diff options
Diffstat (limited to 'plugin/readability')
| -rwxr-xr-x | plugin/readability/BeautifulSoup.py | 2000 | ||||
| -rwxr-xr-x | plugin/readability/README | 24 | ||||
| -rwxr-xr-x | plugin/readability/__init__.py | 2 | ||||
| -rwxr-xr-x | plugin/readability/page_parser.py | 145 | ||||
| -rwxr-xr-x | plugin/readability/readability.py | 343 | ||||
| -rwxr-xr-x | plugin/readability/setup.py | 22 | ||||
| -rwxr-xr-x | plugin/readability/url_helpers.py | 52 |
7 files changed, 0 insertions, 2588 deletions
diff --git a/plugin/readability/BeautifulSoup.py b/plugin/readability/BeautifulSoup.py deleted file mode 100755 index 34204e7..0000000 --- a/plugin/readability/BeautifulSoup.py +++ /dev/null @@ -1,2000 +0,0 @@ -"""Beautiful Soup -Elixir and Tonic -"The Screen-Scraper's Friend" -http://www.crummy.com/software/BeautifulSoup/ - -Beautiful Soup parses a (possibly invalid) XML or HTML document into a -tree representation. It provides methods and Pythonic idioms that make -it easy to navigate, search, and modify the tree. - -A well-formed XML/HTML document yields a well-formed data -structure. An ill-formed XML/HTML document yields a correspondingly -ill-formed data structure. If your document is only locally -well-formed, you can use this library to find and process the -well-formed part of it. - -Beautiful Soup works with Python 2.2 and up. It has no external -dependencies, but you'll have more success at converting data to UTF-8 -if you also install these three packages: - -* chardet, for auto-detecting character encodings - http://chardet.feedparser.org/ -* cjkcodecs and iconv_codec, which add more encodings to the ones supported - by stock Python. - http://cjkpython.i18n.org/ - -Beautiful Soup defines classes for two main parsing strategies: - - * BeautifulStoneSoup, for parsing XML, SGML, or your domain-specific - language that kind of looks like XML. - - * BeautifulSoup, for parsing run-of-the-mill HTML code, be it valid - or invalid. This class has web browser-like heuristics for - obtaining a sensible parse tree in the face of common HTML errors. - -Beautiful Soup also defines a class (UnicodeDammit) for autodetecting -the encoding of an HTML or XML document, and converting it to -Unicode. Much of this code is taken from Mark Pilgrim's Universal Feed Parser. - -For more than you ever wanted to know about Beautiful Soup, see the -documentation: -http://www.crummy.com/software/BeautifulSoup/documentation.html - -Here, have some legalese: - -Copyright (c) 2004-2009, Leonard Richardson - -All rights reserved. - -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are -met: - - * Redistributions of source code must retain the above copyright - notice, this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above - copyright notice, this list of conditions and the following - disclaimer in the documentation and/or other materials provided - with the distribution. - - * Neither the name of the the Beautiful Soup Consortium and All - Night Kosher Bakery nor the names of its contributors may be - used to endorse or promote products derived from this software - without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS -"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT -LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR -A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR -CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, -EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, -PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR -PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF -LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING -NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS -SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE, DAMMIT. - -""" -from __future__ import generators - -__author__ = "Leonard Richardson (leonardr@segfault.org)" -__version__ = "3.1.0.1" -__copyright__ = "Copyright (c) 2004-2009 Leonard Richardson" -__license__ = "New-style BSD" - -import codecs -import markupbase -import types -import re -from HTMLParser import HTMLParser, HTMLParseError -try: - from htmlentitydefs import name2codepoint -except ImportError: - name2codepoint = {} -try: - set -except NameError: - from sets import Set as set - -#These hacks make Beautiful Soup able to parse XML with namespaces -markupbase._declname_match = re.compile(r'[a-zA-Z][-_.:a-zA-Z0-9]*\s*').match - -DEFAULT_OUTPUT_ENCODING = "utf-8" - -# First, the classes that represent markup elements. - -def sob(unicode, encoding): - """Returns either the given Unicode string or its encoding.""" - if encoding is None: - return unicode - else: - return unicode.encode(encoding) - -class PageElement: - """Contains the navigational information for some part of the page - (either a tag or a piece of text)""" - - def setup(self, parent=None, previous=None): - """Sets up the initial relations between this element and - other elements.""" - self.parent = parent - self.previous = previous - self.next = None - self.previousSibling = None - self.nextSibling = None - if self.parent and self.parent.contents: - self.previousSibling = self.parent.contents[-1] - self.previousSibling.nextSibling = self - - def replaceWith(self, replaceWith): - oldParent = self.parent - myIndex = self.parent.contents.index(self) - if hasattr(replaceWith, 'parent') and replaceWith.parent == self.parent: - # We're replacing this element with one of its siblings. - index = self.parent.contents.index(replaceWith) - if index and index < myIndex: - # Furthermore, it comes before this element. That - # means that when we extract it, the index of this - # element will change. - myIndex = myIndex - 1 - self.extract() - oldParent.insert(myIndex, replaceWith) - - def extract(self): - """Destructively rips this element out of the tree.""" - if self.parent: - try: - self.parent.contents.remove(self) - except ValueError: - pass - - #Find the two elements that would be next to each other if - #this element (and any children) hadn't been parsed. Connect - #the two. - lastChild = self._lastRecursiveChild() - nextElement = lastChild.next - - if self.previous: - self.previous.next = nextElement - if nextElement: - nextElement.previous = self.previous - self.previous = None - lastChild.next = None - - self.parent = None - if self.previousSibling: - self.previousSibling.nextSibling = self.nextSibling - if self.nextSibling: - self.nextSibling.previousSibling = self.previousSibling - self.previousSibling = self.nextSibling = None - return self - - def _lastRecursiveChild(self): - "Finds the last element beneath this object to be parsed." - lastChild = self - while hasattr(lastChild, 'contents') and lastChild.contents: - lastChild = lastChild.contents[-1] - return lastChild - - def insert(self, position, newChild): - if (isinstance(newChild, basestring) - or isinstance(newChild, unicode)) \ - and not isinstance(newChild, NavigableString): - newChild = NavigableString(newChild) - - position = min(position, len(self.contents)) - if hasattr(newChild, 'parent') and newChild.parent != None: - # We're 'inserting' an element that's already one - # of this object's children. - if newChild.parent == self: - index = self.find(newChild) - if index and index < position: - # Furthermore we're moving it further down the - # list of this object's children. That means that - # when we extract this element, our target index - # will jump down one. - position = position - 1 - newChild.extract() - - newChild.parent = self - previousChild = None - if position == 0: - newChild.previousSibling = None - newChild.previous = self - else: - previousChild = self.contents[position-1] - newChild.previousSibling = previousChild - newChild.previousSibling.nextSibling = newChild - newChild.previous = previousChild._lastRecursiveChild() - if newChild.previous: - newChild.previous.next = newChild - - newChildsLastElement = newChild._lastRecursiveChild() - - if position >= len(self.contents): - newChild.nextSibling = None - - parent = self - parentsNextSibling = None - while not parentsNextSibling: - parentsNextSibling = parent.nextSibling - parent = parent.parent - if not parent: # This is the last element in the document. - break - if parentsNextSibling: - newChildsLastElement.next = parentsNextSibling - else: - newChildsLastElement.next = None - else: - nextChild = self.contents[position] - newChild.nextSibling = nextChild - if newChild.nextSibling: - newChild.nextSibling.previousSibling = newChild - newChildsLastElement.next = nextChild - - if newChildsLastElement.next: - newChildsLastElement.next.previous = newChildsLastElement - self.contents.insert(position, newChild) - - def append(self, tag): - """Appends the given tag to the contents of this tag.""" - self.insert(len(self.contents), tag) - - def findNext(self, name=None, attrs={}, text=None, **kwargs): - """Returns the first item that matches the given criteria and - appears after this Tag in the document.""" - return self._findOne(self.findAllNext, name, attrs, text, **kwargs) - - def findAllNext(self, name=None, attrs={}, text=None, limit=None, - **kwargs): - """Returns all items that match the given criteria and appear - after this Tag in the document.""" - return self._findAll(name, attrs, text, limit, self.nextGenerator, - **kwargs) - - def findNextSibling(self, name=None, attrs={}, text=None, **kwargs): - """Returns the closest sibling to this Tag that matches the - given criteria and appears after this Tag in the document.""" - return self._findOne(self.findNextSiblings, name, attrs, text, - **kwargs) - - def findNextSiblings(self, name=None, attrs={}, text=None, limit=None, - **kwargs): - """Returns the siblings of this Tag that match the given - criteria and appear after this Tag in the document.""" - return self._findAll(name, attrs, text, limit, - self.nextSiblingGenerator, **kwargs) - fetchNextSiblings = findNextSiblings # Compatibility with pre-3.x - - def findPrevious(self, name=None, attrs={}, text=None, **kwargs): - """Returns the first item that matches the given criteria and - appears before this Tag in the document.""" - return self._findOne(self.findAllPrevious, name, attrs, text, **kwargs) - - def findAllPrevious(self, name=None, attrs={}, text=None, limit=None, - **kwargs): - """Returns all items that match the given criteria and appear - before this Tag in the document.""" - return self._findAll(name, attrs, text, limit, self.previousGenerator, - **kwargs) - fetchPrevious = findAllPrevious # Compatibility with pre-3.x - - def findPreviousSibling(self, name=None, attrs={}, text=None, **kwargs): - """Returns