Server IP : 103.119.228.120 / Your IP : 3.143.241.253 Web Server : Apache System : Linux v8.techscape8.com 3.10.0-1160.119.1.el7.tuxcare.els2.x86_64 #1 SMP Mon Jul 15 12:09:18 UTC 2024 x86_64 User : nobody ( 99) PHP Version : 5.6.40 Disable Function : shell_exec,symlink,system,exec,proc_get_status,proc_nice,proc_terminate,define_syslog_variables,syslog,openlog,closelog,escapeshellcmd,passthru,ocinum cols,ini_alter,leak,listen,chgrp,apache_note,apache_setenv,debugger_on,debugger_off,ftp_exec,dl,dll,myshellexec,proc_open,socket_bind,proc_close,escapeshellarg,parse_ini_filepopen,fpassthru,exec,passthru,escapeshellarg,escapeshellcmd,proc_close,proc_open,ini_alter,popen,show_source,proc_nice,proc_terminate,proc_get_status,proc_close,pfsockopen,leak,apache_child_terminate,posix_kill,posix_mkfifo,posix_setpgid,posix_setsid,posix_setuid,dl,symlink,shell_exec,system,dl,passthru,escapeshellarg,escapeshellcmd,myshellexec,c99_buff_prepare,c99_sess_put,fpassthru,getdisfunc,fx29exec,fx29exec2,is_windows,disp_freespace,fx29sh_getupdate,fx29_buff_prepare,fx29_sess_put,fx29shexit,fx29fsearch,fx29ftpbrutecheck,fx29sh_tools,fx29sh_about,milw0rm,imagez,sh_name,myshellexec,checkproxyhost,dosyayicek,c99_buff_prepare,c99_sess_put,c99getsource,c99sh_getupdate,c99fsearch,c99shexit,view_perms,posix_getpwuid,posix_getgrgid,posix_kill,parse_perms,parsesort,view_perms_color,set_encoder_input,ls_setcheckboxall,ls_reverse_all,rsg_read,rsg_glob,selfURL,dispsecinfo,unix2DosTime,addFile,system,get_users,view_size,DirFiles,DirFilesWide,DirPrintHTMLHeaders,GetFilesTotal,GetTitles,GetTimeTotal,GetMatchesCount,GetFileMatchesCount,GetResultFiles,fs_copy_dir,fs_copy_obj,fs_move_dir,fs_move_obj,fs_rmdir,SearchText,getmicrotime MySQL : ON | cURL : ON | WGET : ON | Perl : ON | Python : ON | Sudo : ON | Pkexec : ON Directory : /lib64/python2.7/site-packages/lxml/html/ |
Upload File : |
# Copyright (c) 2004 Ian Bicking. All rights reserved. # # Redistribution and use in source and binary forms, with or without # modification, are permitted provided that the following conditions are # met: # # 1. Redistributions of source code must retain the above copyright # notice, this list of conditions and the following disclaimer. # # 2. 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. # # 3. Neither the name of Ian Bicking 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 IAN BICKING 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. """The ``lxml.html`` tool set for HTML handling. """ import sys import re try: from urlparse import urljoin except ImportError: # Python 3 from urllib.parse import urljoin import copy from lxml import etree from lxml.html import defs from lxml.html._setmixin import SetMixin try: from collections import MutableMapping as DictMixin except ImportError: # Python < 2.6 from UserDict import DictMixin try: set except NameError: # Python 2.3 from sets import Set as set try: bytes except NameError: # Python < 2.6 bytes = str try: unicode except NameError: # Python 3 unicode = str try: basestring except NameError: # Python 3 basestring = (str, bytes) def __fix_docstring(s): if not s: return s import sys if sys.version_info[0] >= 3: sub = re.compile(r"^(\s*)u'", re.M).sub else: sub = re.compile(r"^(\s*)b'", re.M).sub return sub(r"\1'", s) __all__ = [ 'document_fromstring', 'fragment_fromstring', 'fragments_fromstring', 'fromstring', 'tostring', 'Element', 'defs', 'open_in_browser', 'submit_form', 'find_rel_links', 'find_class', 'make_links_absolute', 'resolve_base_href', 'iterlinks', 'rewrite_links', 'open_in_browser', 'parse'] XHTML_NAMESPACE = "http://www.w3.org/1999/xhtml" _rel_links_xpath = etree.XPath("descendant-or-self::a[@rel]|descendant-or-self::x:a[@rel]", namespaces={'x':XHTML_NAMESPACE}) _options_xpath = etree.XPath("descendant-or-self::option|descendant-or-self::x:option", namespaces={'x':XHTML_NAMESPACE}) _forms_xpath = etree.XPath("descendant-or-self::form|descendant-or-self::x:form", namespaces={'x':XHTML_NAMESPACE}) #_class_xpath = etree.XPath(r"descendant-or-self::*[regexp:match(@class, concat('\b', $class_name, '\b'))]", {'regexp': 'http://exslt.org/regular-expressions'}) _class_xpath = etree.XPath("descendant-or-self::*[@class and contains(concat(' ', normalize-space(@class), ' '), concat(' ', $class_name, ' '))]") _id_xpath = etree.XPath("descendant-or-self::*[@id=$id]") _collect_string_content = etree.XPath("string()") _css_url_re = re.compile(r'url\(('+'["][^"]*["]|'+"['][^']*[']|"+r'[^)]*)\)', re.I) _css_import_re = re.compile(r'@import "(.*?)"') _label_xpath = etree.XPath("//label[@for=$id]|//x:label[@for=$id]", namespaces={'x':XHTML_NAMESPACE}) _archive_re = re.compile(r'[^ ]+') def _unquote_match(s, pos): if s[:1] == '"' and s[-1:] == '"' or s[:1] == "'" and s[-1:] == "'": return s[1:-1], pos+1 else: return s,pos def _transform_result(typ, result): """Convert the result back into the input type. """ if issubclass(typ, bytes): return tostring(result, encoding='utf-8') elif issubclass(typ, unicode): return tostring(result, encoding=unicode) else: return result def _nons(tag): if isinstance(tag, basestring): if tag[0] == '{' and tag[1:len(XHTML_NAMESPACE)+1] == XHTML_NAMESPACE: return tag.split('}')[-1] return tag class HtmlMixin(object): def base_url(self): """ Returns the base URL, given when the page was parsed. Use with ``urlparse.urljoin(el.base_url, href)`` to get absolute URLs. """ return self.getroottree().docinfo.URL base_url = property(base_url, doc=base_url.