
reason = "too short content length %s and no images" % content_length
|
|
|
|
|
to_remove = True
|
|
|
|
|
elif content_length < MIN_LEN and counts["img"] > 2:
|
|
|
|
|
reason = "too short content length %s and too much images" % content_length
|
|
|
|
|
to_remove = True
|
|
|
|
|
elif weight < 25 and link_density > 0.2:
|
|
|
|
|
reason = "too many links %.3f for its weight %s" % (
|
|
|
|
@ -496,6 +505,9 @@ class Document:
|
|
|
|
|
elif (counts["embed"] == 1 and content_length < 75) or counts["embed"] > 1:
|
|
|
|
|
reason = "<embed>s with too short content length, or too many <embed>s"
|
|
|
|
|
to_remove = True
|
|
|
|
|
elif not content_length:
|
|
|
|
|
reason = "no content"
|
|
|
|
|
to_remove = True
|
|
|
|
|
# if el.tag == 'div' and counts['img'] >= 1 and to_remove:
|
|
|
|
|
# imgs = el.findall('.//img')
|
|
|
|
|
# valid_img = False
|
|
|
|
@ -523,35 +535,38 @@ class Document:
|
|
|
|
|
#self.debug(sib.text_content())
|
|
|
|
|
sib_content_length = text_length(sib)
|
|
|
|
|
if sib_content_length:
|
|
|
|
|
i =+ 1
|
|
|
|
|
i += 1
|
|
|
|
|
siblings.append(sib_content_length)
|
|
|
|
|
if i == x:
|
|
|
|
|
if i >= x:
|
|
|
|
|
break
|
|
|
|
|
for sib in el.itersiblings(preceding=True):
|
|
|
|
|
#self.debug(sib.text_content())
|
|
|
|
|
sib_content_length = text_length(sib)
|
|
|
|
|
if sib_content_length:
|
|
|
|
|
j =+ 1
|
|
|
|
|
j += 1
|
|
|
|
|
siblings.append(sib_content_length)
|
|
|
|
|
if j == x:
|
|
|
|
|
if j >= x:
|
|
|
|
|
break
|
|
|
|
|
#self.debug(str(siblings))
|
|
|
|
|
if siblings and sum(siblings) > 1000:
|
|
|
|
|
to_remove = False
|
|
|
|
|
self.debug("Allowing %s" % describe(el))
|
|
|
|
|
log.info("Allowing %s" % describe(el))
|
|
|
|
|
for desnode in self.tags(el, "table", "ul", "div"):
|
|
|
|
|
allowed[desnode] = True
|
|
|
|
|
|
|
|
|
|
if to_remove:
|
|
|
|
|
self.debug("Cleaned %6.3f %s with weight %s cause it has %s." %
|
|
|
|
|
(content_score, describe(el), weight, reason))
|
|
|
|
|
log.info("Cleaned %s (score=%6.3f, weight=%s) cause it has %s: %s" %
|
|
|
|
|
(describe(el), content_score, weight, reason, text_content(el)))
|
|
|
|
|
#print tounicode(el)
|
|
|
|
|
#self.debug("pname %s pweight %.3f" %(pname, pweight))
|
|
|
|
|
el.drop_tree()
|
|
|
|
|
else:
|
|
|
|
|
log.info("Not cleaned %s of length %s: %s" %
|
|
|
|
|
(describe(el), content_length, text_content(el)))
|
|
|
|
|
|
|
|
|
|
for el in ([node] + [n for n in node.iter()]):
|
|
|
|
|
if not self.options.get('attributes', None):
|
|
|
|
|
#el.attrib = {} #FIXME:Checkout the effects of disabling this
|
|
|
|
|
# el.attrib = {} #FIXME:Checkout the effects of disabling this
|
|
|
|
|
pass
|
|
|
|
|
|
|
|
|
|
self.html = node
|
|
|
|
@ -584,12 +599,23 @@ class HashableElement():
|
|
|
|
|
def __getattr__(self, tag):
|
|
|
|
|
return getattr(self.node, tag)
|
|
|
|
|
|
|
|
|
|
VERBOSITY = {
|
|
|
|
|
1: logging.WARNING,
|
|
|
|
|
2: logging.INFO,