the closest sibling to this Tag that matches the - given criteria and appears before this Tag in the document.""" - return self._findOne(self.findPreviousSiblings, name, attrs, text, - **kwargs) - - def findPreviousSiblings(self, name=None, attrs={}, text=None, - limit=None, **kwargs): - """Returns the siblings of this Tag that match the given - criteria and appear before this Tag in the document.""" - return self._findAll(name, attrs, text, limit, - self.previousSiblingGenerator, **kwargs) - fetchPreviousSiblings = findPreviousSiblings # Compatibility with pre-3.x - - def findParent(self, name=None, attrs={}, **kwargs): - """Returns the closest parent of this Tag that matches the given - criteria.""" - # NOTE: We can't use _findOne because findParents takes a different - # set of arguments. - r = None - l = self.findParents(name, attrs, 1) - if l: - r = l[0] - return r - - def findParents(self, name=None, attrs={}, limit=None, **kwargs): - """Returns the parents of this Tag that match the given - criteria.""" - - return self._findAll(name, attrs, None, limit, self.parentGenerator, - **kwargs) - fetchParents = findParents # Compatibility with pre-3.x - - #These methods do the real heavy lifting. - - def _findOne(self, method, name, attrs, text, **kwargs): - r = None - l = method(name, attrs, text, 1, **kwargs) - if l: - r = l[0] - return r - - def _findAll(self, name, attrs, text, limit, generator, **kwargs): - "Iterates over a generator looking for things that match." - - if isinstance(name, SoupStrainer): - strainer = name - else: - # Build a SoupStrainer - strainer = SoupStrainer(name, attrs, text, **kwargs) - results = ResultSet(strainer) - g = generator() - while True: - try: - i = g.next() - except StopIteration: - break - if i: - found = strainer.search(i) - if found: - results.append(found) - if limit and len(results) >= limit: - break - return results - - #These Generators can be used to navigate starting from both - #NavigableStrings and Tags. - def nextGenerator(self): - i = self - while i: - i = i.next - yield i - - def nextSiblingGenerator(self): - i = self - while i: - i = i.nextSibling - yield i - - def previousGenerator(self): - i = self - while i: - i = i.previous - yield i - - def previousSiblingGenerator(self): - i = self - while i: - i = i.previousSibling - yield i - - def parentGenerator(self): - i = self - while i: - i = i.parent - yield i - - # Utility methods - def substituteEncoding(self, str, encoding=None): - encoding = encoding or "utf-8" - return str.replace("%SOUP-ENCODING%", encoding) - - def toEncoding(self, s, encoding=None): - """Encodes an object to a string in some encoding, or to Unicode. - .""" - if isinstance(s, unicode): - if encoding: - s = s.encode(encoding) - elif isinstance(s, str): - if encoding: - s = s.encode(encoding) - else: - s = unicode(s) - else: - if encoding: - s = self.toEncoding(str(s), encoding) - else: - s = unicode(s) - return s - -class NavigableString(unicode, PageElement): - - def __new__(cls, value): - """Create a new NavigableString. - - When unpickling a NavigableString, this method is called with - the string in DEFAULT_OUTPUT_ENCODING. That encoding needs to be - passed in to the superclass's __new__ or the superclass won't know - how to handle non-ASCII characters. - """ - if isinstance(value, unicode): - return unicode.__new__(cls, value) - return unicode.__new__(cls, value, DEFAULT_OUTPUT_ENCODING) - - def __getnewargs__(self): - return (unicode(self),) - - def __getattr__(self, attr): - """text.string gives you text. This is for backwards - compatibility for Navigable*String, but for CData* it lets you - get the string without the CData wrapper.""" - if attr == 'string': - return self - else: - raise AttributeError, "'%s' object has no attribute '%s'" % (self.__class__.__name__, attr) - - def encode(self, encoding=DEFAULT_OUTPUT_ENCODING): - return self.decode().encode(encoding) - - def decodeGivenEventualEncoding(self, eventualEncoding): - return self - -class CData(NavigableString): - - def decodeGivenEventualEncoding(self, eventualEncoding): - return u'<![CDATA[' + self + u']]>' - -class ProcessingInstruction(NavigableString): - - def decodeGivenEventualEncoding(self, eventualEncoding): - output = self - if u'%SOUP-ENCODING%' in output: - output = self.substituteEncoding(output, eventualEncoding) - return u'<?' + output + u'?>' - -class Comment(NavigableString): - def decodeGivenEventualEncoding(self, eventualEncoding): - return u'<!--' + self + u'-->' - -class Declaration(NavigableString): - def decodeGivenEventualEncoding(self, eventualEncoding): - return u'<!' + self + u'>' - -class Tag(PageElement): - - """Represents a found HTML tag with its attributes and contents.""" - - def _invert(h): - "Cheap function to invert a hash." - i = {} - for k,v in h.items(): - i[v] = k - return i - - XML_ENTITIES_TO_SPECIAL_CHARS = { "apos" : "'", - "quot" : '"', - "amp" : "&", - "lt" : "<", - "gt" : ">" } - - XML_SPECIAL_CHARS_TO_ENTITIES = _invert(XML_ENTITIES_TO_SPECIAL_CHARS) - - def _convertEntities(self, match): - """Used in a call to re.sub to replace HTML, XML, and numeric - entities with the appropriate Unicode characters. If HTML - entities are being converted, any unrecognized entities are - escaped.""" - x = match.group(1) - if self.convertHTMLEntities and x in name2codepoint: - return unichr(name2codepoint[x]) - elif x in self.XML_ENTITIES_TO_SPECIAL_CHARS: - if self.convertXMLEntities: - return self.XML_ENTITIES_TO_SPECIAL_CHARS[x] - else: - return u'&%s;' % x - elif len(x) > 0 and x[0] == '#': - # Handle numeric entities - if len(x) > 1 and x[1] == 'x': - return unichr(int(x[2:], 16)) - else: - return unichr(int(x[1:])) - - elif self.escapeUnrecognizedEntities: - return u'&%s;' % x - else: - return u'&%s;' % x - - def __init__(self, parser, name, attrs=None, parent=None, - previous=None): - "Basic constructor." - - # We don't actually store the parser object: that lets extracted - # chunks be garbage-collected - self.parserClass = parser.__class__ - self.isSelfClosing = parser.isSelfClosingTag(name) - self.name = name - if attrs == None: - attrs = [] - self.attrs = attrs - self.contents = [] - self.setup(parent, previous) - self.hidden = False - self.containsSubstitutions = False - self.convertHTMLEntities = parser.convertHTMLEntities - self.convertXMLEntities = parser.convertXMLEntities - self.escapeUnrecognizedEntities = parser.escapeUnrecognizedEntities - - def convert(kval): - "Converts HTML, XML and numeric entities in the attribute value." - k, val = kval - if val is None: - return kval - return (k, re.sub("&(#\d+|#x[0-9a-fA-F]+|\w+);", - self._convertEntities, val)) - self.attrs = map(convert, self.attrs) - - def get(self, key, default=None): - """Returns the value of the 'key' attribute for the tag, or - the value given for 'default' if it doesn't have that - attribute.""" - return self._getAttrMap().get(key, default) - - def has_key(self, key): - return self._getAttrMap().has_key(key) - - def __getitem__(self, key): - """tag[key] returns the value of the 'key' attribute for the tag, - and throws an exception if it's not there.""" - return self._getAttrMap()[key] - - def __iter__(self): - "Iterating over a tag iterates over its contents." - return iter(self.contents) - - def __len__(self): - "The length of a tag is the length of its list of contents." - return len(self.contents) - - def __contains__(self, x): - return x in self.contents - - def __nonzero__(self): - "A tag is non-None even if it has no contents." - return True - - def __setitem__(self, key, value): - """Setting tag[key] sets the value of the 'key' attribute for the - tag.""" - self._getAttrMap() - self.attrMap[key] = value - found = False - for i in range(0, len(self.attrs)): - if self.attrs[i][0] == key: - self.attrs[i] = (key, value) - found = True - if not found: - self.attrs.append((key, value)) - self._getAttrMap()[key] = value - - def __delitem__(self, key): - "Deleting tag[key] deletes all 'key' attributes for the tag." - for item in self.attrs: - if item[0] == key: - self.attrs.remove(item) - #We don't break because bad HTML can define the same - #attribute multiple times. - self._getAttrMap() - if self.attrMap.has_key(key): - del self.attrMap[key] - - def __call__(self, *args, **kwargs): - """Calling a tag like a function is the same as calling its - findAll() method. Eg. tag('a') returns a list of all the A tags - found within this tag.""" - return apply(self.findAll, args, kwargs) - - def __getattr__(self, tag): - #print "Getattr %s.%s" % (self.__class__, tag) - if len(tag) > 3 and tag.rfind('Tag') == len(tag)-3: - return self.find(tag[:-3]) - elif tag.find('__') != 0: - return self.find(tag) - raise AttributeError, "'%s' object has no attribute '%s'" % (self.__class__, tag) - - def __eq__(self, other): - """Returns true iff this tag has the same name, the same attributes, - and the same contents (recursively) as the given tag. - - NOTE: right now this will return false if two tags have the - same attributes in a different order. Should this be fixed?""" - if not hasattr(other, 'name') or not hasattr(other, 'attrs') or not hasattr(other, 'contents') or self.name != other.name or self.attrs != other.attrs or len(self) != len(other): - return False - for i in range(0, len(self.contents)): - if self.contents[i] != other.contents[i]: - return False - return True - - def __ne__(self, other): - """Returns true iff this tag is not identical to the other tag, - as defined in __eq__.""" - return not self == other - - def __repr__(self, encoding=DEFAULT_OUTPUT_ENCODING): - """Renders this tag as a string.""" - return self.decode(eventualEncoding=encoding) - - BARE_AMPERSAND_OR_BRACKET = re.compile("([<>]|" - + "&(?!