__doc__) def forms(self): """ Return a list of all the forms """ return _forms_xpath(self) forms = property(forms, doc=forms.__doc__) def body(self): """ Return the <body> element. Can be called from a child element to get the document's head. """ return self.xpath('//body|//x:body', namespaces={'x':XHTML_NAMESPACE})[0] body = property(body, doc=body.__doc__) def head(self): """ Returns the <head> element. Can be called from a child element to get the document's head. """ return self.xpath('//head|//x:head', namespaces={'x':XHTML_NAMESPACE})[0] head = property(head, doc=head.__doc__) def _label__get(self): """ Get or set any <label> element associated with this element. """ id = self.get('id') if not id: return None result = _label_xpath(self, id=id) if not result: return None else: return result[0] def _label__set(self, label): id = self.get('id') if not id: raise TypeError( "You cannot set a label for an element (%r) that has no id" % self) if _nons(label.tag) != 'label': raise TypeError( "You can only assign label to a label element (not %r)" % label) label.set('for', id) def _label__del(self): label = self.label if label is not None: del label.attrib['for'] label = property(_label__get, _label__set, _label__del, doc=_label__get.__doc__) def drop_tree(self): """ Removes this element from the tree, including its children and text. The tail text is joined to the previous element or parent. """ parent = self.getparent() assert parent is not None if self.tail: previous = self.getprevious() if previous is None: parent.text = (parent.text or '') + self.tail else: previous.tail = (previous.tail or '') + self.tail parent.remove(self) def drop_tag(self): """ Remove the tag, but not its children or text. The children and text are merged into the parent. Example:: >>> h = fragment_fromstring('<div>Hello <b>World!</b></div>') >>> h.find('.//b').drop_tag() >>> print(tostring(h, encoding=unicode)) <div>Hello World!</div> """ parent = self.getparent() assert parent is not None previous = self.getprevious() if self.text and isinstance(self.tag, basestring): # not a Comment, etc. if previous is None: parent.text = (parent.text or '') + self.text else: previous.tail = (previous.tail or '') + self.text if self.tail: if len(self): last = self[-1] last.tail = (last.tail or '') + self.tail elif previous is None: parent.text = (parent.text or '') + self.tail else: previous.tail = (previous.tail or '') + self.tail index = parent.index(self) parent[index:index+1] = self[:] def find_rel_links(self, rel): """ Find any links like ``<a rel="{rel}">...</a>``; returns a list of elements. """ rel = rel.lower() return [el for el in _rel_links_xpath(self) if el.get('rel').lower() == rel] def find_class(self, class_name): """ Find any elements with the given class name. """ return _class_xpath(self, class_name=class_name) def get_element_by_id(self, id, *default): """ Get the first element in a document with the given id. If none is found, return the default argument if provided or raise KeyError otherwise. Note that there can be more than one element with the same id, and this isn't uncommon in HTML documents found in the wild. Browsers return only the first match, and this function does the same. """ try: # FIXME: should this check for multiple matches? # browsers just return the first one return _id_xpath(self, id=id)[0] except IndexError: if default: return default[0] else: raise KeyError(id) def text_content(self): """ Return the text content of the tag (and the text in any children). """ return _collect_string_content(self) def cssselect(self, expr, translator='html'): """ Run the CSS expression on this element and its children, returning a list of the results. Equivalent to lxml.cssselect.CSSSelect(expr, translator='html')(self) -- note that pre-compiling the expression can provide a substantial speedup. """ # Do the import here to make the dependency optional. from lxml.cssselect import CSSSelector return CSSSelector(expr, translator=translator)(self) ######################################## ## Link functions ######################################## def make_links_absolute(self, base_url=None, resolve_base_href=True): """ Make all links in the document absolute, given the ``base_url`` for the document (the full URL where the document came from), or if no ``base_url`` is given, then the ``.base_url`` of the document. If ``resolve_base_href`` is true, then any ``<base href>`` tags in the document are used *and* removed from the document. If it is false then any such tag is ignored. """ if base_url