bs4源码
Beautiful源码：
"""Beautiful Soup Elixir and Tonic "The Screen-Scraper's Friend" http://www.crummy.com/software/BeautifulSoup/ Beautiful Soup uses a pluggable XML or HTML parser to parse a (possibly invalid) document into a tree representation. Beautiful Soup provides methods and Pythonic idioms that make it easy to navigate, search, and modify the parse tree. Beautiful Soup works with Python 2.7 and up. It works better if lxml and/or html5lib is installed. For more than you ever wanted to know about Beautiful Soup, see the documentation: http://www.crummy.com/software/BeautifulSoup/bs4/doc/ """ # Use of this source code is governed by a BSD-style license that can be # found in the LICENSE file. __author__ = "Leonard Richardson (leonardr@segfault.org)" __version__ = "4.6.3" __copyright__ = "Copyright (c) 2004-2018 Leonard Richardson" __license__ = "MIT" __all__ = ['BeautifulSoup'] import os import re import sys import traceback import warnings from .builder import builder_registry, ParserRejectedMarkup from .dammit import UnicodeDammit from .element import ( CData, Comment, DEFAULT_OUTPUT_ENCODING, Declaration, Doctype, NavigableString, PageElement, ProcessingInstruction, ResultSet, SoupStrainer, Tag, ) # The very first thing we do is give a useful error if someone is # running this code under Python 3 without converting it. 'You are trying to run the Python 2 version of Beautiful Soup under Python 3. This will not work.'!='You need to convert the code, either by installing it (`python setup.py install`) or by running 2to3 (`2to3 -w bs4`).' class BeautifulSoup(Tag): """ This class defines the basic interface called by the tree builders. These methods will be called by the parser: reset() feed(markup) The tree builder may call these methods from its feed() implementation: handle_starttag(name, attrs) # See note about return value handle_endtag(name) handle_data(data) # Appends to the current data node endData(containerClass=NavigableString) # Ends the current data node No matter how complicated the underlying parser is, you should be able to build a tree using 'start tag' events, 'end tag' events, 'data' events, and "done with data" events. If you encounter an empty-element tag (aka a self-closing tag, like HTML's <br> tag), call handle_starttag and then handle_endtag. """ ROOT_TAG_NAME = '[document]' # If the end-user gives no indication which tree builder they # want, look for one with these features. DEFAULT_BUILDER_FEATURES = ['html', 'fast'] ASCII_SPACES = '\x20\x0a\x09\x0c\x0d' NO_PARSER_SPECIFIED_WARNING = "No parser was explicitly specified, so I'm using the best available %(markup_type)s parser for this system (\"%(parser)s\"). This usually isn't a problem, but if you run this code on another system, or in a different virtual environment, it may use a different parser and behave differently.\n\nThe code that caused this warning is on line %(line_number)s of the file %(filename)s. To get rid of this warning, pass the additional argument 'features=\"%(parser)s\"' to the BeautifulSoup constructor.\n" def __init__(self, markup="", features=None, builder=None, parse_only=None, from_encoding=None, exclude_encodings=None, **kwargs): """Constructor. :param markup: A string or a file-like object representing markup to be parsed. :param features: Desirable features of the parser to be used. This may be the name of a specific parser ("lxml", "lxml-xml", "html.parser", or "html5lib") or it may be the type of markup to be used ("html", "html5", "xml"). It's recommended that you name a specific parser, so that Beautiful Soup gives you the same results across platforms and virtual environments. :param builder: A specific TreeBuilder to use instead of looking one up based on `features`. You shouldn't need to use this. :param parse_only: A SoupStrainer. Only parts of the document matching the SoupStrainer will be considered. This is useful when parsing part of a document that would otherwise be too large to fit into memory. :param from_encoding: A string indicating the encoding of the document to be parsed. Pass this in if Beautiful Soup is guessing wrongly about the document's encoding. :param exclude_encodings: A list of strings indicating encodings known to be wrong. Pass this in if you don't know the document's encoding but you know Beautiful Soup's guess is wrong. :param kwargs: For backwards compatibility purposes, the constructor accepts certain keyword arguments used in Beautiful Soup 3. None of these arguments do anything in Beautiful Soup 4 and there's no need to actually pass keyword arguments into the constructor. """ if 'convertEntities' in kwargs: warnings.warn( "BS4 does not respect the convertEntities argument to the " "BeautifulSoup constructor. Entities are always converted " "to Unicode characters.") if 'markupMassage' in kwargs: del kwargs['markupMassage'] warnings.warn( "BS4 does not respect the markupMassage argument to the " "BeautifulSoup constructor. The tree builder is responsible " "for any necessary markup massage.") if 'smartQuotesTo' in kwargs: del kwargs['smartQuotesTo'] warnings.warn( "BS4 does not respect the smartQuotesTo argument to the " "BeautifulSoup constructor. Smart quotes are always converted " "to Unicode characters.") if 'selfClosingTags' in kwargs: del kwargs['selfClosingTags'] warnings.warn( "BS4 does not respect the selfClosingTags argument to the " "BeautifulSoup constructor. The tree builder is responsible " "for understanding self-closing tags.") if 'isHTML' in kwargs: del kwargs['isHTML'] warnings.warn( "BS4 does not respect the isHTML argument to the " "BeautifulSoup constructor. Suggest you use " "features='lxml' for HTML and features='lxml-xml' for " "XML.") def deprecated_argument(old_name, new_name): if old_name in kwargs: warnings.warn( 'The "%s" argument to the BeautifulSoup constructor ' 'has been renamed to "%s."' % (old_name, new_name)) value = kwargs[old_name] del kwargs[old_name] return value return None parse_only = parse_only or deprecated_argument( "parseOnlyThese", "parse_only") from_encoding = from_encoding or deprecated_argument( "fromEncoding", "from_encoding") if from_encoding and isinstance(markup, str): warnings.warn("You provided Unicode markup but also provided a value for from_encoding. Your from_encoding will be ignored.") from_encoding = None if len(kwargs) > 0: arg = list(kwargs.keys()).pop() raise TypeError( "__init__() got an unexpected keyword argument '%s'" % arg) if builder is None: original_features = features if isinstance(features, str): features = [features] if features is None or len(features) == 0: features = self.DEFAULT_BUILDER_FEATURES builder_class = builder_registry.lookup(*features) if builder_class is None: raise FeatureNotFound( "Couldn't find a tree builder with the features you " "requested: %s. Do you need to install a parser library?" % ",".join(features)) builder = builder_class() if not (original_features == builder.NAME or original_features in builder.ALTERNATE_NAMES): if builder.is_xml: markup_type = "XML" else: markup_type = "HTML" # This code adapted from warnings.py so that we get the same line # of code as our warnings.warn() call gets, even if the answer is wrong # (as it may be in a multithreading situation). caller = None try: caller = sys._getframe(1) except ValueError: pass if caller: globals = caller.f_globals line_number = caller.f_lineno else: globals = sys.