#\d+;|#x[0-9a-fA-F]+;|\w+;)" - + ")") - - def _sub_entity(self, x): - """Used with a regular expression to substitute the - appropriate XML entity for an XML special character.""" - return "&" + self.XML_SPECIAL_CHARS_TO_ENTITIES[x.group(0)[0]] + ";" - - def __unicode__(self): - return self.decode() - - def __str__(self): - return self.encode() - - def encode(self, encoding=DEFAULT_OUTPUT_ENCODING, - prettyPrint=False, indentLevel=0): - return self.decode(prettyPrint, indentLevel, encoding).encode(encoding) - - def decode(self, prettyPrint=False, indentLevel=0, - eventualEncoding=DEFAULT_OUTPUT_ENCODING): - """Returns a string or Unicode representation of this tag and - its contents. To get Unicode, pass None for encoding.""" - - attrs = [] - if self.attrs: - for key, val in self.attrs: - fmt = '%s="%s"' - if isString(val): - if (self.containsSubstitutions - and eventualEncoding is not None - and '%SOUP-ENCODING%' in val): - val = self.substituteEncoding(val, eventualEncoding) - - # The attribute value either: - # - # * Contains no embedded double quotes or single quotes. - # No problem: we enclose it in double quotes. - # * Contains embedded single quotes. No problem: - # double quotes work here too. - # * Contains embedded double quotes. No problem: - # we enclose it in single quotes. - # * Embeds both single _and_ double quotes. This - # can't happen naturally, but it can happen if - # you modify an attribute value after parsing - # the document. Now we have a bit of a - # problem. We solve it by enclosing the - # attribute in single quotes, and escaping any - # embedded single quotes to XML entities. - if '"' in val: - fmt = "%s='%s'" - if "'" in val: - # TODO: replace with apos when - # appropriate. - val = val.replace("'", "&squot;") - - # Now we're okay w/r/t quotes. But the attribute - # value might also contain angle brackets, or - # ampersands that aren't part of entities. We need - # to escape those to XML entities too. - val = self.BARE_AMPERSAND_OR_BRACKET.sub(self._sub_entity, val) - if val is None: - # Handle boolean attributes. - decoded = key - else: - decoded = fmt % (key, val) - attrs.append(decoded) - close = '' - closeTag = '' - if self.isSelfClosing: - close = ' /' - else: - closeTag = '</%s>' % self.name - - indentTag, indentContents = 0, 0 - if prettyPrint: - indentTag = indentLevel - space = (' ' * (indentTag-1)) - indentContents = indentTag + 1 - contents = self.decodeContents(prettyPrint, indentContents, - eventualEncoding) - if self.hidden: - s = contents - else: - s = [] - attributeString = '' - if attrs: - attributeString = ' ' + ' '.join(attrs) - if prettyPrint: - s.append(space) - s.append('<%s%s%s>' % (self.name, attributeString, close)) - if prettyPrint: - s.append("\n") - s.append(contents) - if prettyPrint and contents and contents[-1] != "\n": - s.append("\n") - if prettyPrint and closeTag: - s.append(space) - s.append(closeTag) - if prettyPrint and closeTag and self.nextSibling: - s.append("\n") - s = ''.join(s) - return s - - def decompose(self): - """Recursively destroys the contents of this tree.""" - contents = [i for i in self.contents] - for i in contents: - if isinstance(i, Tag): - i.decompose() - else: - i.extract() - self.extract() - - def prettify(self, encoding=DEFAULT_OUTPUT_ENCODING): - return self.encode(encoding, True) - - def encodeContents(self, encoding=DEFAULT_OUTPUT_ENCODING, - prettyPrint=False, indentLevel=0): - return self.decodeContents(prettyPrint, indentLevel).encode(encoding) - - def decodeContents(self, prettyPrint=False, indentLevel=0, - eventualEncoding=DEFAULT_OUTPUT_ENCODING): - """Renders the contents of this tag as a string in the given - encoding. If encoding is None, returns a Unicode string..""" - s=[] - for c in self: - text = None - if isinstance(c, NavigableString): - text = c.decodeGivenEventualEncoding(eventualEncoding) - elif isinstance(c, Tag): - s.append(c.decode(prettyPrint, indentLevel, eventualEncoding)) - if text and prettyPrint: - text = text.strip() - if text: - if prettyPrint: - s.append(" " * (indentLevel-1)) - s.append(text) - if prettyPrint: - s.append("\n") - return ''.join(s) - - #Soup methods - - def find(self, name=None, attrs={}, recursive=True, text=None, - **kwargs): - """Return only the first child of this Tag matching the given - criteria.""" - r = None - l = self.findAll(name, attrs, recursive, text, 1, **kwargs) - if l: - r = l[0] - return r - findChild = find - - def findAll(self, name=None, attrs={}, recursive=True, text=None, - limit=None, **kwargs): - """Extracts a list of Tag objects that match the given - criteria. You can specify the name of the Tag and any - attributes you want the Tag to have. - - The value of a key-value pair in the 'attrs' map can be a - string, a list of strings, a regular expression object, or a - callable that takes a string and returns whether or not the - string matches for some custom definition of 'matches'. The - same is true of the tag name.""" - generator = self.recursiveChildGenerator - if not recursive: - generator = self.childGenerator - return self._findAll(name, attrs, text, limit, generator, **kwargs) - findChildren = findAll - - # Pre-3.x compatibility methods. Will go away in 4.0. - first = find - fetch = findAll - - def fetchText(self, text=None, recursive=True, limit=None): - return self.findAll(text=text, recursive=recursive, limit=limit) - - def firstText(self, text=None, recursive=True): - return self.find(text=text, recursive=recursive) - - # 3.x compatibility methods. Will go away in 4.0. - def renderContents(self, encoding=DEFAULT_OUTPUT_ENCODING, - prettyPrint=False, indentLevel=0): - if encoding is None: - return self.decodeContents(prettyPrint, indentLevel, encoding) - else: - return self.encodeContents(encoding, prettyPrint, indentLevel) - - - #Private methods - - def _getAttrMap(self): - """Initializes a map representation of this tag's attributes, - if not already initialized.""" - if not getattr(self, 'attrMap'): - self.attrMap = {} - for (key, value) in self.attrs: - self.attrMap[key] = value - return self.attrMap - - #Generator methods - def recursiveChildGenerator(self): - if not len(self.contents): - raise StopIteration - stopNode = self._lastRecursiveChild().next - current = self.contents[0] - while current is not stopNode: - yield current - current = current.next - - def childGenerator(self): - if not len(self.contents): - raise StopIteration - current = self.contents[0] - while current: - yield current - current = current.nextSibling - raise StopIteration - -# Next, a couple classes to represent queries and their results. -class SoupStrainer: - """Encapsulates a number of ways of matching a markup element (tag or - text).""" - - def __init__(self, name=None, attrs={}, text=None, **kwargs): - self.name = name - if isString(attrs): - kwargs['class'] = attrs - attrs = None - if kwargs: - if attrs: - attrs = attrs.copy() - attrs.update(kwargs) - else: - attrs = kwargs - self.attrs = attrs - self.text = text - - def __str__(self): - if self.text: - return self.text - else: - return "%s|%s" % (self.name, self.attrs) - - def searchTag(self, markupName=None, markupAttrs={}): - found = None - markup = None - if isinstance(markupName, Tag): - markup = markupName - markupAttrs = markup - callFunctionWithTagData = callable(self.name) \ - and not isinstance(markupName, Tag) - - if (not self.name) \ - or callFunctionWithTagData \ - or (markup and self._matches(markup, self.name)) \ - or (not markup and self._matches(markupName, self.name)): - if callFunctionWithTagData: - match = self.name(markupName, markupAttrs) - else: - match = True - markupAttrMap = None - for attr, matchAgainst in self.attrs.items(): - if not markupAttrMap: - if hasattr(markupAttrs, 'get'): - markupAttrMap = markupAttrs - else: - markupAttrMap = {} - for k,v in markupAttrs: - markupAttrMap[k] = v - attrValue = markupAttrMap.get(attr) - if not self._matches(attrValue, matchAgainst): - match = False - break - if match: - if markup: - found = markup - else: - found = markupName - return found - - def search(self, markup): - #print 'looking for %s in %s' % (self, markup) - found = None - # If given a list of items, scan it for a text element that - # matches. - if isList(markup) and not isinstance(markup, Tag): - for element in markup: - if isinstance(element, NavigableString) \ - and self.search(element): - found = element - break - # If it's a Tag, make sure its name or attributes match. - # Don't bother with Tags if we're searching for text. - elif isinstance(markup, Tag): - if not self.text: - found = self.searchTag(markup) - # If it's text, make sure the text matches. - elif isinstance(markup, NavigableString) or \ - isString(markup): - if self._matches(markup, self.text): - found = markup - else: - raise Exception, "I don't know how to match against a %s" \ - % markup.__class__ - return found - - def _matches(self, markup, matchAgainst): - #print "Matching %s against %s" % (markup, matchAgainst) - result = False - if matchAgainst == True and type(matchAgainst) == types.BooleanType: - result = markup != None - elif callable(matchAgainst): - result = matchAgainst(markup) - else: - #Custom match methods take the tag as an argument, but all - #other ways of matching match the tag name as a string. - if isinstance(markup, Tag): - markup = markup.name - if markup is not None and not isString(markup): - markup = unicode(markup) - #Now we know that chunk is either a string, or None. - if hasattr(matchAgainst, 'match'): - # It's a regexp object. - result = markup and matchAgainst.search(markup) - elif (isList(matchAgainst) - and (markup is not None or not isString(matchAgainst))): - result = markup in matchAgainst - elif hasattr(matchAgainst, 'items'): - result = markup.has_key(matchAgainst) - elif matchAgainst and isString(markup): - if isinstance(markup, unicode): - matchAgainst = unicode(matchAgainst) - else: - matchAgainst = str(matchAgainst) - - if not result: - result = matchAgainst == markup - return result - -class ResultSet(list): - """A ResultSet is just a list that keeps track of the SoupStrainer - that created it.""" - def __init__(self, source): - list.__init__([]) - self.source = source - -# Now, some helper functions. - -def isList(l): - """Convenience method that works with all 2.x versions of Python - to determine whether or not something is listlike.""" - return ((hasattr(l, '__iter__') and not isString(l)) - or (type(l) in (types.ListType, types.TupleType))) - -def isString(s): - """Convenience method that works with all 2.x versions of Python - to determine whether or not something is stringlike.""" - try: - return isinstance(s, unicode) or isinstance(s, basestring) - except NameError: - return isinstance(s, str) - -def buildTagMap(default, *args): - """Turns a list of maps, lists, or scalars into a single map. - Used to build the SELF_CLOSING_TAGS, NESTABLE_TAGS, and - NESTING_RESET_TAGS maps out of lists and partial maps.""" - built = {} - for portion in args: - if hasattr(portion, 'items'): - #It's a map. Merge it. - for k,v in portion.items(): - built[k] = v - elif isList(portion) and not isString(portion): - #It's a list. Map each item to the default. - for k in portion: - built[k] = default - else: - #It's a scalar. Map it to the default. - built[portion] = default - return built - -# Now, the parser classes. - -class HTMLParserBuilder(HTMLParser): - - def __init__(self, soup): - HTMLParser.