is None: base_url = self.base_url if base_url is None: raise TypeError( "No base_url given, and the document has no base_url") if resolve_base_href: self.resolve_base_href() def link_repl(href): return urljoin(base_url, href) self.rewrite_links(link_repl) def resolve_base_href(self): """ Find any ``<base href>`` tag in the document, and apply its values to all links found in the document. Also remove the tag once it has been applied. """ base_href = None basetags = self.xpath('//base[@href]|//x:base[@href]', namespaces={'x':XHTML_NAMESPACE}) for b in basetags: base_href = b.get('href') b.drop_tree() if not base_href: return self.make_links_absolute(base_href, resolve_base_href=False) def iterlinks(self): """ Yield (element, attribute, link, pos), where attribute may be None (indicating the link is in the text). ``pos`` is the position where the link occurs; often 0, but sometimes something else in the case of links in stylesheets or style tags. Note: <base href> is *not* taken into account in any way. The link you get is exactly the link in the document. Note: multiple links inside of a single text string or attribute value are returned in reversed order. This makes it possible to replace or delete them from the text string value based on their reported text positions. Otherwise, a modification at one text position can change the positions of links reported later on. """ link_attrs = defs.link_attrs for el in self.iter(): attribs = el.attrib tag = _nons(el.tag) if tag != 'object': for attrib in link_attrs: if attrib in attribs: yield (el, attrib, attribs[attrib], 0) elif tag == 'object': codebase = None ## <object> tags have attributes that are relative to ## codebase if 'codebase' in attribs: codebase = el.get('codebase') yield (el, 'codebase', codebase, 0) for attrib in 'classid', 'data': if attrib in attribs: value = el.get(attrib) if codebase is not None: value = urljoin(codebase, value) yield (el, attrib, value, 0) if 'archive' in attribs: for match in _archive_re.finditer(el.get('archive')): value = match.group(0) if codebase is not None: value = urljoin(codebase, value) yield (el, 'archive', value, match.start()) if tag == 'param': valuetype = el.get('valuetype') or '' if valuetype.lower() == 'ref': ## FIXME: while it's fine we *find* this link, ## according to the spec we aren't supposed to ## actually change the value, including resolving ## it. It can also still be a link, even if it ## doesn't have a valuetype="ref" (which seems to be the norm) ## http://www.w3.org/TR/html401/struct/objects.html#adef-valuetype yield (el, 'value', el.get('value'), 0) if tag == 'style' and el.text: urls = [ _unquote_match(match.group(1), match.start(1)) for match in _css_url_re.finditer(el.text) ] + [ (match.group(1), match.start(1)) for match in _css_import_re.finditer(el.text) ] if urls: # sort by start pos to bring both match sets back into order urls = [ (start, url) for (url, start) in urls ] urls.sort() # reverse the list to report correct positions despite # modifications urls.reverse() for start, url in urls: yield (el, None, url, start) if 'style' in attribs: urls = list(_css_url_re.finditer(attribs['style'])) if urls: # return in reversed order to simplify in-place modifications for match in urls[::-1]: url, start = _unquote_match(match.group(1), match.start(1)) yield (el, 'style', url, start) def rewrite_links(self, link_repl_func, resolve_base_href=True, base_href=None): """ Rewrite all the links in the document. For each link ``link_repl_func(link)`` will be called, and the return value will replace the old link. Note that links may not be absolute (unless you first called ``make_links_absolute()``), and may be internal (e.g., ``'#anchor'``). They can also be values like ``'mailto:email'`` or ``'javascript:expr'``. If you give ``base_href`` then all links passed to ``link_repl_func()`` will take that into account. If the ``link_repl_func`` returns None, the attribute or tag text will be removed completely. """ if base_href is not None: # FIXME: this can be done in one pass with a wrapper # around link_repl_func self.make_links_absolute(base_href, resolve_base_href=resolve_base_href) elif resolve_base_href: self.resolve_base_href() for el, attrib, link, pos in self.iterlinks(): new_link = link_repl_func(link.strip()) if new_link == link: continue if new_link is None: # Remove the attribute or element content if attrib is None: el.text = '' else: del el.attrib[attrib] continue if attrib is None: new = el.text[:pos] + new_link + el.text[pos+len(link):] el.text = new else: cur = el.attrib[attrib] if not pos and len(cur) == len(link): # Most common case el.attrib[attrib] = new_link else: new = cur[:pos] + new_link + cur[pos+len(link):] el.attrib[attrib] = new class _MethodFunc(object): """ An object that represents a method on an element as a function; the function takes either an element or an HTML string. It returns whatever the function normally returns, or if the function works in-place (and so returns None) it returns a serialized form of the resulting document. """ def __init__(self, name, copy=False, source_class=HtmlMixin): self.name = name self.copy = copy self.