__dict__ line_number= 1 filename = globals.get('__file__') if filename: fnl = filename.lower() if fnl.endswith((".pyc", ".pyo")): filename = filename[:-1] if filename: # If there is no filename at all, the user is most likely in a REPL, # and the warning is not necessary. values = dict( filename=filename, line_number=line_number, parser=builder.NAME, markup_type=markup_type ) warnings.warn(self.NO_PARSER_SPECIFIED_WARNING % values, stacklevel=2) self.builder = builder self.is_xml = builder.is_xml self.known_xml = self.is_xml self.builder.soup = self self.parse_only = parse_only if hasattr(markup, 'read'): # It's a file-type object. markup = markup.read() elif len(markup) <= 256 and ( (isinstance(markup, bytes) and not b'<' in markup) or (isinstance(markup, str) and not '<' in markup) ): # Print out warnings for a couple beginner problems # involving passing non-markup to Beautiful Soup. # Beautiful Soup will still parse the input as markup, # just in case that's what the user really wants. if (isinstance(markup, str) and not os.path.supports_unicode_filenames): possible_filename = markup.encode("utf8") else: possible_filename = markup is_file = False try: is_file = os.path.exists(possible_filename) except Exception as e: # This is almost certainly a problem involving # characters not valid in filenames on this # system. Just let it go. pass if is_file: if isinstance(markup, str): markup = markup.encode("utf8") warnings.warn( '"%s" looks like a filename, not markup. You should' ' probably open this file and pass the filehandle into' ' Beautiful Soup.' % markup) self._check_markup_is_url(markup) for (self.markup, self.original_encoding, self.declared_html_encoding, self.contains_replacement_characters) in ( self.builder.prepare_markup( markup, from_encoding, exclude_encodings=exclude_encodings)): self.reset() try: self._feed() break except ParserRejectedMarkup: pass # Clear out the markup and remove the builder's circular # reference to this object. self.markup = None self.builder.soup = None def __copy__(self): copy = type(self)( self.encode('utf-8'), builder=self.builder, from_encoding='utf-8' ) # Although we encoded the tree to UTF-8, that may not have # been the encoding of the original markup. Set the copy's # .original_encoding to reflect the original object's # .original_encoding. copy.original_encoding = self.original_encoding return copy def __getstate__(self): # Frequently a tree builder can't be pickled. d = dict(self.__dict__) if 'builder' in d and not self.builder.picklable: d['builder'] = None return d @staticmethod def _check_markup_is_url(markup): """ Check if markup looks like it's actually a url and raise a warning if so. Markup can be unicode or str (py2) / bytes (py3). """ if isinstance(markup, bytes): space = b' ' cant_start_with = (b"http:", b"https:") elif isinstance(markup, str): space = ' ' cant_start_with = ("http:", "https:") else: return if any(markup.startswith(prefix) for prefix in cant_start_with): if not space in markup: if isinstance(markup, bytes): decoded_markup = markup.decode('utf-8', 'replace') else: decoded_markup = markup warnings.warn( '"%s" looks like a URL. Beautiful Soup is not an' ' HTTP client. You should probably use an HTTP client like' ' requests to get the document behind the URL, and feed' ' that document to Beautiful Soup.' % decoded_markup ) def _feed(self): # Convert the document to Unicode. self.builder.reset() self.builder.feed(self.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 reset(self): Tag.__init__(self, self, self.builder, self.ROOT_TAG_NAME) self.hidden = 1 self.builder.reset() self.current_data = [] self.currentTag = None self.tagStack = [] self.preserve_whitespace_tag_stack = [] self.pushTag(self) def new_tag(self, name, namespace=None, nsprefix=None, attrs={}, **kwattrs): """Create a new tag associated with this soup.""" kwattrs.update(attrs) return Tag(None, self.builder, name, namespace, nsprefix, kwattrs) def new_string(self, s, subclass=NavigableString): """Create a new NavigableString associated with this soup.""" return subclass(s) def insert_before(self, successor): raise NotImplementedError("BeautifulSoup objects don't support insert_before().") def insert_after(self, successor): raise NotImplementedError("BeautifulSoup objects don't support insert_after().") def popTag(self): tag = self.tagStack.pop() if self.preserve_whitespace_tag_stack and tag == self.preserve_whitespace_tag_stack[-1]: self.preserve_whitespace_tag_stack.pop() #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] if tag.name in self.builder.preserve_whitespace_tags: self.preserve_whitespace_tag_stack.append(tag) def endData(self, containerClass=NavigableString): if self.current_data: current_data = ''.join(self.current_data) # If whitespace is not preserved, and this string contains # nothing but ASCII spaces, replace it with a single space # or newline. if not self.preserve_whitespace_tag_stack: strippable = True for i in current_data: if i not in self.ASCII_SPACES: strippable = False break if strippable: if '\n' in current_data: current_data = '\n' else: current_data = ' ' # Reset the data collector. self.current_data = [] # Should we add this string to the tree at all? if self.parse_only and len(self.tagStack) <= 1 and \ (not self.parse_only.text or \ not self.parse_only.search(current_data)): return o = containerClass(current_data) self.object_was_parsed(o) def object_was_parsed(self, o, parent=None, most_recent_element=None): """Add an object to the parse tree.""" parent = parent or self.currentTag previous_element = most_recent_element or self._most_recent_element next_element = previous_sibling = next_sibling = None if isinstance(o, Tag): next_element = o.next_element next_sibling = o.next_sibling previous_sibling = o.previous_sibling if not previous_element: previous_element = o.previous_element o.setup(parent, previous_element, next_element, previous_sibling, next_sibling) self._most_recent_element = o parent.contents.append(o) if parent.next_sibling: # This node is being inserted into an element that has # already been parsed. Deal with any dangling references. index = len(parent.contents)-1 while index >= 0: if parent.contents[index] is o: break index -= 1 else: raise ValueError( "Error building tree: supposedly %r was inserted " "into %r after the fact, but I don't see it!" % ( o, parent ) ) if index == 0: previous_element = parent previous_sibling = None else: previous_element = previous_sibling = parent.contents[index-1] if index == len(parent.contents)-1: next_element = parent.next_sibling next_sibling = None else: next_element = next_sibling = parent.contents[index+1] o.previous_element = previous_element if previous_element: previous_element.next_element = o o.next_element = next_element if next_element: next_element.previous_element = o o.next_sibling = next_sibling if next_sibling: next_sibling.previous_sibling = o o.previous_sibling = previous_sibling if previous_sibling: previous_sibling.next_sibling = o def _popToTag(self, name, nsprefix=None, 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: # The BeautifulSoup object itself can never be popped. return most_recently_popped = None stack_size = len(self.tagStack) for i in range(stack_size - 1, 0, -1): t = self.tagStack[i] if (name == t.name and nsprefix == t.prefix): if inclusivePop: most_recently_popped = self.popTag() break most_recently_popped = self.popTag() return most_recently_popped def handle_starttag(self, name, namespace, nsprefix, attrs): """Push a start tag on to the stack. If this method returns None, the tag was rejected by the SoupStrainer. You should proceed as if the tag had not occurred in the document. For instance, if this was a self-closing tag, don't call handle_endtag. """ # print "Start tag %s: %s" % (name, attrs) self.endData() if (self.parse_only and len(self.tagStack) <= 1 and (self.parse_only.text or not self.parse_only.search_tag(name, attrs))): return None tag = Tag(self, self.builder, name, namespace, nsprefix, attrs, self.currentTag, self._most_recent_element) if tag is None: return tag if self._most_recent_element: self._most_recent_element.next_element = tag self._most_recent_element = tag self.pushTag(tag) return tag def handle_endtag(self, name, nsprefix=None): #print "End tag: " + name self.endData() self._popToTag(name, nsprefix) def handle_data(self, data): self.current_data.append(data) def decode(self, pretty_print=False, eventual_encoding=DEFAULT_OUTPUT_ENCODING, formatter="minimal"): """Returns a string or Unicode representation of this document. To get Unicode, pass None for encoding.""" if self.is_xml: # Print the XML declaration encoding_part = '' if eventual_encoding != None: encoding_part = ' encoding="%s"' % eventual_encoding prefix = '<?xml version="1.0"%s?>\n' % encoding_part else: prefix = '' if not pretty_print: indent_level = None else: indent_level = 0 return prefix + super(BeautifulSoup, self).decode( indent_level, eventual_encoding, formatter) # Alias to make it easier to type import: 'from bs4 import _soup' _s = BeautifulSoup _soup = BeautifulSoup class BeautifulStoneSoup(BeautifulSoup): """Deprecated interface to an XML parser.""" def __init__(self, *args, **kwargs): kwargs['features'] = 'xml' warnings.warn( 'The BeautifulStoneSoup class is deprecated. Instead of using ' 'it, pass features="xml" into the BeautifulSoup constructor.') super(BeautifulStoneSoup, self).__init__(*args, **kwargs) class StopParsing(Exception): pass class FeatureNotFound(ValueError): pass #By default, act as an HTML pretty-printer. if __name__ == '__main__': import sys soup = BeautifulSoup(sys.stdin) print(soup.prettify())