__init__(self) - self.soup = soup - - # We inherit feed() and reset(). - - def handle_starttag(self, name, attrs): - if name == 'meta': - self.soup.extractCharsetFromMeta(attrs) - else: - self.soup.unknown_starttag(name, attrs) - - def handle_endtag(self, name): - self.soup.unknown_endtag(name) - - def handle_data(self, content): - self.soup.handle_data(content) - - def _toStringSubclass(self, text, subclass): - """Adds a certain piece of text to the tree as a NavigableString - subclass.""" - self.soup.endData() - self.handle_data(text) - self.soup.endData(subclass) - - def handle_pi(self, text): - """Handle a processing instruction as a ProcessingInstruction - object, possibly one with a %SOUP-ENCODING% slot into which an - encoding will be plugged later.""" - if text[:3] == "xml": - text = u"xml version='1.0' encoding='%SOUP-ENCODING%'" - self._toStringSubclass(text, ProcessingInstruction) - - def handle_comment(self, text): - "Handle comments as Comment objects." - self._toStringSubclass(text, Comment) - - def handle_charref(self, ref): - "Handle character references as data." - if self.soup.convertEntities: - data = unichr(int(ref)) - else: - data = '&#%s;' % ref - self.handle_data(data) - - def handle_entityref(self, ref): - """Handle entity references as data, possibly converting known - HTML and/or XML entity references to the corresponding Unicode - characters.""" - data = None - if self.soup.convertHTMLEntities: - try: - data = unichr(name2codepoint[ref]) - except KeyError: - pass - - if not data and self.soup.convertXMLEntities: - data = self.soup.XML_ENTITIES_TO_SPECIAL_CHARS.get(ref) - - if not data and self.soup.convertHTMLEntities and \ - not self.soup.XML_ENTITIES_TO_SPECIAL_CHARS.get(ref): - # TODO: We've got a problem here. We're told this is - # an entity reference, but it's not an XML entity - # reference or an HTML entity reference. Nonetheless, - # the logical thing to do is to pass it through as an - # unrecognized entity reference. - # - # Except: when the input is "&carol;" this function - # will be called with input "carol". When the input is - # "AT&T", this function will be called with input - # "T". We have no way of knowing whether a semicolon - # was present originally, so we don't know whether - # this is an unknown entity or just a misplaced - # ampersand. - # - # The more common case is a misplaced ampersand, so I - # escape the ampersand and omit the trailing semicolon. - data = "&%s" % ref - if not data: - # This case is different from the one above, because we - # haven't already gone through a supposedly comprehensive - # mapping of entities to Unicode characters. We might not - # have gone through any mapping at all. So the chances are - # very high that this is a real entity, and not a - # misplaced ampersand. - data = "&%s;" % ref - self.handle_data(data) - - def handle_decl(self, data): - "Handle DOCTYPEs and the like as Declaration objects." - self._toStringSubclass(data, Declaration) - - def parse_declaration(self, i): - """Treat a bogus SGML declaration as raw data. Treat a CDATA - declaration as a CData object.""" - j = None - if self.rawdata[i:i+9] == '<![CDATA[': - k = self.rawdata.find(']]>', i) - if k == -1: - k = len(self.rawdata) - data = self.rawdata[i+9:k] - j = k+3 - self._toStringSubclass(data, CData) - else: - try: - j = HTMLParser.parse_declaration(self, i) - except HTMLParseError: - toHandle = self.rawdata[i:] - self.handle_data(toHandle) - j = i + len(toHandle) - return j - - -class BeautifulStoneSoup(Tag): - - """This class contains the basic parser and search code. It defines - a parser that knows nothing about tag behavior except for the - following: - - You can't close a tag without closing all the tags it encloses. - That is, "<foo><bar></foo>" actually means - "<foo><bar></bar></foo>". - - [Another possible explanation is "<foo><bar /></foo>", but since - this class defines no SELF_CLOSING_TAGS, it will never use that - explanation.] - - This class is useful for parsing XML or made-up markup languages, - or when BeautifulSoup makes an assumption counter to what you were - expecting.""" - - SELF_CLOSING_TAGS = {} - NESTABLE_TAGS = {} - RESET_NESTING_TAGS = {} - QUOTE_TAGS = {} - PRESERVE_WHITESPACE_TAGS = [] - - MARKUP_MASSAGE = [(re.compile('(<[^<>]*)/>'), - lambda x: x.group(1) + ' />'), - (re.compile('<!\s+([^<>]*)>'), - lambda x: '<!' + x.group(1) + '>') - ] - - ROOT_TAG_NAME = u'[document]' - - HTML_ENTITIES = "html" - XML_ENTITIES = "xml" - XHTML_ENTITIES = "xhtml" - # TODO: This only exists for backwards-compatibility - ALL_ENTITIES = XHTML_ENTITIES - - # Used when determining whether a text node is all whitespace and - # can be replaced with a single space. A text node that contains - # fancy Unicode spaces (usually non-breaking) should be left - # alone. - STRIP_ASCII_SPACES = { 9: None, 10: None, 12: None, 13: None, 32: None, } - - def __init__(self, markup="", parseOnlyThese=None, fromEncoding=None, - markupMassage=True, smartQuotesTo=XML_ENTITIES, - convertEntities=None, selfClosingTags=None, isHTML=False, - builder=HTMLParserBuilder): - """The Soup object is initialized as the 'root tag', and the - provided markup (which can be a string or a file-like object) - is fed into the underlying parser. - - HTMLParser will process most bad HTML, and the BeautifulSoup - class has some tricks for dealing with some HTML that kills - HTMLParser, but Beautiful Soup can nonetheless choke or lose data - if your data uses self-closing tags or declarations - incorrectly. - - By default, Beautiful Soup uses regexes to sanitize input, - avoiding the vast majority of these problems. If the problems - don't apply to you, pass in False for markupMassage, and - you'll get better performance. - - The default parser massage techniques fix the two most common - instances of invalid HTML that choke HTMLParser: - - <br/> (No space between name of closing tag and tag close) - <! --Comment--> (Extraneous whitespace in declaration) - - You can pass in a custom list of (RE object, replace method) - tuples to get Beautiful Soup to scrub your input the way you - want.""" - - self.parseOnlyThese = parseOnlyThese - self.fromEncoding = fromEncoding - self.smartQuotesTo = smartQuotesTo - self.convertEntities = convertEntities - # Set the rules for how we'll deal with the entities we - # encounter - if self.convertEntities: - # It doesn't make sense to convert encoded characters to - # entities even while you're converting entities to Unicode. - # Just convert it all to Unicode. - self.smartQuotesTo = None - if convertEntities == self.HTML_ENTITIES: - self.convertXMLEntities = False - self.convertHTMLEntities = True - self.escapeUnrecognizedEntities = True - elif convertEntities == self.XHTML_ENTITIES: - self.convertXMLEntities = True - self.convertHTMLEntities = True - self.escapeUnrecognizedEntities = False - elif convertEntities == self.XML_ENTITIES: - self.convertXMLEntities = True - self.convertHTMLEntities = False - self.escapeUnrecognizedEntities = False - else: - self.convertXMLEntities = False - self.convertHTMLEntities = False - self.escapeUnrecognizedEntities = False - - self.instanceSelfClosingTags = buildTagMap(None, selfClosingTags) - self.builder = builder(self) - self.reset() - - if hasattr(markup, 'read'): # It's a file-type object. - markup = markup.read() - self.markup = markup - self.markupMassage = markupMassage - try: - self._feed(isHTML=isHTML) - except StopParsing: - pass - self.markup = None # The markup can now be GCed. - self.builder = None # So can the builder. - - def _feed(self, inDocumentEncoding=None, isHTML=False): - # Convert the document to Unicode. - markup = self.markup - if isinstance(markup, unicode): - if not hasattr(self, 'originalEncoding'): - self.originalEncoding = None - else: - dammit = UnicodeDammit\ - (markup, [self.fromEncoding, inDocumentEncoding], - smartQuotesTo=self.smartQuotesTo, isHTML=isHTML) - markup = dammit.unicode - self.originalEncoding = dammit.originalEncoding - self.declaredHTMLEncoding = dammit.declaredHTMLEncoding - if markup: - if self.markupMassage: - if not isList(self.markupMassage): - self.markupMassage = self.MARKUP_MASSAGE - for fix, m in self.markupMassage: - markup = fix.sub(m, markup) - # TODO: We get rid of markupMassage so that the - # soup object can be deepcopied later on. Some - # Python installations can't copy regexes. If anyone - # was relying on the existence of markupMassage, this - # might cause problems. - del(self.markupMassage) - self.builder.reset() - - self.builder.feed(markup) - # Close out any unfinished strings and close all the open tags. - self.endData() - while self.currentTag.name != self.ROOT_TAG_NAME: - self.popTag() - - def isSelfClosingTag(self, name): - """Returns true iff the given string is the name of a - self-closing tag according to this parser.""" - return self.SELF_CLOSING_TAGS.has_key(name) \ - or self.instanceSelfClosingTags.has_key(name) - - def reset(self): - Tag.