__doc__ = getattr(source_class, self.name).__doc__ def __call__(self, doc, *args, **kw): result_type = type(doc) if isinstance(doc, basestring): if 'copy' in kw: raise TypeError( "The keyword 'copy' can only be used with element inputs to %s, not a string input" % self.name) doc = fromstring(doc, **kw) else: if 'copy' in kw: make_a_copy = kw.pop('copy') else: make_a_copy = self.copy if make_a_copy: doc = copy.deepcopy(doc) meth = getattr(doc, self.name) result = meth(*args, **kw) # FIXME: this None test is a bit sloppy if result is None: # Then return what we got in return _transform_result(result_type, doc) else: return result find_rel_links = _MethodFunc('find_rel_links', copy=False) find_class = _MethodFunc('find_class', copy=False) make_links_absolute = _MethodFunc('make_links_absolute', copy=True) resolve_base_href = _MethodFunc('resolve_base_href', copy=True) iterlinks = _MethodFunc('iterlinks', copy=False) rewrite_links = _MethodFunc('rewrite_links', copy=True) class HtmlComment(etree.CommentBase, HtmlMixin): pass class HtmlElement(etree.ElementBase, HtmlMixin): pass class HtmlProcessingInstruction(etree.PIBase, HtmlMixin): pass class HtmlEntity(etree.EntityBase, HtmlMixin): pass class HtmlElementClassLookup(etree.CustomElementClassLookup): """A lookup scheme for HTML Element classes. To create a lookup instance with different Element classes, pass a tag name mapping of Element classes in the ``classes`` keyword argument and/or a tag name mapping of Mixin classes in the ``mixins`` keyword argument. The special key '*' denotes a Mixin class that should be mixed into all Element classes. """ _default_element_classes = {} def __init__(self, classes=None, mixins=None): etree.CustomElementClassLookup.__init__(self) if classes is None: classes = self._default_element_classes.copy() if mixins: mixers = {} for name, value in mixins: if name == '*': for n in classes.keys(): mixers.setdefault(n, []).append(value) else: mixers.setdefault(name, []).append(value) for name, mix_bases in mixers.items(): cur = classes.get(name, HtmlElement) bases = tuple(mix_bases + [cur]) classes[name] = type(cur.__name__, bases, {}) self._element_classes = classes def lookup(self, node_type, document, namespace, name): if node_type == 'element': return self._element_classes.get(name.lower(), HtmlElement) elif node_type == 'comment': return HtmlComment elif node_type == 'PI': return HtmlProcessingInstruction elif node_type == 'entity': return HtmlEntity # Otherwise normal lookup return None ################################################################################ # parsing ################################################################################ _looks_like_full_html_unicode = re.compile( unicode(r'^\s*<(?:html|!doctype)'), re.I).match _looks_like_full_html_bytes = re.compile( r'^\s*<(?:html|!doctype)'.encode('ascii'), re.I).match def document_fromstring(html, parser=None, **kw): if parser is None: parser = html_parser value = etree.fromstring(html, parser, **kw) if value is None: raise etree.ParserError( "Document is empty") return value def fragments_fromstring(html, no_leading_text=False, base_url=None, parser=None, **kw): """ Parses several HTML elements, returning a list of elements. The first item in the list may be a string (though leading whitespace is removed). If no_leading_text is true, then it will be an error if there is leading text, and it will always be a list of only elements. base_url will set the document's base_url attribute (and the tree's docinfo.URL) """ if parser is None: parser = html_parser # FIXME: check what happens when you give html with a body, head, etc. if isinstance(html, bytes): if not _looks_like_full_html_bytes(html): html = '<html><body>%s</body></html>'.encode('ascii') % html else: if not _looks_like_full_html_unicode(html): html = '<html><body>%s</body></html>' % html doc = document_fromstring(html, parser=parser, base_url=base_url, **kw) assert _nons(doc.tag) == 'html' bodies = [e for e in doc if _nons(e.tag) == 'body'] assert len(bodies) == 1, ("too many bodies: %r in %r" % (bodies, html)) body = bodies[0] elements = [] if no_leading_text and body.text and body.text.strip(): raise etree.ParserError( "There is leading text: %r" % body.text) if body.text and body.text.strip(): elements.append(body.text) elements.extend(body) # FIXME: removing the reference to the parent artificial document # would be nice return elements def fragment_fromstring(html, create_parent=False, base_url=None, parser=None, **kw): """ Parses a single HTML element; it is an error if there is more than one element, or if anything but whitespace precedes or follows the element. If create_parent is true (or is a tag name) then a parent node will be created to encapsulate the HTML in a single element. In this case, leading or trailing text is allowed. base_url will set the