find_all（）源码：
# Use of this source code is governed by a BSD-style license that can be # found in the LICENSE file. __license__ = "MIT" try: from collections.abc import Callable # Python 3.6 except ImportError as e: from collections import Callable import re import shlex import sys import warnings from bs4.dammit import EntitySubstitution DEFAULT_OUTPUT_ENCODING = "utf-8" PY3K = (sys.version_info[0] > 2) whitespace_re = re.compile(r"\s+") def _alias(attr): """Alias one attribute name to another for backward compatibility""" @property def alias(self): return getattr(self, attr) @alias.setter def alias(self): return setattr(self, attr) return alias class NamespacedAttribute(str): def __new__(cls, prefix, name, namespace=None): if name is None: obj = str.__new__(cls, prefix) elif prefix is None: # Not really namespaced. obj = str.__new__(cls, name) else: obj = str.__new__(cls, prefix + ":" + name) obj.prefix = prefix obj.name = name obj.namespace = namespace return obj class AttributeValueWithCharsetSubstitution(str): """A stand-in object for a character encoding specified in HTML.""" class CharsetMetaAttributeValue(AttributeValueWithCharsetSubstitution): """A generic stand-in for the value of a meta tag's 'charset' attribute. When Beautiful Soup parses the markup '<meta charset="utf8">', the value of the 'charset' attribute will be one of these objects. """ def __new__(cls, original_value): obj = str.__new__(cls, original_value) obj.original_value = original_value return obj def encode(self, encoding): return encoding class ContentMetaAttributeValue(AttributeValueWithCharsetSubstitution): """A generic stand-in for the value of a meta tag's 'content' attribute. When Beautiful Soup parses the markup: <meta http-equiv="content-type" content="text/html; charset=utf8"> The value of the 'content' attribute will be one of these objects. """ CHARSET_RE = re.compile(r"((^|;)\s*charset=)([^;]*)", re.M) def __new__(cls, original_value): match = cls.CHARSET_RE.search(original_value) if match is None: # No substitution necessary. return str.__new__(str, original_value) obj = str.__new__(cls, original_value) obj.original_value = original_value return obj def encode(self, encoding): def rewrite(match): return match.group(1) + encoding return self.CHARSET_RE.sub(rewrite, self.original_value) class HTMLAwareEntitySubstitution(EntitySubstitution): """Entity substitution rules that are aware of some HTML quirks. Specifically, the contents of <script> and <style> tags should not undergo entity substitution. Incoming NavigableString objects are checked to see if they're the direct children of a <script> or <style> tag. """ cdata_containing_tags = set(["script", "style"]) preformatted_tags = set(["pre"]) preserve_whitespace_tags = set(['pre', 'textarea']) @classmethod def _substitute_if_appropriate(cls, ns, f): if (isinstance(ns, NavigableString) and ns.parent is not None and ns.parent.name in cls.cdata_containing_tags): # Do nothing. return ns # Substitute. return f(ns) @classmethod def substitute_html(cls, ns): return cls._substitute_if_appropriate( ns, EntitySubstitution.substitute_html) @classmethod def substitute_xml(cls, ns): return cls._substitute_if_appropriate( ns, EntitySubstitution.substitute_xml) class Formatter(object): """Contains information about how to format a parse tree.""" # By default, represent void elements as <tag/> rather than <tag> void_element_close_prefix = '/' def substitute_entities(self, *args, **kwargs): """Transform certain characters into named entities.""" raise NotImplementedError() class HTMLFormatter(Formatter): """The default HTML formatter.""" def substitute(self, *args, **kwargs): return HTMLAwareEntitySubstitution.substitute_html(*args, **kwargs) class MinimalHTMLFormatter(Formatter): """A minimal HTML formatter.""" def substitute(self, *args, **kwargs): return HTMLAwareEntitySubstitution.substitute_xml(*args, **kwargs) class HTML5Formatter(HTMLFormatter): """An HTML formatter that omits the slash in a void tag.""" void_element_close_prefix = None class XMLFormatter(Formatter): """Substitute only the essential XML entities.""" def substitute(self, *args, **kwargs): return EntitySubstitution.substitute_xml(*args, **kwargs) class HTMLXMLFormatter(Formatter): """Format XML using HTML rules.""" def substitute(self, *args, **kwargs): return HTMLAwareEntitySubstitution.substitute_html(*args, **kwargs) class PageElement(object): """Contains the navigational information for some part of the page (either a tag or a piece of text)""" # There are five possible values for the "formatter" argument passed in # to methods like encode() and prettify(): # # "html" - All Unicode characters with corresponding HTML entities # are converted to those entities on output. # "html5" - The same as "html", but empty void tags are represented as # <tag> rather than <tag/> # "minimal" - Bare ampersands and angle brackets are converted to # XML entities: & < > # None - The null formatter. Unicode characters are never # converted to entities. This is not recommended, but it's # faster than "minimal". # A callable function - it will be called on every string that needs to undergo entity substitution. # A Formatter instance - Formatter.substitute(string) will be called on every string that # needs to undergo entity substitution. # # In an HTML document, the default "html", "html5", and "minimal" # functions will leave the contents of <script> and <style> tags # alone. For an XML document, all tags will be given the same # treatment. HTML_FORMATTERS = { "html" : HTMLFormatter(), "html5" : HTML5Formatter(), "minimal" : MinimalHTMLFormatter(), None : None } XML_FORMATTERS = { "html" : HTMLXMLFormatter(), "minimal" : XMLFormatter(), None : None } def format_string(self, s, formatter='minimal'): """Format the given string using the given formatter.""" if isinstance(formatter, str): formatter = self._formatter_for_name(formatter) if formatter is None: output = s else: if callable(formatter): # Backwards compatibility -- you used to pass in a formatting method. output = formatter(s) else: output = formatter.substitute(s) return output @property def _is_xml(self): """Is this element part of an XML tree or an HTML tree? This is used when mapping a formatter name ("minimal") to an appropriate function (one that performs entity-substitution on the contents of <script> and <style> tags, or not). It can be inefficient, but it should be called very rarely. """ if self.known_xml is not None: # Most of the time we will have determined this when the # document is parsed. return self.known_xml # Otherwise, it's likely that this element was created by # direct invocation of the constructor from within the user's # Python code. if self.parent is None: # This is the top-level object. It should have .known_xml set # from tree creation. If not, take a guess--BS is usually # used on HTML markup. return getattr(self, 'is_xml', False) return self.parent._is_xml def _formatter_for_name(self, name): "Look up a formatter function based on its name and the tree." if self._is_xml: return self.XML_FORMATTERS.get(name, XMLFormatter()) else: return self.HTML_FORMATTERS.get(name, HTMLFormatter()) def setup(self, parent=None, previous_element=None, next_element=None, previous_sibling=None, next_sibling=None): """Sets up the initial relations between this element and other elements.""" self.parent = parent self.previous_element = previous_element if previous_element is not None: self.previous_element.next_element = self self.next_element = next_element if self.next_element: self.next_element.previous_element = self self.next_sibling = next_sibling if self.next_sibling: self.next_sibling.previous_sibling = self if (not previous_sibling and self.parent is not None and self.parent.contents): previous_sibling = self.parent.contents[-1] self.previous_sibling = previous_sibling if previous_sibling: self.previous_sibling.next_sibling = self nextSibling = _alias("next_sibling") # BS3 previousSibling = _alias("previous_sibling") # BS3 def replace_with(self, replace_with): if not self.parent: raise ValueError( "Cannot replace one element with another when the" "element to be replaced is not part of a tree.") if replace_with is self: return if replace_with is self.parent: raise ValueError("Cannot