|
|
|
|
|
3: logging.DEBUG
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def main():
|
|
|
|
|
from optparse import OptionParser
|
|
|
|
|
parser = OptionParser(usage="%prog: [options] [file]")
|
|
|
|
|
parser.add_option('-v', '--verbose', action='store_true')
|
|
|
|
|
parser.add_option('-v', '--verbose', action='count', default=0)
|
|
|
|
|
parser.add_option('-b', '--browser', default=None, action='store_true', help="open in browser")
|
|
|
|
|
parser.add_option('-l', '--log', default=None, help="use filename for logs (appended)")
|
|
|
|
|
parser.add_option('-u', '--url', default=None, help="use URL instead of a local file")
|
|
|
|
|
parser.add_option('-s', '--show-xpath', default=None, help="show xpath")
|
|
|
|
|
parser.add_option('-x', '--xpath', default=None, help="use xpath")
|
|
|
|
|
parser.add_option('-t', '--support-text', default=None, help="use this support text")
|
|
|
|
|
parser.add_option('-p', '--positive-keywords', default=None, help="positive keywords (separated with comma)", action='store')
|
|
|
|
|
parser.add_option('-n', '--negative-keywords', default=None, help="negative keywords (separated with comma)", action='store')
|
|
|
|
|
(options, args) = parser.parse_args()
|
|
|
|
@ -598,20 +624,33 @@ def main():
|
|
|
|
|
parser.print_help()
|
|
|
|
|
sys.exit(1)
|
|
|
|
|
|
|
|
|
|
if options.verbose:
|
|
|
|
|
logging.basicConfig(level=VERBOSITY[options.verbose], filename=options.log,
|
|
|
|
|
format='%(asctime)s: %(levelname)s: %(message)s (at %(filename)s: %(lineno)d)')
|
|
|
|
|
|
|
|
|
|
file = None
|
|
|
|
|
if options.url:
|
|
|
|
|
import urllib
|
|
|
|
|
file = urllib.urlopen(options.url)
|
|
|
|
|
else:
|
|
|
|
|
file = open(args[0], 'rt')
|
|
|
|
|
enc = sys.__stdout__.encoding or 'utf-8' # XXX: this hack could not always work, better to set PYTHONIOENCODING
|
|
|
|
|
html = file.read() # bytes object
|
|
|
|
|
|
|
|
|
|
encoding = get_encoding(html)
|
|
|
|
|
html = html.decode(encoding)
|
|
|
|
|
try:
|
|
|
|
|
print Document(file.read(),
|
|
|
|
|
debug=options.verbose,
|
|
|
|
|
url=options.url,
|
|
|
|
|
positive_keywords = options.positive_keywords,
|
|
|
|
|
negative_keywords = options.negative_keywords,
|
|
|
|
|
).summary().encode(enc, 'replace')
|
|
|
|
|
doc = Document(html, url=options.url,
|
|
|
|
|
positive_keywords=options.positive_keywords,
|
|
|
|
|
negative_keywords=options.negative_keywords)
|
|
|
|
|
if options.browser:
|
|
|
|
|
result = 'Title: ' + doc.short_title() + '<br/>' + doc.summary()
|
|
|
|
|
open_in_browser(result)
|
|
|
|
|
else:
|
|
|
|
|
# XXX: a hack, better to set PYTHONIOENCODING explicitly
|
|
|
|
|
output_encoding = sys.__stdout__.encoding or 'utf-8'
|
|
|
|
|
|
|
|
|
|
print 'Title:', doc.short_title().encode(output_encoding, 'replace')
|
|
|
|
|
print doc.summary().encode(output_encoding, 'replace')
|
|
|
|
|
finally:
|
|
|
|
|
file.close()
|
|
|
|
|
|
|
|
|
|