document's base_url attribute (and the tree's docinfo.URL) """ if parser is None: parser = html_parser accept_leading_text = bool(create_parent) elements = fragments_fromstring( html, parser=parser, no_leading_text=not accept_leading_text, base_url=base_url, **kw) if create_parent: if not isinstance(create_parent, basestring): create_parent = 'div' new_root = Element(create_parent) if elements: if isinstance(elements[0], basestring): new_root.text = elements[0] del elements[0] new_root.extend(elements) return new_root if not elements: raise etree.ParserError('No elements found') if len(elements) > 1: raise etree.ParserError( "Multiple elements found (%s)" % ', '.join([_element_name(e) for e in elements])) el = elements[0] if el.tail and el.tail.strip(): raise etree.ParserError( "Element followed by text: %r" % el.tail) el.tail = None return el def fromstring(html, base_url=None, parser=None, **kw): """ Parse the html, returning a single element/document. This tries to minimally parse the chunk of text, without knowing if it is a fragment or a document. base_url will set the document's base_url attribute (and the tree's docinfo.URL) """ if parser is None: parser = html_parser if isinstance(html, bytes): is_full_html = _looks_like_full_html_bytes(html) else: is_full_html = _looks_like_full_html_unicode(html) doc = document_fromstring(html, parser=parser, base_url=base_url, **kw) if is_full_html: return doc # otherwise, lets parse it out... bodies = doc.findall('body') if not bodies: bodies = doc.findall('{%s}body' % XHTML_NAMESPACE) if bodies: body = bodies[0] if len(bodies) > 1: # Somehow there are multiple bodies, which is bad, but just # smash them into one body for other_body in bodies[1:]: if other_body.text: if len(body): body[-1].tail = (body[-1].tail or '') + other_body.text else: body.text = (body.text or '') + other_body.text body.extend(other_body) # We'll ignore tail # I guess we are ignoring attributes too other_body.drop_tree() else: body = None heads = doc.findall('head') if not heads: heads = doc.findall('{%s}head' % XHTML_NAMESPACE) if heads: # Well, we have some sort of structure, so lets keep it all head = heads[0] if len(heads) > 1: for other_head in heads[1:]: head.extend(other_head) # We don't care about text or tail in a head other_head.drop_tree() return doc if body is None: return doc if (len(body) == 1 and (not body.text or not body.text.strip()) and (not body[-1].tail or not body[-1].tail.strip())): # The body has just one element, so it was probably a single # element passed in return body[0] # Now we have a body which represents a bunch of tags which have the # content that was passed in. We will create a fake container, which # is the body tag, except <body> implies too much structure. if _contains_block_level_tag(body): body.tag = 'div' else: body.tag = 'span' return body def parse(filename_or_url, parser=None, base_url=None, **kw): """ Parse a filename, URL, or file-like object into an HTML document tree. Note: this returns a tree, not an element. Use ``parse(...).getroot()`` to get the document root. You can override the base URL with the ``base_url`` keyword. This is most useful when parsing from a file-like object. """ if parser is None: parser = html_parser return etree.parse(filename_or_url, parser, base_url=base_url, **kw) def _contains_block_level_tag(el): # FIXME: I could do this with XPath, but would that just be # unnecessarily slow? for el in el.iter(): if _nons(el.tag) in defs.block_tags: return True return False def _element_name(el): if isinstance(el, etree.CommentBase): return 'comment' elif isinstance(el, basestring): return 'string' else: return _nons(el.tag) ################################################################################ # form handling ################################################################################ class FormElement(HtmlElement): """ Represents a <form> element. """ def inputs(self): """ Returns an accessor for all the input elements in the form. See `InputGetter` for more information about the object. """ return InputGetter(self) inputs = property(inputs, doc=inputs.__doc__) def _fields__get(self): """ Dictionary-like object that represents all the fields in this form. You can set values in this dictionary to effect the form. """ return FieldsDict(self.inputs) def _fields__set(self, value): prev_keys = self.fields.keys() for key, value in value.items(): if key in prev_keys: prev_keys.remove(key) self.fields[key] = value for key in prev_keys: if key is None: # Case of an unnamed input; these aren't really # expressed in form_values() anyway. continue self.fields[key] = None fields = property(_fields__get, _fields__set, doc=_fields__get.