__init__(self, self, self.ROOT_TAG_NAME) - self.hidden = 1 - self.builder.reset() - self.currentData = [] - self.currentTag = None - self.tagStack = [] - self.quoteStack = [] - self.pushTag(self) - - def popTag(self): - tag = self.tagStack.pop() - # Tags with just one string-owning child get the child as a - # 'string' property, so that soup.tag.string is shorthand for - # soup.tag.contents[0] - if len(self.currentTag.contents) == 1 and \ - isinstance(self.currentTag.contents[0], NavigableString): - self.currentTag.string = self.currentTag.contents[0] - - #print "Pop", tag.name - if self.tagStack: - self.currentTag = self.tagStack[-1] - return self.currentTag - - def pushTag(self, tag): - #print "Push", tag.name - if self.currentTag: - self.currentTag.contents.append(tag) - self.tagStack.append(tag) - self.currentTag = self.tagStack[-1] - - def endData(self, containerClass=NavigableString): - if self.currentData: - currentData = u''.join(self.currentData) - if (currentData.translate(self.STRIP_ASCII_SPACES) == '' and - not set([tag.name for tag in self.tagStack]).intersection( - self.PRESERVE_WHITESPACE_TAGS)): - if '\n' in currentData: - currentData = '\n' - else: - currentData = ' ' - self.currentData = [] - if self.parseOnlyThese and len(self.tagStack) <= 1 and \ - (not self.parseOnlyThese.text or \ - not self.parseOnlyThese.search(currentData)): - return - o = containerClass(currentData) - o.setup(self.currentTag, self.previous) - if self.previous: - self.previous.next = o - self.previous = o - self.currentTag.contents.append(o) - - - def _popToTag(self, name, inclusivePop=True): - """Pops the tag stack up to and including the most recent - instance of the given tag. If inclusivePop is false, pops the tag - stack up to but *not* including the most recent instqance of - the given tag.""" - #print "Popping to %s" % name - if name == self.ROOT_TAG_NAME: - return - - numPops = 0 - mostRecentTag = None - for i in range(len(self.tagStack)-1, 0, -1): - if name == self.tagStack[i].name: - numPops = len(self.tagStack)-i - break - if not inclusivePop: - numPops = numPops - 1 - - for i in range(0, numPops): - mostRecentTag = self.popTag() - return mostRecentTag - - def _smartPop(self, name): - - """We need to pop up to the previous tag of this type, unless - one of this tag's nesting reset triggers comes between this - tag and the previous tag of this type, OR unless this tag is a - generic nesting trigger and another generic nesting trigger - comes between this tag and the previous tag of this type. - - Examples: - <p>Foo<b>Bar *<p>* should pop to 'p', not 'b'. - <p>Foo<table>Bar *<p>* should pop to 'table', not 'p'. - <p>Foo<table><tr>Bar *<p>* should pop to 'tr', not 'p'. - - <li><ul><li> *<li>* should pop to 'ul', not the first 'li'. - <tr><table><tr> *<tr>* should pop to 'table', not the first 'tr' - <td><tr><td> *<td>* should pop to 'tr', not the first 'td' - """ - - nestingResetTriggers = self.NESTABLE_TAGS.get(name) - isNestable = nestingResetTriggers != None - isResetNesting = self.RESET_NESTING_TAGS.has_key(name) - popTo = None - inclusive = True - for i in range(len(self.tagStack)-1, 0, -1): - p = self.tagStack[i] - if (not p or p.name == name) and not isNestable: - #Non-nestable tags get popped to the top or to their - #last occurance. - popTo = name - break - if (nestingResetTriggers != None - and p.name in nestingResetTriggers) \ - or (nestingResetTriggers == None and isResetNesting - and self.RESET_NESTING_TAGS.has_key(p.name)): - - #If we encounter one of the nesting reset triggers - #peculiar to this tag, or we encounter another tag - #that causes nesting to reset, pop up to but not - #including that tag. - popTo = p.name - inclusive = False - break - p = p.parent - if popTo: - self._popToTag(popTo, inclusive) - - def unknown_starttag(self, name, attrs, selfClosing=0): - #print "Start tag %s: %s" % (name, attrs) - if self.quoteStack: - #This is not a real tag. - #print "<%s> is not real!" % name - attrs = ''.join(map(lambda(x, y): ' %s="%s"' % (x, y), attrs)) - self.handle_data('<%s%s>' % (name, attrs)) - return - self.endData() - - if not self.isSelfClosingTag(name) and not selfClosing: - self._smartPop(name) - - if self.parseOnlyThese and len(self.tagStack) <= 1 \ - and (self.parseOnlyThese.text or not self.parseOnlyThese.searchTag(name, attrs)): - return - - tag = Tag(self, name, attrs, self.currentTag, self.previous) - if self.previous: - self.previous.next = tag - self.previous = tag - self.pushTag(tag) - if selfClosing or self.isSelfClosingTag(name): - self.popTag() - if name in self.QUOTE_TAGS: - #print "Beginning quote (%s)" % name - self.quoteStack.append(name) - self.literal = 1 - return tag - - def unknown_endtag(self, name): - #print "End tag %s" % name - if self.quoteStack and self.quoteStack[-1] != name: - #This is not a real end tag. - #print "</%s> is not real!" % name - self.handle_data('</%s>' % name) - return - self.endData() - self._popToTag(name) - if self.quoteStack and self.quoteStack[-1] == name: - self.quoteStack.pop() - self.literal = (len(self.quoteStack) > 0) - - def handle_data(self, data): - self.currentData.append(data) - - def extractCharsetFromMeta(self, attrs): - self.unknown_starttag('meta', attrs) - - -class BeautifulSoup(BeautifulStoneSoup): - - """This parser knows the following facts about HTML: - - * Some tags have no closing tag and should be interpreted as being - closed as soon as they are encountered. - - * The text inside some tags (ie. 'script') may contain tags which - are not really part of the document and which should be parsed - as text, not tags. If you want to parse the text as tags, you can - always fetch it and parse it explicitly. - - * Tag nesting rules: - - Most tags can't be nested at all. For instance, the occurance of - a <p> tag should implicitly close the previous <p> tag. - - <p>Para1<p>Para2 - should be transformed into: - <p>Para1</p><p>Para2 - - Some tags can be nested arbitrarily. For instance, the occurance - of a <blockquote> tag should _not_ implicitly close the previous - <blockquote> tag. - - Alice said: <blockquote>Bob said: <blockquote>Blah - should NOT be transformed into: - Alice said: <blockquote>Bob said: </blockquote><blockquote>Blah - - Some tags can be nested, but the nesting is reset by the - interposition of other tags. For instance, a <tr> tag should - implicitly close the previous <tr> tag within the same <table>, - but not close a <tr> tag in another table. - - <table><tr>Blah<tr>Blah - should be transformed into: - <table><tr>Blah</tr><tr>Blah - but, - <tr>Blah<table><tr>Blah - should NOT be transformed into - <tr>Blah<table></tr><tr>Blah - - Differing assumptions about tag nesting rules are a major source - of problems with the BeautifulSoup class. If BeautifulSoup is not - treating as nestable a tag your page author treats as nestable, - try ICantBelieveItsBeautifulSoup, MinimalSoup, or - BeautifulStoneSoup before writing your own subclass.""" - - def __init__(self, *args, **kwargs): - if not kwargs.has_key('smartQuotesTo'): - kwargs['smartQuotesTo'] = self.HTML_ENTITIES - kwargs['isHTML'] = True - BeautifulStoneSoup.__init__(self, *args, **kwargs) - - SELF_CLOSING_TAGS = buildTagMap(None, - ['br' , 'hr', 'input', 'img', 'meta', - 'spacer', 'link', 'frame', 'base']) - - PRESERVE_WHITESPACE_TAGS = set(['pre', 'textarea']) - - QUOTE_TAGS = {'script' : None, 'textarea' : None} - - #According to the HTML standard, each of these inline tags can - #contain another tag of the same type. Furthermore, it's common - #to actually use these tags this way. - NESTABLE_INLINE_TAGS = ['span', 'font', 'q', 'object', 'bdo', 'sub', 'sup', - 'center'] - - #According to the HTML standard, these block tags can contain - #another tag of the same type. Furthermore, it's common - #to actually use these tags this way. - NESTABLE_BLOCK_TAGS = ['blockquote', 'div', 'fieldset', 'ins', 'del'] - - #Lists can contain other lists, but there are restrictions. - NESTABLE_LIST_TAGS = { 'ol' : [], - 'ul' : [], - 'li' : ['ul', 'ol'], - 'dl' : [], - 'dd' : ['dl'], - 'dt' : ['dl'] } - - #Tables can contain other tables, but there are restrictions. - NESTABLE_TABLE_TAGS = {'table' : [], - 'tr' : ['table', 'tbody', 'tfoot', 'thead'], - 'td' : ['tr'], - 'th' : ['tr'], - 'thead' : ['table'], - 'tbody' : ['table'], - 'tfoot' : ['table'], - } - - NON_NESTABLE_BLOCK_TAGS = ['address', 'form', 'p', 'pre'] - - #If one of these tags is encountered, all tags up to the next tag of - #this type are popped. - RESET_NESTING_TAGS = buildTagMap(None, NESTABLE_BLOCK_TAGS, 'noscript', - NON_NESTABLE_BLOCK_TAGS, - NESTABLE_LIST_TAGS, - NESTABLE_TABLE_TAGS) - - NESTABLE_TAGS = buildTagMap([], NESTABLE_INLINE_TAGS, NESTABLE_BLOCK_TAGS, - NESTABLE_LIST_TAGS, NESTABLE_TABLE_TAGS) - - # Used to detect the charset in a META tag; see start_meta - CHARSET_RE = re.compile("((^|;)\s*charset=)([^;]*)", re.M) - - def extractCharsetFromMeta(self, attrs): - """Beautiful Soup can detect a charset included in a META tag, - try to convert the document to that charset, and re-parse the - document from the beginning.""" - httpEquiv = None - contentType = None - contentTypeIndex = None - tagNeedsEncodingSubstitution = False - - for i in range(0, len(attrs)): - key, value = attrs[i] - key = key.lower() - if key == 'http-equiv': - httpEquiv = value - elif key == 'content': - contentType = value - contentTypeIndex = i - - if httpEquiv and contentType: # It's an interesting meta tag. - match = self.CHARSET_RE.search(contentType) - if match: - if (self.declaredHTMLEncoding is not None or - self.originalEncoding == self.fromEncoding): - # An HTML encoding was sniffed while converting - # the document to Unicode, or an HTML encoding was - # sniffed during a previous pass through the - # document, or an encoding was specified - # explicitly and it worked. Rewrite the meta tag. - def rewrite(match): - return match.group(1) + "%SOUP-ENCODING%" - newAttr = self.CHARSET_RE.sub(rewrite, contentType) - attrs[contentTypeIndex] = (attrs[contentTypeIndex][0], - newAttr) - tagNeedsEncodingSubstitution = True - else: - # This is our first pass through the document. - # Go through it again with the encoding information. - newCharset = match.group(3) - if newCharset and newCharset != self.originalEncoding: - self.declaredHTMLEncoding = newCharset - self._feed(self.declaredHTMLEncoding) - raise StopParsing - pass - tag = self.unknown_starttag("meta", attrs) - if tag and tagNeedsEncodingSubstitution: - tag.containsSubstitutions = True - - -class StopParsing(Exception): - pass - -class ICantBelieveItsBeautifulSoup(BeautifulSoup): - - """The BeautifulSoup class is oriented towards skipping over - common HTML errors like unclosed tags. However, sometimes it makes - errors of its own. For instance, consider this fragment: - - <b>Foo<b>Bar</b></b> - - This is perfectly valid (if bizarre) HTML. However, the - BeautifulSoup class will implicitly close the first b tag when it - encounters the second 'b'. It will think the author wrote - "<b>Foo<b>Bar", and didn't close the first 'b' tag, because - there's no real-world reason to bold something that's already - bold. When it encounters '</b></b>' it will close two more 'b' - tags, for a grand total of three tags closed instead of two. This - can throw off the rest of your document structure. The same is - true of a number of other tags, listed below. - - It's much more common for someone to forget to close a 'b' tag - than to actually use nested 'b' tags, and the BeautifulSoup class - handles the common case. This class handles the not-co-common - case: where you can't believe someone wrote what they did, but - it's valid HTML and BeautifulSoup screwed up by assuming it - wouldn't be.""" - - I_CANT_BELIEVE_THEYRE_NESTABLE_INLINE_TAGS = \ - ['em', 'big', 'i', 'small', 'tt', 'abbr', 'acronym', 'strong', - 'cite', 'code', 'dfn', 'kbd', 'samp', 'strong', 'var', 'b', - 'big'] - - I_CANT_BELIEVE_THEYRE_NESTABLE_BLOCK_TAGS = ['noscript'] - - NESTABLE_TAGS = buildTagMap([], BeautifulSoup.NESTABLE_TAGS, - I_CANT_BELIEVE_THEYRE_NESTABLE_BLOCK_TAGS, - I_CANT_BELIEVE_THEYRE_NESTABLE_INLINE_TAGS) - -class