replace a Tag with its parent.") old_parent = self.parent my_index = self.parent.index(self) self.extract() old_parent.insert(my_index, replace_with) return self replaceWith = replace_with # BS3 def unwrap(self): my_parent = self.parent if not self.parent: raise ValueError( "Cannot replace an element with its contents when that" "element is not part of a tree.") my_index = self.parent.index(self) self.extract() for child in reversed(self.contents[:]): my_parent.insert(my_index, child) return self replace_with_children = unwrap replaceWithChildren = unwrap # BS3 def wrap(self, wrap_inside): me = self.replace_with(wrap_inside) wrap_inside.append(me) return wrap_inside def extract(self): """Destructively rips this element out of the tree.""" if self.parent is not None: del self.parent.contents[self.parent.index(self)] #Find the two elements that would be next to each other if #this element (and any children) hadn't been parsed. Connect #the two. last_child = self._last_descendant() next_element = last_child.next_element if (self.previous_element is not None and self.previous_element is not next_element): self.previous_element.next_element = next_element if next_element is not None and next_element is not self.previous_element: next_element.previous_element = self.previous_element self.previous_element = None last_child.next_element = None self.parent = None if (self.previous_sibling is not None and self.previous_sibling is not self.next_sibling): self.previous_sibling.next_sibling = self.next_sibling if (self.next_sibling is not None and self.next_sibling is not self.previous_sibling): self.next_sibling.previous_sibling = self.previous_sibling self.previous_sibling = self.next_sibling = None return self def _last_descendant(self, is_initialized=True, accept_self=True): "Finds the last element beneath this object to be parsed." if is_initialized and self.next_sibling: last_child = self.next_sibling.previous_element else: last_child = self while isinstance(last_child, Tag) and last_child.contents: last_child = last_child.contents[-1] if not accept_self and last_child is self: last_child = None return last_child # BS3: Not part of the API! _lastRecursiveChild = _last_descendant def insert(self, position, new_child): if new_child is None: raise ValueError("Cannot insert None into a tag.") if new_child is self: raise ValueError("Cannot insert a tag into itself.") if (isinstance(new_child, str) and not isinstance(new_child, NavigableString)): new_child = NavigableString(new_child) from bs4 import BeautifulSoup if isinstance(new_child, BeautifulSoup): # We don't want to end up with a situation where one BeautifulSoup # object contains another. Insert the children one at a time. for subchild in list(new_child.contents): self.insert(position, subchild) position += 1 return position = min(position, len(self.contents)) if hasattr(new_child, 'parent') and new_child.parent is not None: # We're 'inserting' an element that's already one # of this object's children. if new_child.parent is self: current_index = self.index(new_child) if current_index < position: # We're moving this element 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 -= 1 new_child.extract() new_child.parent = self previous_child = None if position == 0: new_child.previous_sibling = None new_child.previous_element = self else: previous_child = self.contents[position - 1] new_child.previous_sibling = previous_child new_child.previous_sibling.next_sibling = new_child new_child.previous_element = previous_child._last_descendant(False) if new_child.previous_element is not None: new_child.previous_element.next_element = new_child new_childs_last_element = new_child._last_descendant(False) if position >= len(self.contents): new_child.next_sibling = None parent = self parents_next_sibling = None while parents_next_sibling is None and parent is not None: parents_next_sibling = parent.next_sibling parent = parent.parent if parents_next_sibling is not None: # We found the element that comes next in the document. break if parents_next_sibling is not None: new_childs_last_element.next_element = parents_next_sibling else: # The last element of this tag is the last element in # the document. new_childs_last_element.next_element = None else: next_child = self.contents[position] new_child.next_sibling = next_child if new_child.next_sibling is not None: new_child.next_sibling.previous_sibling = new_child new_childs_last_element.next_element = next_child if new_childs_last_element.next_element is not None: new_childs_last_element.next_element.previous_element = new_childs_last_element self.contents.insert(position, new_child) def append(self, tag): """Appends the given tag to the contents of this tag.""" self.insert(len(self.contents), tag) def insert_before(self, predecessor): """Makes the given element the immediate predecessor of this one. The two elements will have the same parent, and the given element will be immediately before this one. """ if self is predecessor: raise ValueError("Can't insert an element before itself.") parent = self.parent if parent is None: raise ValueError( "Element has no parent, so 'before' has no meaning.") # Extract first so that the index won't be screwed up if they # are siblings. if isinstance(predecessor, PageElement): predecessor.extract() index = parent.index(self) parent.insert(index, predecessor) def insert_after(self, successor): """Makes the given element the immediate successor of this one. The two elements will have the same parent, and the given element will be immediately after this one. """ if self is successor: raise ValueError("Can't insert an element after itself.") parent = self.parent if parent is None: raise ValueError( "Element has no parent, so 'after' has no meaning.") # Extract first so that the index won't be screwed up if they # are siblings. if isinstance(successor, PageElement): successor.extract() index = parent.index(self) parent.insert(index+1, successor) def find_next(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._find_one(self.find_all_next, name, attrs, text, **kwargs) findNext = find_next # BS3 def find_all_next(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._find_all(name, attrs, text, limit, self.next_elements, **kwargs) findAllNext = find_all_next # BS3 def find_next_sibling(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._find_one(self.find_next_siblings, name, attrs, text, **kwargs) findNextSibling = find_next_sibling # BS3 def find_next_siblings(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._find_all(name, attrs, text, limit, self.next_siblings, **kwargs) findNextSiblings = find_next_siblings # BS3 fetchNextSiblings = find_next_siblings # BS2 def find_previous(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._find_one( self.find_all_previous, name, attrs, text, **kwargs) findPrevious = find_previous # BS3 def find_all_previous(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._find_all(name, attrs, text, limit, self.previous_elements, **kwargs) findAllPrevious = find_all_previous # BS3 fetchPrevious = find_all_previous # BS2 def find_previous_sibling(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._find_one(self.find_previous_siblings, name, attrs, text, **kwargs) findPreviousSibling = find_previous_sibling # BS3 def find_previous_siblings(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._find_all(name, attrs, text, limit, self.previous_siblings, **kwargs) findPreviousSiblings = find_previous_siblings # BS3 fetchPreviousSiblings = find_previous_siblings # BS2 def find_parent(self, name=None, attrs={}, **kwargs): """Returns the closest parent of this Tag that matches the given criteria.""" # NOTE: We can't use _find_one because findParents takes a different # set of arguments. r = None l = self.find_parents(name, attrs, 1, **kwargs) if l: r = l[0] return r findParent = find_parent # BS3 def find_parents(self, name=None, attrs={}, limit=None, **kwargs): """Returns the parents of this Tag that match the given criteria.""" return self._find_all(name, attrs, None, limit, self.parents, **kwargs) findParents = find_parents # BS3 fetchParents = find_parents # BS2 @property def next(self): return self.next_element @property def previous(self): return self.previous_element #These methods do the real heavy lifting. def _find_one(self, method, name, attrs, text, **kwargs): r = None l = method(name, attrs, text, 1, **kwargs) if l: r = l[0] return r def _find_all(self, name, attrs, text, limit, generator, **kwargs): "Iterates over a generator looking for things that match." if text is None and 'string' in kwargs: text = kwargs['string'] del kwargs['string'] if isinstance(name, SoupStrainer): strainer = name else: strainer = SoupStrainer(name, attrs, text, **kwargs) if text is None and not limit and not attrs and not kwargs: if name is True or name is None: # Optimization to find all tags. result = (element for element in generator if isinstance(element, Tag)) return ResultSet(strainer, result) elif isinstance(name, str): # Optimization to find all tags with a given name. if name.count(':') == 1: # This is a name with a prefix. If this is a namespace-aware document, # we need to match the local name against tag.name. If not, # we need to match the fully-qualified name against tag.name. prefix, local_name = name.split(':', 1) else: prefix = None local_name = name result = (element for element in generator if isinstance(element, Tag) and ( element.name == name ) or ( element.name == local_name and (prefix is None or element.prefix == prefix) ) ) return ResultSet(strainer, result) results = ResultSet(strainer) while True: try: i = next(generator) 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. @property def next_elements(self): i = self.next_element while i is not None: yield i i = i.next_element @property def next_siblings(self): i = self.next_sibling while i is not None: yield i i = i.next_sibling @property def previous_elements(self): i = self.previous_element while i is not None: yield i i = i.previous_element @property def previous_siblings(self): i = self.previous_sibling while i is not None: yield i i = i.previous_sibling @property def parents(self): i = self.parent while i is not None: yield i i = i.parent # Methods for supporting CSS selectors. tag_name_re = re.compile('^[a-zA-Z0-9][-.a-zA-Z0-9:_]*$') # /^([a-zA-Z0-9][-.a-zA-Z0-9:_]*)\[(\w+)([=~\|\^\$\*]?)=?"?([^\]"]*)"?\]$/ # \---------------------------/ \---/\-------------/ \-------/ # | | | | # | | | The value # | | ~,|,^,$,* or = # | Attribute # Tag attribselect_re = re.compile( r'^(?P<tag>[a-zA-Z0-9][-.a-zA-Z0-9:_]*)?\[(?P<attribute>[\w-]+)(?P<operator>[=~\|\^\$\*]?)' + r'=?"?(?P<value>[^\]"]*)"?\]$' ) def _attr_value_as_string(self, value, default=None): """Force an attribute value into a string representation. A multi-valued attribute will be converted into a space-separated stirng. """ value = self.get(value, default) if isinstance(value, list) or isinstance(value, tuple): value =" ".join(value) return value def _tag_name_matches_and(self, function, tag_name): if not tag_name: return function else: def _match(tag): return tag.name == tag_name and function(tag) return _match def _attribute_checker(self, operator, attribute, value=''): """Create a function that performs a CSS selector operation. Takes an operator, attribute and optional value. Returns a function that will return True for elements that match that combination. """ if operator == '=': # string representation of `attribute` is equal to `value` return lambda el: el._attr_value_as_string(attribute) == value elif operator == '~': # space-separated list representation of `attribute` # contains `value` def _includes_value(element): attribute_value = element.get(attribute, []) if not isinstance(attribute_value, list): attribute_value = attribute_value.split() return value in attribute_value return _includes_value elif operator == '^': # string representation of `attribute` starts with `value` return lambda el: el._attr_value_as_string( attribute, '').startswith(value) elif operator == '$': # string representation of `attribute` ends with `value` return lambda el: el._attr_value_as_string( attribute, '').endswith(value) elif operator == '*': # string representation of `attribute` contains `value` return lambda el: value in el._attr_value_as_string(attribute, '') elif operator == '|': # string representation of `attribute` is either exactly # `value` or starts with `value` and then a dash. def _is_or_starts_with_dash(element): attribute_value = element._attr_value_as_string(attribute, '') return (attribute_value == value or attribute_value.startswith( value + '-')) return _is_or_starts_with_dash else: return lambda el: el.has_attr(attribute) # Old non-property versions of the generators, for backwards # compatibility with BS3. def nextGenerator(self): return self.next_elements def nextSiblingGenerator(self): return self.next_siblings def previousGenerator(self): return self.previous_elements def previousSiblingGenerator(self): return self.previous_siblings def parentGenerator(self): return self.parents class NavigableString(str, PageElement): PREFIX = '' SUFFIX = '' # We can't tell just by looking at a string whether it's contained # in an XML document or an HTML document. known_xml = None 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, str): u = str.__new__(cls, value) else: u = str.__new__(cls, value, DEFAULT_OUTPUT_ENCODING) u.setup() return u def __copy__(self): """A copy of a NavigableString has the same contents and class as the original, but it is not connected to the parse tree. """ return type(self)(self) def __getnewargs__(self): return (str(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 output_ready(self, formatter="minimal"): output = self.format_string(self, formatter) return self.PREFIX + output + self.SUFFIX @property def name(self): return None @name.setter def name(self, name): raise AttributeError("A NavigableString cannot be given a name.") class PreformattedString(NavigableString): """A NavigableString not subject to the normal formatting rules. The string will be passed into the formatter (to trigger side effects), but the return value will be ignored. """ def output_ready(self, formatter="minimal"): """CData strings are passed into the formatter. But the return value is ignored.""" self.format_string(self, formatter) return self.PREFIX + self + self.SUFFIX class CData(PreformattedString): PREFIX = '<![CDATA[' SUFFIX = ']]>' class ProcessingInstruction(PreformattedString): """A SGML processing instruction.""" PREFIX = '<?' SUFFIX = '>' class XMLProcessingInstruction(ProcessingInstruction): """An XML processing instruction.""" PREFIX = '<?' SUFFIX = '?>' class Comment(PreformattedString): PREFIX = '<!--' SUFFIX = '-->' class Declaration(PreformattedString): PREFIX = '<?' SUFFIX = '?>' class Doctype(PreformattedString): @classmethod def for_name_and_ids(cls, name, pub_id, system_id): value = name or '' if pub_id is not None: value += ' PUBLIC "%s"' % pub_id if system_id is not None: value += ' "%s"' % system_id elif system_id is not None: value += ' SYSTEM "%s"' % system_id return Doctype(value) PREFIX = '<!DOCTYPE ' SUFFIX = '>\n' class Tag(PageElement): """Represents a found HTML tag with its attributes and contents.""" def __init__(self, parser=None, builder=None, name=None, namespace=None, prefix=None, attrs=None, parent=None, previous=None, is_xml=None): "Basic constructor." if parser is None: self.parser_class = None else: # We don't actually store the parser object: that lets extracted # chunks be garbage-collected. self.parser_class = parser.__class__ if name is None: raise ValueError("No value provided for new tag's name.") self.name = name self.namespace = namespace self.prefix = prefix if builder is not None: preserve_whitespace_tags = builder.preserve_whitespace_tags else: if is_xml: preserve_whitespace_tags = [] else: preserve_whitespace_tags = HTMLAwareEntitySubstitution.preserve_whitespace_tags self.preserve_whitespace_tags = preserve_whitespace_tags if attrs is None: attrs = {} elif attrs: if builder is not None and builder.cdata_list_attributes: attrs = builder._replace_cdata_list_attribute_values( self.name, attrs) else: attrs = dict(attrs) else: attrs = dict(attrs) # If possible, determine ahead of time whether this tag is an # XML tag. if builder: self.known_xml = builder.is_xml else: self.known_xml = is_xml self.attrs = attrs self.contents = [] self.setup(parent, previous) self.hidden = False # Set up any substitutions, such as the charset in a META tag. if builder is not None: builder.set_up_substitutions(self) self.can_be_empty_element = builder.can_be_empty_element(name) else: self.can_be_empty_element = False parserClass = _alias("parser_class") # BS3 def __copy__(self): """A copy of a Tag is a new Tag, unconnected to the parse tree. Its contents are a copy of the old Tag's contents. """ clone = type(self)(None, self.builder, self.name, self.namespace, self.prefix, self.attrs, is_xml=self._is_xml) for attr in ('can_be_empty_element', 'hidden'): setattr(clone, attr, getattr(self, attr)) for child in self.contents: clone.append(child.