__doc__) def _name(self): if self.get('name'): return self.get('name') elif self.get('id'): return '#' + self.get('id') forms = list(self.body.iter('form')) if not forms: forms = list(self.body.iter('{%s}form' % XHTML_NAMESPACE)) return str(forms.index(self)) def form_values(self): """ Return a list of tuples of the field values for the form. This is suitable to be passed to ``urllib.urlencode()``. """ results = [] for el in self.inputs: name = el.name if not name: continue tag = _nons(el.tag) if tag == 'textarea': results.append((name, el.value)) elif tag == 'select': value = el.value if el.multiple: for v in value: results.append((name, v)) elif value is not None: results.append((name, el.value)) else: assert tag == 'input', ( "Unexpected tag: %r" % el) if el.checkable and not el.checked: continue if el.type in ('submit', 'image', 'reset'): continue value = el.value if value is not None: results.append((name, el.value)) return results def _action__get(self): """ Get/set the form's ``action`` attribute. """ base_url = self.base_url action = self.get('action') if base_url and action is not None: return urljoin(base_url, action) else: return action def _action__set(self, value): self.set('action', value) def _action__del(self): if 'action' in self.attrib: del self.attrib['action'] action = property(_action__get, _action__set, _action__del, doc=_action__get.__doc__) def _method__get(self): """ Get/set the form's method. Always returns a capitalized string, and defaults to ``'GET'`` """ return self.get('method', 'GET').upper() def _method__set(self, value): self.set('method', value.upper()) method = property(_method__get, _method__set, doc=_method__get.__doc__) HtmlElementClassLookup._default_element_classes['form'] = FormElement def submit_form(form, extra_values=None, open_http=None): """ Helper function to submit a form. Returns a file-like object, as from ``urllib.urlopen()``. This object also has a ``.geturl()`` function, which shows the URL if there were any redirects. You can use this like:: form = doc.forms[0] form.inputs['foo'].value = 'bar' # etc response = form.submit() doc = parse(response) doc.make_links_absolute(response.geturl()) To change the HTTP requester, pass a function as ``open_http`` keyword argument that opens the URL for you. The function must have the following signature:: open_http(method, URL, values) The action is one of 'GET' or 'POST', the URL is the target URL as a string, and the values are a sequence of ``(name, value)`` tuples with the form data. """ values = form.form_values() if extra_values: if hasattr(extra_values, 'items'): extra_values = extra_values.items() values.extend(extra_values) if open_http is None: open_http = open_http_urllib if form.action: url = form.action else: url = form.base_url return open_http(form.method, url, values) def open_http_urllib(method, url, values): if not url: raise ValueError("cannot submit, no URL provided") ## FIXME: should test that it's not a relative URL or something try: from urllib import urlencode, urlopen except ImportError: # Python 3 from urllib.request import urlopen from urllib.parse import urlencode if method == 'GET': if '?' in url: url += '&' else: url += '?' url += urlencode(values) data = None else: data = urlencode(values) return urlopen(url, data) class FieldsDict(DictMixin): def __init__(self, inputs): self.inputs = inputs def __getitem__(self, item): return self.inputs[item].value def __setitem__(self, item, value): self.inputs[item].value = value def __delitem__(self, item): raise KeyError( "You cannot remove keys from ElementDict") def keys(self): return self.inputs.keys() def __contains__(self, item): return item in self.inputs def __iter__(self): return iter(self.inputs.keys()) def __len__(self): return len(self.inputs) def __repr__(self): return '<%s for form %s>' % ( self.__class__.__name__, self.inputs.form._name()) class InputGetter(object): """ An accessor that represents all the input fields in a form. You can get fields by name from this, with ``form.inputs['field_name']``. If there are a set of checkboxes with the same name, they are returned as a list (a `CheckboxGroup` which also allows value setting). Radio inputs are handled similarly. You can also iterate over this to get all input elements. This won't return the same thing as if you get all the names, as checkboxes and radio elements are returned individually. """ _name_xpath = etree.XPath(".//*[@name = $name and (local-name(.) = 'select' or local-name(.) = 'input' or local-name(.) = 'textarea')]") _all_xpath = etree.XPath(".//*[local-name() = 'select' or local-name() = 'input' or local-name() = 'textarea']") def __init__(self, form): self.form = form def __repr__(self): return '<%s for form %s>' % ( self.__class__.