MinimalSoup(BeautifulSoup): - """The MinimalSoup class is for parsing HTML that contains - pathologically bad markup. It makes no assumptions about tag - nesting, but it does know which tags are self-closing, that - <script> tags contain Javascript and should not be parsed, that - META tags may contain encoding information, and so on. - - This also makes it better for subclassing than BeautifulStoneSoup - or BeautifulSoup.""" - - RESET_NESTING_TAGS = buildTagMap('noscript') - NESTABLE_TAGS = {} - -class BeautifulSOAP(BeautifulStoneSoup): - """This class will push a tag with only a single string child into - the tag's parent as an attribute. The attribute's name is the tag - name, and the value is the string child. An example should give - the flavor of the change: - - <foo><bar>baz</bar></foo> - => - <foo bar="baz"><bar>baz</bar></foo> - - You can then access fooTag['bar'] instead of fooTag.barTag.string. - - This is, of course, useful for scraping structures that tend to - use subelements instead of attributes, such as SOAP messages. Note - that it modifies its input, so don't print the modified version - out. - - I'm not sure how many people really want to use this class; let me - know if you do. Mainly I like the name.""" - - def popTag(self): - if len(self.tagStack) > 1: - tag = self.tagStack[-1] - parent = self.tagStack[-2] - parent._getAttrMap() - if (isinstance(tag, Tag) and len(tag.contents) == 1 and - isinstance(tag.contents[0], NavigableString) and - not parent.attrMap.has_key(tag.name)): - parent[tag.name] = tag.contents[0] - BeautifulStoneSoup.popTag(self) - -#Enterprise class names! It has come to our attention that some people -#think the names of the Beautiful Soup parser classes are too silly -#and "unprofessional" for use in enterprise screen-scraping. We feel -#your pain! For such-minded folk, the Beautiful Soup Consortium And -#All-Night Kosher Bakery recommends renaming this file to -#"RobustParser.py" (or, in cases of extreme enterprisiness, -#"RobustParserBeanInterface.class") and using the following -#enterprise-friendly class aliases: -class RobustXMLParser(BeautifulStoneSoup): - pass -class RobustHTMLParser(BeautifulSoup): - pass -class RobustWackAssHTMLParser(ICantBelieveItsBeautifulSoup): - pass -class RobustInsanelyWackAssHTMLParser(MinimalSoup): - pass -class SimplifyingSOAPParser(BeautifulSOAP): - pass - -###################################################### -# -# Bonus library: Unicode, Dammit -# -# This class forces XML data into a standard format (usually to UTF-8 -# or Unicode). It is heavily based on code from Mark Pilgrim's -# Universal Feed Parser. It does not rewrite the XML or HTML to -# reflect a new encoding: that happens in BeautifulStoneSoup.handle_pi -# (XML) and BeautifulSoup.start_meta (HTML). - -# Autodetects character encodings. -# Download from http://chardet.feedparser.org/ -try: - import chardet -# import chardet.constants -# chardet.constants._debug = 1 -except ImportError: - chardet = None - -# cjkcodecs and iconv_codec make Python know about more character encodings. -# Both are available from http://cjkpython.i18n.org/ -# They're built in if you use Python 2.4. -try: - import cjkcodecs.aliases -except ImportError: - pass -try: - import iconv_codec -except ImportError: - pass - -class UnicodeDammit: - """A class for detecting the encoding of a *ML document and - converting it to a Unicode string. If the source encoding is - windows-1252, can replace MS smart quotes with their HTML or XML - equivalents.""" - - # This dictionary maps commonly seen values for "charset" in HTML - # meta tags to the corresponding Python codec names. It only covers - # values that aren't in Python's aliases and can't be determined - # by the heuristics in find_codec. - CHARSET_ALIASES = { "macintosh" : "mac-roman", - "x-sjis" : "shift-jis" } - - def __init__(self, markup, overrideEncodings=[], - smartQuotesTo='xml', isHTML=False): - self.declaredHTMLEncoding = None - self.markup, documentEncoding, sniffedEncoding = \ - self._detectEncoding(markup, isHTML) - self.smartQuotesTo = smartQuotesTo - self.triedEncodings = [] - if markup == '' or isinstance(markup, unicode): - self.originalEncoding = None - self.unicode = unicode(markup) - return - - u = None - for proposedEncoding in overrideEncodings: - u = self._convertFrom(proposedEncoding) - if u: break - if not u: - for proposedEncoding in (documentEncoding, sniffedEncoding): - u = self._convertFrom(proposedEncoding) - if u: break - - # If no luck and we have auto-detection library, try that: - if not u and chardet and not isinstance(self.markup, unicode): - u = self._convertFrom(chardet.detect(self.markup)['encoding']) - - # As a last resort, try utf-8 and windows-1252: - if not u: - for proposed_encoding in ("utf-8", "windows-1252"): - u = self._convertFrom(proposed_encoding) - if u: break - - self.unicode = u - if not u: self.originalEncoding = None - - def _subMSChar(self, match): - """Changes a MS smart quote character to an XML or HTML - entity.""" - orig = match.group(1) - sub = self.MS_CHARS.get(orig) - if type(sub) == types.TupleType: - if self.smartQuotesTo == 'xml': - sub = '&#x'.encode() + sub[1].encode() + ';'.encode() - else: - sub = '&'.encode() + sub[0].encode() + ';'.encode() - else: - sub = sub.encode() - return sub - - def _convertFrom(self, proposed): - proposed = self.find_codec(proposed) - if not proposed or proposed in self.triedEncodings: - return None - self.triedEncodings.append(proposed) - markup = self.markup - - # Convert smart quotes to HTML if coming from an encoding - # that might have them. - if self.smartQuotesTo and proposed.lower() in("windows-1252", - "iso-8859-1", - "iso-8859-2"): - smart_quotes_re = "([\x80-\x9f])" - smart_quotes_compiled = re.compile(smart_quotes_re) - markup = smart_quotes_compiled.sub(self._subMSChar, markup) - - try: - # print "Trying to convert document to %s" % proposed - u = self._toUnicode(markup, proposed) - self.markup = u - self.originalEncoding = proposed - except Exception, e: - # print "That didn't work!" - # print e - return None - #print "Correct encoding: %s" % proposed - return self.markup - - def _toUnicode(self, data, encoding): - '''Given a string and its encoding, decodes the string into Unicode. - %encoding is a string recognized by encodings.aliases''' - - # strip Byte Order Mark (if present) - if (len(data) >= 4) and (data[:2] == '\xfe\xff') \ - and (data[2:4] != '\x00\x00'): - encoding = 'utf-16be' - data = data[2:] - elif (len(data) >= 4) and (data[:2] == '\xff\xfe') \ - and (data[2:4] != '\x00\x00'): - encoding = 'utf-16le' - data = data[2:] - elif data[:3] == '\xef\xbb\xbf': - encoding = 'utf-8' - data = data[3:] - elif data[:4] == '\x00\x00\xfe\xff': - encoding = 'utf-32be' - data = data[4:] - elif data[:4] == '\xff\xfe\x00\x00': - encoding = 'utf-32le' - data = data[4:] - newdata = unicode(data, encoding) - return newdata - - def _detectEncoding(self, xml_data, isHTML=False): - """Given a document, tries to detect its XML encoding.""" - xml_encoding = sniffed_xml_encoding = None - try: - if xml_data[:4] == '\x4c\x6f\xa7\x94': - # EBCDIC - xml_data = self._ebcdic_to_ascii(xml_data) - elif xml_data[:4] == '\x00\x3c\x00\x3f': - # UTF-16BE - sniffed_xml_encoding = 'utf-16be' - xml_data = unicode(xml_data, 'utf-16be').encode('utf-8') - elif (len(xml_data) >= 4) and (xml_data[:2] == '\xfe\xff') \ - and (xml_data[2:4] != '\x00\x00'): - # UTF-16BE with BOM - sniffed_xml_encoding = 'utf-16be' - xml_data = unicode(xml_data[2:], 'utf-16be').encode('utf-8') - elif xml_data[:4] == '\x3c\x00\x3f\x00': - # UTF-16LE - sniffed_xml_encoding = 'utf-16le' - xml_data = unicode(xml_data, 'utf-16le').encode('utf-8') - elif (len(xml_data) >= 4) and (xml_data[:2] == '\xff\xfe') and \ - (xml_data[2:4] != '\x00\x00'): - # UTF-16LE with BOM - sniffed_xml_encoding = 'utf-16le' - xml_data = unicode(xml_data[2:], 'utf-16le').encode('utf-8') - elif xml_data[:4] == '\x00\x00\x00\x3c': - # UTF-32BE - sniffed_xml_encoding = 'utf-32be' - xml_data = unicode(xml_data, 'utf-32be').encode('utf-8') - elif xml_data[:4] == '\x3c\x00\x00\x00': - # UTF-32LE - sniffed_xml_encoding = 'utf-32le' - xml_data = unicode(xml_data, 'utf-32le').encode('utf-8') - elif xml_data[:4] == '\x00\x00\xfe\xff': - # UTF-32BE with BOM - sniffed_xml_encoding = 'utf-32be' - xml_data = unicode(xml_data[4:], 'utf-32be').encode('utf-8') - elif xml_data[:4] == '\xff\xfe\x00\x00': - # UTF-32LE with BOM - sniffed_xml_encoding = 'utf-32le' - xml_data = unicode(xml_data[4:], 'utf-32le').encode('utf-8') - elif xml_data[:3] == '\xef\xbb\xbf': - # UTF-8 with BOM - sniffed_xml_encoding = 'utf-8' - xml_data = unicode(xml_data[3:], 'utf-8').encode('utf-8') - else: - sniffed_xml_encoding = 'ascii' - pass - except: - xml_encoding_match = None - xml_encoding_re = '^<\?.*encoding=[\'"](.*?)[\'"].*\?>'.encode() - xml_encoding_match = re.compile(xml_encoding_re).match(xml_data) - if not xml_encoding_match and isHTML: - meta_re = '<\s*meta[^>]+charset=([^>]*?)[;\'">]'.encode() - regexp = re.compile(meta_re, re.I) - xml_encoding_match = regexp.search(xml_data) - if xml_encoding_match is not None: - xml_encoding = xml_encoding_match.groups()[0].decode( - 'ascii').lower() - if isHTML: - self.declaredHTMLEncoding = xml_encoding - if sniffed_xml_encoding and \ - (xml_encoding in ('iso-10646-ucs-2', 'ucs-2', 'csunicode', - 'iso-10646-ucs-4', 'ucs-4', 'csucs4', - 'utf-16', 'utf-32', 'utf_16', 'utf_32', - 'utf16', 'u16')): - xml_encoding = sniffed_xml_encoding - return xml_data, xml_encoding, sniffed_xml_encoding - - - def find_codec(self, charset): - return self._codec(self.CHARSET_ALIASES.get(charset, charset)) \ - or (charset and self._codec(charset.replace("-", ""))) \ - or (charset and self._codec(charset.replace("-", "_"))) \ - or charset - - def _codec(self, charset): - if not charset: return charset - codec = None - try: - codecs.lookup(charset) - codec = charset - except (LookupError, ValueError): - pass - return codec - - EBCDIC_TO_ASCII_MAP = None - def _ebcdic_to_ascii(self, s): - c = self.