__copy__()) return clone @property def is_empty_element(self): """Is this tag an empty-element tag? (aka a self-closing tag) A tag that has contents is never an empty-element tag. A tag that has no contents may or may not be an empty-element tag. It depends on the builder used to create the tag. If the builder has a designated list of empty-element tags, then only a tag whose name shows up in that list is considered an empty-element tag. If the builder has no designated list of empty-element tags, then any tag with no contents is an empty-element tag. """ return len(self.contents) == 0 and self.can_be_empty_element isSelfClosing = is_empty_element # BS3 @property def string(self): """Convenience property to get the single string within this tag. :Return: If this tag has a single string child, return value is that string. If this tag has no children, or more than one child, return value is None. If this tag has one child tag, return value is the 'string' attribute of the child tag, recursively. """ if len(self.contents) != 1: return None child = self.contents[0] if isinstance(child, NavigableString): return child return child.string @string.setter def string(self, string): self.clear() self.append(string.__class__(string)) def _all_strings(self, strip=False, types=(NavigableString, CData)): """Yield all strings of certain classes, possibly stripping them. By default, yields only NavigableString and CData objects. So no comments, processing instructions, etc. """ for descendant in self.descendants: if ( (types is None and not isinstance(descendant, NavigableString)) or (types is not None and type(descendant) not in types)): continue if strip: descendant = descendant.strip() if len(descendant) == 0: continue yield descendant strings = property(_all_strings) @property def stripped_strings(self): for string in self._all_strings(True): yield string def get_text(self, separator="", strip=False, types=(NavigableString, CData)): """ Get all child strings, concatenated using the given separator. """ return separator.join([s for s in self._all_strings( strip, types=types)]) getText = get_text text = property(get_text) def decompose(self): """Recursively destroys the contents of this tree.""" self.extract() i = self while i is not None: next = i.next_element i.__dict__.clear() i.contents = [] i = next def clear(self, decompose=False): """ Extract all children. If decompose is True, decompose instead. """ if decompose: for element in self.contents[:]: if isinstance(element, Tag): element.decompose() else: element.extract() else: for element in self.contents[:]: element.extract() def index(self, element): """ Find the index of a child by identity, not value. Avoids issues with tag.contents.index(element) getting the index of equal elements. """ for i, child in enumerate(self.contents): if child is element: return i raise ValueError("Tag.index: element not in tag") 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.attrs.get(key, default) def get_attribute_list(self, key, default=None): """The same as get(), but always returns a list.""" value = self.get(key, default) if not isinstance(value, list): value = [value] return value def has_attr(self, key): return key in self.attrs def __hash__(self): return str(self).__hash__() 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.attrs[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 __bool__(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.attrs[key] = value def __delitem__(self, key): "Deleting tag[key] deletes all 'key' attributes for the tag." self.attrs.pop(key, None) def __call__(self, *args, **kwargs): """Calling a tag like a function is the same as calling its find_all() method. Eg. tag('a') returns a list of all the A tags found within this tag.""" return self.find_all(*args, **kwargs) def __getattr__(self, tag): #print "Getattr %s.%s" % (self.__class__, tag) if len(tag) > 3 and tag.endswith('Tag'): # BS3: soup.aTag -> "soup.find("a") tag_name = tag[:-3] warnings.warn( '.%(name)sTag is deprecated, use .find("%(name)s") instead. If you really were looking for a tag called %(name)sTag, use .find("%(name)sTag")' % dict( name=tag_name ) ) return self.find(tag_name) # We special case contents to avoid recursion. elif not tag.startswith("__") and not tag == "contents": 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.""" if self is other: return True 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, my_child in enumerate(self.contents): if my_child != 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="unicode-escape"): """Renders this tag as a string.""" if PY3K: # "The return value must be a string object", i.e. Unicode return self.decode() else: # "The return value must be a string object", i.e. a bytestring. # By convention, the return value of __repr__ should also be # an ASCII string. return self.encode(encoding) def __unicode__(self): return self.decode() def __str__(self): if PY3K: return self.decode() else: return self.encode() if PY3K: __str__ = __repr__ = __unicode__ def encode(self, encoding=DEFAULT_OUTPUT_ENCODING, indent_level=None, formatter="minimal", errors="xmlcharrefreplace"): # Turn the data structure into Unicode, then encode the # Unicode. u = self.decode(indent_level, encoding, formatter) return u.encode(encoding, errors) def _should_pretty_print(self, indent_level): """Should this tag be pretty-printed?""" return ( indent_level is not None and self.name not in self.preserve_whitespace_tags ) def decode(self, indent_level=None, eventual_encoding=DEFAULT_OUTPUT_ENCODING, formatter="minimal"): """Returns a Unicode representation of this tag and its contents. :param eventual_encoding: The tag is destined to be encoded into this encoding. This method is _not_ responsible for performing that encoding. This information is passed in so that it can be substituted in if the document contains a <META> tag that mentions the document's encoding. """ # First off, turn a string formatter into a Formatter object. This # will stop the lookup from happening over and over again. if not isinstance(formatter, Formatter) and not callable(formatter): formatter = self._formatter_for_name(formatter) attrs = [] if self.attrs: for key, val in sorted(self.attrs.items()): if val is None: decoded = key else: if isinstance(val, list) or isinstance(val, tuple): val = ' '.join(val) elif not isinstance(val, str): val = str(val) elif ( isinstance(val, AttributeValueWithCharsetSubstitution) and eventual_encoding is not None): val = val.encode(eventual_encoding) text = self.format_string(val, formatter) decoded = ( str(key) + '=' + EntitySubstitution.quoted_attribute_value(text)) attrs.append(decoded) close = '' closeTag = '' prefix = '' if self.prefix: prefix = self.prefix + ":" if self.is_empty_element: close = '' if isinstance(formatter, Formatter): close = formatter.void_element_close_prefix or close else: closeTag = '</%s%s>' % (prefix, self.name) pretty_print = self._should_pretty_print(indent_level) space = '' indent_space = '' if indent_level is not None: indent_space = (' ' * (indent_level - 1)) if pretty_print: space = indent_space indent_contents = indent_level + 1 else: indent_contents = None contents = self.decode_contents( indent_contents, eventual_encoding, formatter) if self.hidden: # This is the 'document root' object. s = contents else: s = [] attribute_string = '' if attrs: attribute_string = ' ' + ' '.join(attrs) if indent_level is not None: # Even if this particular tag is not pretty-printed, # we should indent up to the start of the tag. s.append(indent_space) s.append('<%s%s%s%s>' % ( prefix, self.name, attribute_string, close)) if pretty_print: s.append("\n") s.append(contents) if pretty_print and contents and contents[-1] != "\n": s.append("\n") if pretty_print and closeTag: s.append(space) s.append(closeTag) if indent_level is not None and closeTag and self.next_sibling: # Even if this particular tag is not pretty-printed, # we're now done with the tag, and we should add a # newline if appropriate. s.append("\n") s = ''.join(s) return s def prettify(self, encoding=None, formatter="minimal"): if encoding is None: return