__name__, self.form._name()) ## FIXME: there should be more methods, and it's unclear if this is ## a dictionary-like object or list-like object def __getitem__(self, name): results = self._name_xpath(self.form, name=name) if results: type = results[0].get('type') if type == 'radio' and len(results) > 1: group = RadioGroup(results) group.name = name return group elif type == 'checkbox' and len(results) > 1: group = CheckboxGroup(results) group.name = name return group else: # I don't like throwing away elements like this return results[0] else: raise KeyError( "No input element with the name %r" % name) def __contains__(self, name): results = self._name_xpath(self.form, name=name) return bool(results) def keys(self): names = set() for el in self: names.add(el.name) if None in names: names.remove(None) return list(names) def __iter__(self): ## FIXME: kind of dumb to turn a list into an iterator, only ## to have it likely turned back into a list again :( return iter(self._all_xpath(self.form)) class InputMixin(object): """ Mix-in for all input elements (input, select, and textarea) """ def _name__get(self): """ Get/set the name of the element """ return self.get('name') def _name__set(self, value): self.set('name', value) def _name__del(self): if 'name' in self.attrib: del self.attrib['name'] name = property(_name__get, _name__set, _name__del, doc=_name__get.__doc__) def __repr__(self): type = getattr(self, 'type', None) if type: type = ' type=%r' % type else: type = '' return '<%s %x name=%r%s>' % ( self.__class__.__name__, id(self), self.name, type) class TextareaElement(InputMixin, HtmlElement): """ ``<textarea>`` element. You can get the name with ``.name`` and get/set the value with ``.value`` """ def _value__get(self): """ Get/set the value (which is the contents of this element) """ content = self.text or '' if self.tag.startswith("{%s}" % XHTML_NAMESPACE): serialisation_method = 'xml' else: serialisation_method = 'html' for el in self: # it's rare that we actually get here, so let's not use ''.join() content += etree.tostring(el, method=serialisation_method, encoding=unicode) return content def _value__set(self, value): del self[:] self.text = value def _value__del(self): self.text = '' del self[:] value = property(_value__get, _value__set, _value__del, doc=_value__get.__doc__) HtmlElementClassLookup._default_element_classes['textarea'] = TextareaElement class SelectElement(InputMixin, HtmlElement): """ ``<select>`` element. You can get the name with ``.name``. ``.value`` will be the value of the selected option, unless this is a multi-select element (``<select multiple>``), in which case it will be a set-like object. In either case ``.value_options`` gives the possible values. The boolean attribute ``.multiple`` shows if this is a multi-select. """ def _value__get(self): """ Get/set the value of this select (the selected option). If this is a multi-select, this is a set-like object that represents all the selected options. """ if self.multiple: return MultipleSelectOptions(self) for el in _options_xpath(self): if el.get('selected') is not None: value = el.get('value') if value is None: value = el.text or '' if value: value = value.strip() return value return None def _value__set(self, value): if self.multiple: if isinstance(value, basestring): raise TypeError( "You must pass in a sequence") self.value.clear() self.value.update(value) return if value is not None: value = value.strip() for el in _options_xpath(self): opt_value = el.get('value') if opt_value is None: opt_value = el.text or '' if opt_value: opt_value = opt_value.strip() if opt_value == value: checked_option = el break else: raise ValueError( "There is no option with the value of %r" % value) for el in _options_xpath(self): if 'selected' in el.attrib: del el.attrib['selected'] if value is not None: checked_option.set('selected', '') def _value__del(self): # FIXME: should del be allowed at all? if self.multiple: self.value.clear() else: self.value = None value = property(_value__get, _value__set, _value__del, doc=_value__get.__doc__) def value_options(self): """ All the possible values this select can have (the ``value`` attribute of all the ``<option>`` elements. """ options = [] for el in _options_xpath(self): value = el.get('value') if value is None: value = el.text or '' if value: value = value.strip() options.append(value) return options value_options = property(value_options, doc=value_options.__doc__) def _multiple__get(self): """ Boolean attribute: is there a ``multiple`` attribute on this element. """ return 'multiple' in self.attrib def _multiple__set(self, value): if value: self.set('multiple', '') elif 'multiple' in self.attrib: del self.attrib['multiple'] multiple = property(_multiple__get, _multiple__set, doc=_multiple__get.__doc__) HtmlElementClassLookup._default_element_classes['select'] = SelectElement class MultipleSelectOptions(SetMixin): """ Represents all the selected options in a ``<select multiple>`` element. You can add to this set-like option to select an option, or remove to unselect the option. """ def __init__(self, select): self.select = select def options(self): """ Iterator of all the ``<option>`` elements. """ return iter(_options_xpath(self.select)) options = property(options) def __iter__(self): for option in self.options: if 'selected' in option.attrib: opt_value = option.get('value') if opt_value is None: opt_value = option.text or '' if opt_value: opt_value = opt_value.strip() yield opt_value def add(self, item): for option in self.options: opt_value = option.get('value') if opt_value is None: opt_value = option.text or '' if opt_value: opt_value = opt_value.strip() if opt_value == item: option.set('selected', '') break else: raise ValueError( "There is no option with the value %r" % item) def remove(self, item): for option in self.options: opt_value = option.get('value') if opt_value is None: opt_value = option.text or '' if opt_value: opt_value = opt_value.strip() if opt_value == item: if 'selected' in option.attrib: del option.attrib['selected'] else: raise ValueError( "The option %r is not currently selected" % item) break else: raise ValueError( "There is not option with the value %r" % item) def __repr__(self): return '<%s {%s} for select name=%r>' % ( self.