__class__ - if not c.EBCDIC_TO_ASCII_MAP: - emap = (0,1,2,3,156,9,134,127,151,141,142,11,12,13,14,15, - 16,17,18,19,157,133,8,135,24,25,146,143,28,29,30,31, - 128,129,130,131,132,10,23,27,136,137,138,139,140,5,6,7, - 144,145,22,147,148,149,150,4,152,153,154,155,20,21,158,26, - 32,160,161,162,163,164,165,166,167,168,91,46,60,40,43,33, - 38,169,170,171,172,173,174,175,176,177,93,36,42,41,59,94, - 45,47,178,179,180,181,182,183,184,185,124,44,37,95,62,63, - 186,187,188,189,190,191,192,193,194,96,58,35,64,39,61,34, - 195,97,98,99,100,101,102,103,104,105,196,197,198,199,200, - 201,202,106,107,108,109,110,111,112,113,114,203,204,205, - 206,207,208,209,126,115,116,117,118,119,120,121,122,210, - 211,212,213,214,215,216,217,218,219,220,221,222,223,224, - 225,226,227,228,229,230,231,123,65,66,67,68,69,70,71,72, - 73,232,233,234,235,236,237,125,74,75,76,77,78,79,80,81, - 82,238,239,240,241,242,243,92,159,83,84,85,86,87,88,89, - 90,244,245,246,247,248,249,48,49,50,51,52,53,54,55,56,57, - 250,251,252,253,254,255) - import string - c.EBCDIC_TO_ASCII_MAP = string.maketrans( \ - ''.join(map(chr, range(256))), ''.join(map(chr, emap))) - return s.translate(c.EBCDIC_TO_ASCII_MAP) - - MS_CHARS = { '\x80' : ('euro', '20AC'), - '\x81' : ' ', - '\x82' : ('sbquo', '201A'), - '\x83' : ('fnof', '192'), - '\x84' : ('bdquo', '201E'), - '\x85' : ('hellip', '2026'), - '\x86' : ('dagger', '2020'), - '\x87' : ('Dagger', '2021'), - '\x88' : ('circ', '2C6'), - '\x89' : ('permil', '2030'), - '\x8A' : ('Scaron', '160'), - '\x8B' : ('lsaquo', '2039'), - '\x8C' : ('OElig', '152'), - '\x8D' : '?', - '\x8E' : ('#x17D', '17D'), - '\x8F' : '?', - '\x90' : '?', - '\x91' : ('lsquo', '2018'), - '\x92' : ('rsquo', '2019'), - '\x93' : ('ldquo', '201C'), - '\x94' : ('rdquo', '201D'), - '\x95' : ('bull', '2022'), - '\x96' : ('ndash', '2013'), - '\x97' : ('mdash', '2014'), - '\x98' : ('tilde', '2DC'), - '\x99' : ('trade', '2122'), - '\x9a' : ('scaron', '161'), - '\x9b' : ('rsaquo', '203A'), - '\x9c' : ('oelig', '153'), - '\x9d' : '?', - '\x9e' : ('#x17E', '17E'), - '\x9f' : ('Yuml', ''),} - -####################################################################### - - -#By default, act as an HTML pretty-printer. -if __name__ == '__main__': - import sys - soup = BeautifulSoup(sys.stdin) - print soup.prettify() diff --git a/plugin/readability/README b/plugin/readability/README deleted file mode 100755 index 5e08c96..0000000 --- a/plugin/readability/README +++ /dev/null @@ -1,24 +0,0 @@ -This code is under the Apache License 2.0. http://www.apache.org/licenses/LICENSE-2.0 - -This is a python port of a ruby port of arc90's readability project - -http://lab.arc90.com/experiments/readability/ - -Given a html document, it pulls out the main body text and cleans it up. - -Ruby port by starrhorne and iterationlabs -Python port by gfxmonk - -This port uses BeautifulSoup for the HTML parsing. That means it can be -a little slow, but will work on Google App Engine (unlike libxml-based -libraries) - - -**note**: I don't currently have any plans for using or improving this -library, and it's far from perfect (slow, and almost certainly buggy). -So if you do something cool with it or have a better tool that does -the same job, please let me know and I can link to it from here. - -If you're looking for alternatives / forks, here's the list so far: - - http://www.minvolai.com/blog/decruft-arc90s-readability-in-python/ - - https://github.com/buriy/python-readability diff --git a/plugin/readability/__init__.py b/plugin/readability/__init__.py deleted file mode 100755 index f554b39..0000000 --- a/plugin/readability/__init__.py +++ /dev/null @@ -1,2 +0,0 @@ -from readability import Document, main -from page_parser import ascii, Unparseable diff --git a/plugin/readability/page_parser.py b/plugin/readability/page_parser.py deleted file mode 100755 index 1c80ca7..0000000 --- a/plugin/readability/page_parser.py +++ /dev/null @@ -1,145 +0,0 @@ -import re -from url_helpers import absolute_url -from BeautifulSoup import BeautifulSoup, HTMLParseError, UnicodeDammit -from logging import error - -__all__ = [ - 'Unparseable', - 'parse', - 'get_title', - 'get_body', - 'ascii'] - -def debug(s): pass - -class Unparseable(ValueError): - pass - -def parse(raw_content, base_href=None, notify=lambda x: None): - for parse_method in _parse_methods(): - try: - return parse_method(raw_content, base_href) - except HTMLParseError, e: - notify("parsing (%s) failed: %s" % (parse_method.__name__, e)) - continue - raise Unparseable() - -def get_title(soup): - title = unicode(getattr(soup.title, 'string', '')) - if not title: - return None - return normalize_spaces(title) - - -def get_body(soup): - [ elem.extract() for elem in soup.findAll(['script', 'link', 'style']) ] - raw_html = unicode(soup.body or soup) - cleaned = clean_attributes(raw_html) - try: - BeautifulSoup(cleaned) - return cleaned - except HTMLParseError: - error("cleansing broke html content: %s\n---------\n%s" % (raw_html,cleaned)) - return raw_html - -def ascii(s): - return s.decode('ascii', 'ignore') - -class Replacement(object): - def __init__(self, desc, regex, replacement): - self.desc = desc - self.regex = regex - self.replacement = replacement - - def apply(self, content): -# # useful for debugging: -# try: -# print self. desc + ':' + str(self.regex.findall(content)) -# except RuntimeError: pass - return self.regex.sub(self.replacement, content) - -def beautiful_soup(content, base_href): - soup = BeautifulSoup(content) - if base_href: - _fix_references(soup, base_href) - return soup - - -def _make_absolute_links(soup, base_href): - for link in soup.findAll('a', attrs={'href':True}): - link['href'] = absolute_url(link['href'], base_href) - -def _make_absolute_images(soup, base_href): - for img in soup.findAll('img', attrs={'src':True}): - img['src'] = absolute_url(img['src'], base_href) - -def _fix_references(soup, base_href): - _make_absolute_links(soup, base_href) - _make_absolute_images(soup, base_href) - -# a bunch of regexes to hack around lousy html -dodgy_regexes = ( - Replacement('javascript', - regex=re.compile('<script.*?</script[^>]*>', re.DOTALL | re.IGNORECASE), - replacement=''), - - Replacement('double double-quoted attributes', - regex=re.compile('(="[^"]+")"+'), - replacement='\\1'), - - Replacement('unclosed tags', - regex = re.compile('(<[a-zA-Z]+[^>]*)(<[a-zA-Z]+[^<>]*>)'), - replacement='\\1>\\2'), - - Replacement('unclosed (numerical) attribute values', - regex = re.compile('(<[^>]*[a-zA-Z]+\s*=\s*"[0-9]+)( [a-zA-Z]+="\w+"|/?>)'), - replacement='\\1"\\2'), - ) - - -# helpers for parsing -def normalize_spaces(s): - """replace any sequence of whitespace - characters with a single space""" - return ' '.join(s.split()) - -def _remove_crufty_html(content): - for replacement in dodgy_regexes: - content = replacement.apply(content) - return content - -def _parse_methods(): - def unicode_cleansed(content, base_href): - content = UnicodeDammit(content, isHTML=True).markup - cleaned = _remove_crufty_html(content) - debug("Cleaned content: %s" % (cleaned,)) - return beautiful_soup(cleaned, base_href) - - def ascii_cleansed(content, base_href): - content = ascii(content) - cleaned = _remove_crufty_html(content) - debug("Cleaned content: %s" % (cleaned,)) - return beautiful_soup(cleaned, base_href) - - return ( - beautiful_soup, - unicode_cleansed, - ascii_cleansed) - -# strip out a set of nuisance html attributes that can mess up rendering in RSS feeds -bad_attrs = ['width','height','style','[-a-z]*color','background[-a-z]*'] -single_quoted = "'[^']+'" -double_quoted = '"[^"]+"' -non_space = '[^ "\'>]+' -htmlstrip = re.compile("<" # open - "([^>]+) " # prefix - "(?:%s) *" % ('|'.join(bad_attrs),) + # undesirable attributes - '= *(?:%s|%s|%s)' % (non_space, single_quoted, double_quoted) + # value - "([^>]*)" # postfix - ">" # end -, re.I) -def clean_attributes(html): - while htmlstrip.search(html): - html = htmlstrip.sub('<\\1\\2>', html) - return html - diff --git a/plugin/readability/readability.py b/plugin/readability/readability.py deleted file mode 100755 index 9c5f5c6..0000000 --- a/plugin/readability/readability.py +++ /dev/null @@ -1,343 +0,0 @@ -#!/usr/bin/env python -from BeautifulSoup import NavigableString -from page_parser import parse, get_title, get_body, Unparseable -import logging -import re - -REGEXES = { 'unlikelyCandidatesRe': re.compile('combx|comment|disqus|foot|header|menu|meta|nav|rss|shoutbox|sidebar|sponsor',re.I), - 'okMaybeItsACandidateRe': re.compile('and|article|body|column|main',re.I), - 'positiveRe': re.compile('article|body|content|entry|hentry|page|pagination|post|text',re.I), - 'negativeRe': re.compile('combx|comment|contact|foot|footer|footnote|link|media|meta|promo|related|scroll|shoutbox|sponsor|tags|widget',re.I), - 'divToPElementsRe': re.compile('<(a|blockquote|dl|div|img|ol|p|pre|table|ul)',re.I), - 'replaceBrsRe': re.compile('(<br[^>]*>[ \n\r\t]*){2,}',re.I), - 'replaceFontsRe': re.compile('<(\/?)font[^>]*>',re.I), - 'trimRe': re.compile('^\s+|\s+$/'), - 'normalizeRe': re.compile('\s{2,}/'), - 'killBreaksRe': re.compile('(<br\s*\/?>(\s| ?)*){1,}/'), - 'videoRe': re.compile('http:\/\/(www\.)?(youtube|vimeo)\.com', re.I), -} - -from collections import defaultdict -def describe(node): - if not hasattr(node, 'name'): - return "[text]" - return "%s#%s.%s" % ( - node.name, node.get('id', ''), node.get('class','')) - -def _text(node): - return " ".join(node.findAll(text=True)) - -class Document: - TEXT_LENGTH_THRESHOLD = 25 - RETRY_LENGTH = 250 - - def __init__(self, input, notify=None, **options): - self.input = input - self.options = defaultdict(lambda: None) - for k, v in options.items(): - self.options[k] = v - self.notify = notify or logging.info - self.html = None - - def _html(self, force=False): - if force or self.html is None: - self.html = parse(self.input, self.options['url'], notify=self.notify) - return self.html - - def content(self): - return get_body(self._html()) - - def title(self): - return get_title(self._html()) - - def summary(self): - try: - ruthless = True - while True: - self._html(True) - [i.extract() for i in self.tags(self.html, 'script', 'style')] - - if ruthless: self.remove_unlikely_candidates() - self.transform_misused_divs_into_paragraphs() - candidates = self.score_paragraphs(self.options.get('min_text_length', self.TEXT_LENGTH_THRESHOLD)) - best_candidate = self.select_best_candidate(candidates) - if best_candidate: - article = self.get_article(candidates, best_candidate) - else: - if ruthless: - ruthless = False - self.debug("ended up stripping too much - going for a safer parse") - # try again - continue - else: - article = self.html.find('body') or self.html - - cleaned_article = self.sanitize(article, candidates) - of_acceptable_length = len(cleaned_article or '') >= (self.options['retry_length'] or self.RETRY_LENGTH) - if ruthless and not of_acceptable_length: - ruthless = False - continue # try again - else: - return cleaned_article - except StandardError, e: - logging.exception('error getting summary:') - raise Unparseable(str(e)) - - def get_article(self, candidates, best_candidate): - # Now that we have the top candidate, look through its siblings for content that might also be related. - # Things like preambles, content split by ads that we removed, etc. - - sibling_score_threshold = max([10, best_candidate['content_score'] * 0.2]) - output = parse("<div/>") - for sibling in best_candidate['elem'].parent.contents: - if isinstance(sibling, NavigableString): continue - append = False - if sibling is best_candidate['elem']: - append = True - sibling_key = HashableElement(sibling) - if sibling_key in candidates and candidates[sibling_key]['content_score'] >= sibling_score_threshold: - append = True - - if sibling.name == "p": - link_density = self.get_link_density(sibling) - node_content = sibling.string or "" - node_length = len(node_content) - - if node_length > 80 and link_density < 0.25: - append = True - elif node_length < 80 and link_density == 0 and re.search('\.