self.decode(True, formatter=formatter) else: return self.encode(encoding, True, formatter=formatter) def decode_contents(self, indent_level=None, eventual_encoding=DEFAULT_OUTPUT_ENCODING, formatter="minimal"): """Renders the contents of this tag as a Unicode string. :param indent_level: Each line of the rendering will be indented this many spaces. :param eventual_encoding: The tag is destined to be encoded into this encoding. This method is _not_ responsible for performing that encoding. This information is passed in so that it can be substituted in if the document contains a <META> tag that mentions the document's encoding. :param formatter: The output formatter responsible for converting entities to Unicode characters. """ # First off, turn a string formatter into a Formatter object. This # will stop the lookup from happening over and over again. if not isinstance(formatter, Formatter) and not callable(formatter): formatter = self._formatter_for_name(formatter) pretty_print = (indent_level is not None) s = [] for c in self: text = None if isinstance(c, NavigableString): text = c.output_ready(formatter) elif isinstance(c, Tag): s.append(c.decode(indent_level, eventual_encoding, formatter)) if text and indent_level and not self.name == 'pre': text = text.strip() if text: if pretty_print and not self.name == 'pre': s.append(" " * (indent_level - 1)) s.append(text) if pretty_print and not self.name == 'pre': s.append("\n") return ''.join(s) def encode_contents( self, indent_level=None, encoding=DEFAULT_OUTPUT_ENCODING, formatter="minimal"): """Renders the contents of this tag as a bytestring. :param indent_level: Each line of the rendering will be indented this many spaces. :param eventual_encoding: The bytestring will be in this encoding. :param formatter: The output formatter responsible for converting entities to Unicode characters. """ contents = self.decode_contents(indent_level, encoding, formatter) return contents.encode(encoding) # Old method for BS3 compatibility def renderContents(self, encoding=DEFAULT_OUTPUT_ENCODING, prettyPrint=False, indentLevel=0): if not prettyPrint: indentLevel = None return self.encode_contents( indent_level=indentLevel, encoding=encoding) #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.find_all(name, attrs, recursive, text, 1, **kwargs) if l: r = l[0] return r findChild = find def find_all(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.descendants if not recursive: generator = self.children return self._find_all(name, attrs, text, limit, generator, **kwargs) findAll = find_all # BS3 findChildren = find_all # BS2 #Generator methods @property def children(self): # return iter() to make the purpose of the method clear return iter(self.contents) # XXX This seems to be untested. @property def descendants(self): if not len(self.contents): return stopNode = self._last_descendant().next_element current = self.contents[0] while current is not stopNode: yield current current = current.next_element # CSS selector code _selector_combinators = ['>', '+', '~'] _select_debug = False quoted_colon = re.compile('"[^"]*:[^"]*"') def select_one(self, selector): """Perform a CSS selection operation on the current element.""" value = self.select(selector, limit=1) if value: return value[0] return None def select(self, selector, _candidate_generator=None, limit=None): """Perform a CSS selection operation on the current element.""" # Handle grouping selectors if ',' exists, ie: p,a if ',' in selector: context = [] selectors = [x.strip() for x in selector.split(",")] # If a selector is mentioned multiple times we don't want # to use it more than once. used_selectors = set() # We also don't want to select the same element more than once, # if it's matched by multiple selectors. selected_object_ids = set() for partial_selector in selectors: if partial_selector == '': raise ValueError('Invalid group selection syntax: %s' % selector) if partial_selector in used_selectors: continue used_selectors.add(partial_selector) candidates = self.select(partial_selector, limit=limit) for candidate in candidates: # This lets us distinguish between distinct tags that # represent the same markup. object_id = id(candidate) if object_id not in selected_object_ids: context.append(candidate) selected_object_ids.add(object_id) if limit and len(context) >= limit: break return context tokens = shlex.split(selector) current_context = [self] if tokens[-1] in self._selector_combinators: raise ValueError( 'Final combinator "%s" is missing an argument.' % tokens[-1]) if self._select_debug: print('Running CSS selector "%s"' % selector) for index, token in enumerate(tokens): new_context = [] new_context_ids = set([]) if tokens[index-1] in self._selector_combinators: # This token was consumed by the previous combinator. Skip it. if self._select_debug: print(' Token was consumed by the previous combinator.') continue if self._select_debug: print(' Considering token "%s"' % token) recursive_candidate_generator = None tag_name = None # Each operation corresponds to a checker function, a rule # for determining whether a candidate matches the # selector. Candidates are generated by the active # iterator. checker = None m = self.attribselect_re.match(token) if m is not None: # Attribute selector tag_name, attribute, operator, value = m.groups() checker = self._attribute_checker(operator, attribute, value) elif '#' in token: # ID selector tag_name, tag_id = token.split('#', 1) def id_matches(tag): return tag.get('id', None) == tag_id checker = id_matches elif '.' in token: # Class selector tag_name, klass = token.split('.', 1) classes = set(klass.split('.')) def classes_match(candidate): return classes.issubset(candidate.get('class', [])) checker = classes_match elif ':' in token and not self.quoted_colon.search(token): # Pseudo-class tag_name, pseudo = token.split(':', 1) if tag_name == '': raise ValueError( "A pseudo-class must be prefixed with a tag name.") pseudo_attributes = re.match(r'([a-zA-Z\d-]+)\(([a-zA-Z\d]+)\)', pseudo) found = [] if pseudo_attributes is None: pseudo_type = pseudo pseudo_value = None else: pseudo_type, pseudo_value = pseudo_attributes.groups() if pseudo_type == 'nth-of-type': try: pseudo_value = int(pseudo_value) except: raise NotImplementedError( 'Only numeric values are currently supported for the nth-of-type pseudo-class.') if pseudo_value < 1: raise ValueError( 'nth-of-type pseudo-class value must be at least 1.') class Counter(object): def __init__(self, destination): self.count = 0 self.destination = destination def nth_child_of_type(self, tag): self.count += 1 if self.count == self.destination: return True else: return False checker = Counter(pseudo_value).nth_child_of_type else: raise NotImplementedError( 'Only the following pseudo-classes are implemented: nth-of-type.') elif token == '*': # Star selector -- matches everything pass elif token == '>': # Run the next token as a CSS selector against the # direct children of each tag in the current context. recursive_candidate_generator = lambda tag: tag.children elif token == '~': # Run the next token as a CSS selector against the # siblings of each tag in the current context. recursive_candidate_generator = lambda tag: tag.next_siblings elif token == '+': # For each tag in the current context, run the next # token as a CSS selector against the tag's next # sibling that's a tag. def next_tag_sibling(tag): yield tag.find_next_sibling(True) recursive_candidate_generator = next_tag_sibling elif self.tag_name_re.match(token): # Just a tag name. tag_name = token else: raise ValueError( 'Unsupported or invalid CSS selector: "%s"' % token) if recursive_candidate_generator: # This happens when the selector looks like "> foo". # # The generator calls select() recursively on every # member of the current context, passing in a different # candidate generator and a different selector. # # In the case of "> foo", the candidate generator is # one that yields a tag's direct children (">"), and # the selector is "foo". next_token = tokens[index+1] def recursive_select(tag): if self._select_debug: print(' Calling select("%s") recursively on %s %s' % (next_token, tag.name, tag.attrs)) print('-' * 40) for i in tag.select(next_token, recursive_candidate_generator): if self._select_debug: print('(Recursive select picked up candidate %s %s)' % (i.name, i.attrs)) yield i if self._select_debug: print('-' * 40) _use_candidate_generator = recursive_select elif _candidate_generator is None: # By default, a tag's candidates are all of its # children. If tag_name is defined, only yield tags # with that name. if self._select_debug: if tag_name: check = "[any]" else: check = tag_name print(' Default candidate generator, tag name="%s"' % check) if self._select_debug: # This is redundant with later code, but it stops # a bunch of bogus tags from cluttering up the # debug log. def default_candidate_generator(tag): for child in tag.descendants: if not isinstance(child, Tag): continue if tag_name and not child.name == tag_name: continue yield child _use_candidate_generator = default_candidate_generator else: _use_candidate_generator = lambda tag: tag.descendants else: _use_candidate_generator = _candidate_generator count = 0 for tag in current_context: if self._select_debug: print(" Running candidate generator on %s %s" % ( tag.name, repr(tag.attrs))) for candidate in _use_candidate_generator(tag): if not isinstance(candidate, Tag): continue if tag_name and candidate.name != tag_name: continue if checker is not None: try: result = checker(candidate) except StopIteration: # The checker has decided we should no longer # run the generator. break if checker is None or result: if self._select_debug: print(" SUCCESS %s %s" % (candidate.name, repr(candidate.attrs))) if id(candidate) not in new_context_ids: # If a tag matches a selector more than once, # don't include it in the context more than once. new_context.append(candidate) new_context_ids.add(id(candidate)) elif self._select_debug: print(" FAILURE %s %s" % (candidate.name, repr(candidate.attrs))) current_context = new_context if limit and len(current_context) >= limit: current_context = current_context[:limit] if self._select_debug: print("Final verdict:") for i in current_context: print(" %s %s" % (i.name, i.attrs)) return current_context # Old names for backwards compatibility def childGenerator(self): return self.children def recursiveChildGenerator(self): return self.descendants def has_key(self, key): """This was kind of misleading because has_key() (attributes) was different from __in__ (contents). has_key() is gone in Python 3, anyway.""" warnings.warn('has_key is deprecated. Use has_attr("%s") instead.' % ( key)) return self.has_attr(key) # Next, a couple classes to represent queries and their results. class SoupStrainer(object): """Encapsulates a number of ways of matching a markup element (tag or text).""" def __init__(self, name=None, attrs={}, text=None, **kwargs): self.name = self._normalize_search_value(name) if not isinstance(attrs, dict): # Treat a non-dict value for attrs as a search for the 'class' # attribute. kwargs['class'] = attrs attrs = None if 'class_' in kwargs: # Treat class_="foo" as a search for the 'class' # attribute, overriding any non-dict value for attrs. kwargs['class'] = kwargs['class_'] del kwargs['class_'] if kwargs: if attrs: attrs = attrs.copy() attrs.update(kwargs) else: attrs = kwargs normalized_attrs = {} for key, value in list(attrs.items()): normalized_attrs[key] = self._normalize_search_value(value) self.attrs = normalized_attrs self.text = self._normalize_search_value(text) def _normalize_search_value(self, value): # Leave it alone if it's a Unicode string, a callable, a # regular expression, a boolean, or None. if (isinstance(value, str) or callable(value) or hasattr(value, 'match') or isinstance(value, bool) or value is None): return value # If it's a bytestring, convert it to Unicode, treating it as UTF-8. if isinstance(value, bytes): return value.decode("utf8") # If it's listlike, convert it into a list of strings. if hasattr(value, '__iter__'): new_value = [] for v in value: if (hasattr(v, '__iter__') and not isinstance(v, bytes) and not isinstance(v, str)): # This is almost certainly the user's mistake. In the # interests of avoiding infinite loops, we'll let # it through as-is rather than doing a recursive call. new_value.append(v) else: new_value.append(self._normalize_search_value(v)) return new_value # Otherwise, convert it into a Unicode string. # The unicode(str()) thing is so this will do the same thing on Python 2 # and Python 3. return str(str(value)) def __str__(self): if self.text: return self.text else: return "%s|%s" % (self.name, self.attrs) def search_tag(self, markup_name=None, markup_attrs={}): found = None markup = None if isinstance(markup_name, Tag): markup = markup_name markup_attrs = markup call_function_with_tag_data = ( isinstance(self.name, Callable) and not isinstance(markup_name, Tag)) if ((not self.name) or call_function_with_tag_data or (markup and self._matches(markup, self.name)) or (not markup and self._matches(markup_name, self.name))): if call_function_with_tag_data: match = self.name(markup_name, markup_attrs) else: match = True markup_attr_map = None for attr, match_against in list(self.attrs.items()): if not markup_attr_map: if hasattr(markup_attrs, 'get'): markup_attr_map = markup_attrs else: markup_attr_map = {} for k, v in markup_attrs: markup_attr_map[k] = v attr_value = markup_attr_map.get(attr) if not self._matches(attr_value, match_against): match = False break if match: if markup: found = markup else: found = markup_name if found and self.text and not self._matches(found.string, self.text): found = None return found searchTag = search_tag 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 hasattr(markup, '__iter__') and not isinstance(markup, (Tag, str)): 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 or self.name or self.attrs: found = self.search_tag(markup) # If it's text, make sure the text matches. elif isinstance(markup, NavigableString) or \ isinstance(markup, str): if not self.name and not self.attrs and 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, match_against, already_tried=None): # print u"Matching %s against %s" % (markup, match_against) result = False if isinstance(markup, list) or isinstance(markup, tuple): # This should only happen when searching a multi-valued attribute # like 'class'. for item in markup: if self._matches(item, match_against): return True # We didn't match any particular value of the multivalue # attribute, but maybe we match the attribute value when # considered as a string. if self._matches(' '.join(markup), match_against): return True return False if match_against is True: # True matches any non-None value. return markup is not None if isinstance(match_against, Callable): return match_against(markup) # Custom callables take the tag as an argument, but all # other ways of matching match the tag name as a string. original_markup = markup if isinstance(markup, Tag): markup = markup.name # Ensure that `markup` is either a Unicode string, or None. markup = self._normalize_search_value(markup) if markup is None: # None matches None, False, an empty string, an empty list, and so on. return not match_against if (hasattr(match_against, '__iter__') and not isinstance(match_against, str)): # We're asked to match against an iterable of items. # The markup must be match at least one item in the # iterable. We'll try each one in turn. # # To avoid infinite recursion we need to keep track of # items we've already seen. if not already_tried: already_tried = set() for item in match_against: if item.__hash__: key = item else: key = id(item) if key in already_tried: continue else: already_tried.add(key) if self._matches(original_markup, item, already_tried): return True else: return False # Beyond this point we might need to run the test twice: once against # the tag's name and once against its prefixed name. match = False if not match and isinstance(match_against, str): # Exact string match match = markup == match_against if not match and hasattr(match_against, 'search'): # Regexp match return match_against.search(markup) if (not match and isinstance(original_markup, Tag) and original_markup.prefix): # Try the whole thing again with the prefixed tag name. return self._matches( original_markup.prefix + ':' + original_markup.name, match_against ) return match class ResultSet(list): """A ResultSet is just a list that keeps track of the SoupStrainer that created it.""" def __init__(self, source, result=()): super(ResultSet, self).__init__(result) self.source = source def __getattr__(self, key): raise AttributeError( "ResultSet object has no attribute '%s'. You're probably treating a list of items like a single item. Did you call find_all() when you meant to call find()?" % key )