( |$)', node_content): - append = True - - if append: - output.append(sibling) - - if not output: output.append(best_candidate) - return output - - def select_best_candidate(self, candidates): - sorted_candidates = sorted(candidates.values(), key=lambda x: x['content_score'], reverse=True) - self.debug("Top 5 candidates:") - for candidate in sorted_candidates[:5]: - elem = candidate['elem'] - self.debug("Candidate %s with score %s" % (describe(elem), candidate['content_score'])) - - if len(sorted_candidates) == 0: - return None - best_candidate = sorted_candidates[0] - self.debug("Best candidate %s with score %s" % (describe(best_candidate['elem']), best_candidate['content_score'])) - return best_candidate - - def get_link_density(self, elem): - link_length = len("".join([i.text or "" for i in elem.findAll("a")])) - text_length = len(_text(elem)) - return float(link_length) / max(text_length, 1) - - def score_paragraphs(self, min_text_length): - candidates = {} - elems = self.tags(self.html, "p","td") - - for elem in elems: - parent_node = elem.parent - grand_parent_node = parent_node.parent - parent_key = HashableElement(parent_node) - grand_parent_key = HashableElement(grand_parent_node) - - inner_text = _text(elem) - - # If this paragraph is less than 25 characters, don't even count it. - if (not inner_text) or len(inner_text) < min_text_length: - continue - - if parent_key not in candidates: - candidates[parent_key] = self.score_node(parent_node) - if grand_parent_node and grand_parent_key not in candidates: - candidates[grand_parent_key] = self.score_node(grand_parent_node) - - content_score = 1 - content_score += len(inner_text.split(',')) - content_score += min([(len(inner_text) / 100), 3]) - - candidates[parent_key]['content_score'] += content_score - if grand_parent_node: - candidates[grand_parent_key]['content_score'] += content_score / 2.0 - - # Scale the final candidates score based on link density. Good content should have a - # relatively small link density (5% or less) and be mostly unaffected by this operation. - for elem, candidate in candidates.items(): - candidate['content_score'] *= (1 - self.get_link_density(elem)) - self.debug("candidate %s scored %s" % (describe(elem), candidate['content_score'])) - - return candidates - - def class_weight(self, e): - weight = 0 - if e.get('class', None): - if REGEXES['negativeRe'].search(e['class']): - weight -= 25 - - if REGEXES['positiveRe'].search(e['class']): - weight += 25 - - if e.get('id', None): - if REGEXES['negativeRe'].search(e['id']): - weight -= 25 - - if REGEXES['positiveRe'].search(e['id']): - weight += 25 - - return weight - - def score_node(self, elem): - content_score = self.class_weight(elem) - name = elem.name.lower() - if name == "div": - content_score += 5 - elif name == "blockquote": - content_score += 3 - elif name == "form": - content_score -= 3 - elif name == "th": - content_score -= 5 - return { 'content_score': content_score, 'elem': elem } - - def debug(self, *a): - if self.options['debug']: - logging.debug(*a) - - def remove_unlikely_candidates(self): - for elem in self.html.findAll(): - s = "%s%s" % (elem.get('class', ''), elem.get('id', '')) - if REGEXES['unlikelyCandidatesRe'].search(s) and (not REGEXES['okMaybeItsACandidateRe'].search(s)) and elem.name != 'body': - self.debug("Removing unlikely candidate - %s" % (s,)) - elem.extract() - - def transform_misused_divs_into_paragraphs(self): - for elem in self.html.findAll(): - if elem.name.lower() == "div": - # transform <div>s that do not contain other block elements into <p>s - if REGEXES['divToPElementsRe'].search(''.join(map(unicode, elem.contents))): - self.debug("Altering div(#%s.%s) to p" % (elem.get('id', ''), elem.get('class', ''))) - elem.name = "p" - - def tags(self, node, *tag_names): - for tag_name in tag_names: - for e in node.findAll(tag_name): - yield e - - def sanitize(self, node, candidates): - 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: header.extract() - - for elem in self.tags(node, "form", "iframe"): - elem.extract() - - # Conditionally clean <table>s, <ul>s, and <div>s - for el in self.tags(node, "table", "ul", "div"): - weight = self.class_weight(el) - el_key = HashableElement(el) - if el_key in candidates: - content_score = candidates[el_key]['content_score'] - else: - content_score = 0 - name = el.name - - if weight + content_score < 0: - el.extract() - self.debug("Conditionally cleaned %s with weight %s and content score %s because score + content score was less than zero." % - (describe(el), weight, content_score)) - elif len(_text(el).split(",")) < 10: - counts = {} - for kind in ['p', 'img', 'li', 'a', 'embed', 'input']: - counts[kind] = len(el.findAll(kind)) - counts["li"] -= 100 - - content_length = len(_text(el)) # Count the text length excluding any surrounding whitespace - link_density = self.get_link_density(el) - to_remove = False - reason = "" - - if counts["img"] > counts["p"]: - reason = "too many images" - to_remove = True - elif counts["li"] > counts["p"] and name != "ul" and name != "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 < (self.options.get('min_text_length', self.TEXT_LENGTH_THRESHOLD)) and (counts["img"] == 0 or counts["img"] > 2): - reason = "too short a content length without a single image" - to_remove = True - elif weight < 25 and link_density > 0.2: - reason = "too many links for its weight (#{weight})" - to_remove = True - elif weight >= 25 and link_density > 0.5: - reason = "too many links for its weight (#{weight})" - to_remove = True - elif (counts["embed"] == 1 and content_length < 75) or counts["embed"] > 1: - reason = "<embed>s with too short a content length, or too many <embed>s" - to_remove = True - - if to_remove: - self.debug("Conditionally cleaned %s#%s.%s with weight %s and content score %s because it has %s." % - (el.name, el.get('id',''), el.get('class', ''), weight, content_score, reason)) - el.extract() - - for el in ([node] + node.findAll()): - if not (self.options['attributes']): - el.attrMap = {} - - return unicode(node) - -class HashableElement(): - def __init__(self, node): - self.node = node - self._path = None - - def _get_path(self): - if self._path is None: - reverse_path = [] - node = self.node - while node: - node_id = (node.name, tuple(node.attrs), node.string) - reverse_path.append(node_id) - node = node.parent - self._path = tuple(reverse_path) - return self._path - path = property(_get_path) - - def __hash__(self): - return hash(self.path) - - def __eq__(self, other): - return self.path == other.path - - def __getattr__(self, name): - return getattr(self.node, name) - -def main(): - import sys - from optparse import OptionParser - parser = OptionParser(usage="%prog: [options] [file]") - parser.add_option('-v', '--verbose', action='store_true') - parser.add_option('-u', '--url', help="use URL instead of a local file") - (options, args) = parser.parse_args() - - if not (len(args) == 1 or options.url): - parser.print_help() - sys.exit(1) - logging.basicConfig(level=logging.DEBUG) - - file = None - if options.url: - import urllib - file = urllib.urlopen(options.url) - else: - file = open(args[0]) - try: - print Document(file.read(), debug=options.verbose).summary().encode('ascii','ignore') - finally: - file.close() - -if __name__ == '__main__': - main() diff --git a/plugin/readability/setup.py b/plugin/readability/setup.py deleted file mode 100755 index a027449..0000000 --- a/plugin/readability/setup.py +++ /dev/null @@ -1,22 +0,0 @@ -#!/usr/bin/env python -from distutils.core import setup - - -setup( - name="python-readability", - author="Tim Cuthbertson", - author_email="tim3d.junk+github@gmail.com", - description="python port of arc90's readability bookmarklet", - long_description=open("README").read(), - license="Apache License 2.0", - url="http://github.com/gfxmonk/python-readability", - packages=[ - "readability", - ], - classifiers=[ - "Environment :: Web Environment", - "Intended Audience :: Developers", - "Operating System :: OS Independent", - "Programming Language :: Python", - ], -) diff --git a/plugin/readability/url_helpers.py b/plugin/readability/url_helpers.py deleted file mode 100755 index 8234c8d..0000000 --- a/plugin/readability/url_helpers.py +++ /dev/null @@ -1,52 +0,0 @@ -import logging -from urlparse import urlparse - -def host_for_url(url): - """ - >>> host_for_url('http://base/whatever/fdsh') - 'base' - >>> host_for_url('invalid') - """ - host = urlparse(url)[1] - if not host: - logging.error("could not extract host from URL: %r" % (url,)) - return None - return host - -def absolute_url(url, base_href): - """ - >>> absolute_url('foo', 'http://base/whatever/ooo/fdsh') - 'http://base/whatever/ooo/foo' - - >>> absolute_url('foo/bar/', 'http://base') - 'http://base/foo/bar/' - - >>> absolute_url('/foo/bar', 'http://base/whatever/fdskf') - 'http://base/foo/bar' - - >>> absolute_url('\\n/foo/bar', 'http://base/whatever/fdskf') - 'http://base/foo/bar' - - >>> absolute_url('http://localhost/foo', 'http://base/whatever/fdskf') - 'http://localhost/foo' - """ - url = url.strip() - proto = urlparse(url)[0] - if proto: - return url - - base_url_parts = urlparse(base_href) - base_server = '://'.join(base_url_parts[:2]) - if url.startswith('/'): - return base_server + url - else: - path = base_url_parts[2] - if '/' in path: - path = path.rsplit('/', 1)[0] + '/' - else: - path = '/' - return base_server + path + url - -if __name__ == '__main__': - import doctest - doctest.testmod()
\ No newline at end of file |