__class__.__name__, ', '.join([repr(v) for v in self]), self.select.name) class RadioGroup(list): """ This object represents several ``<input type=radio>`` elements that have the same name. You can use this like a list, but also use the property ``.value`` to check/uncheck inputs. Also you can use ``.value_options`` to get the possible values. """ def _value__get(self): """ Get/set the value, which checks the radio with that value (and unchecks any other value). """ for el in self: if 'checked' in el.attrib: return el.get('value') return None def _value__set(self, value): if value is not None: for el in self: if el.get('value') == value: checked_option = el break else: raise ValueError( "There is no radio input with the value %r" % value) for el in self: if 'checked' in el.attrib: del el.attrib['checked'] if value is not None: checked_option.set('checked', '') def _value__del(self): self.value = None value = property(_value__get, _value__set, _value__del, doc=_value__get.__doc__) def value_options(self): """ Returns a list of all the possible values. """ return [el.get('value') for el in self] value_options = property(value_options, doc=value_options.__doc__) def __repr__(self): return '%s(%s)' % ( self.__class__.__name__, list.__repr__(self)) class CheckboxGroup(list): """ Represents a group of checkboxes (``<input type=checkbox>``) that have the same name. In addition to using this like a list, the ``.value`` attribute returns a set-like object that you can add to or remove from to check and uncheck checkboxes. You can also use ``.value_options`` to get the possible values. """ def _value__get(self): """ Return a set-like object that can be modified to check or uncheck individual checkboxes according to their value. """ return CheckboxValues(self) def _value__set(self, value): self.value.clear() if not hasattr(value, '__iter__'): raise ValueError( "A CheckboxGroup (name=%r) must be set to a sequence (not %r)" % (self[0].name, value)) self.value.update(value) def _value__del(self): self.value.clear() value = property(_value__get, _value__set, _value__del, doc=_value__get.__doc__) def value_options(self): """ Returns a list of all the possible values. """ return [el.get('value') for el in self] value_options = property(value_options, doc=value_options.__doc__) def __repr__(self): return '%s(%s)' % ( self.__class__.__name__, list.__repr__(self)) class CheckboxValues(SetMixin): """ Represents the values of the checked checkboxes in a group of checkboxes with the same name. """ def __init__(self, group): self.group = group def __iter__(self): return iter([ el.get('value') for el in self.group if 'checked' in el.attrib]) def add(self, value): for el in self.group: if el.get('value') == value: el.set('checked', '') break else: raise KeyError("No checkbox with value %r" % value) def remove(self, value): for el in self.group: if el.get('value') == value: if 'checked' in el.attrib: del el.attrib['checked'] else: raise KeyError( "The checkbox with value %r was already unchecked" % value) break else: raise KeyError( "No checkbox with value %r" % value) def __repr__(self): return '<%s {%s} for checkboxes name=%r>' % ( self.__class__.__name__, ', '.join([repr(v) for v in self]), self.group.name) class InputElement(InputMixin, HtmlElement): """ Represents an ``<input>`` element. You can get the type with ``.type`` (which is lower-cased and defaults to ``'text'``). Also you can get and set the value with ``.value`` Checkboxes and radios have the attribute ``input.checkable == True`` (for all others it is false) and a boolean attribute ``.checked``. """ ## FIXME: I'm a little uncomfortable with the use of .checked def _value__get(self): """ Get/set the value of this element, using the ``value`` attribute. Also, if this is a checkbox and it has no value, this defaults to ``'on'``. If it is a checkbox or radio that is not checked, this returns None. """ if self.checkable: if self.checked: return self.get('value') or 'on' else: return None return self.get('value') def _value__set(self, value): if self.checkable: if not value: self.checked = False else: self.checked = True if isinstance(value, basestring): self.set('value', value) else: self.set('value', value) def _value__del(self): if self.checkable: self.checked = False else: if 'value' in self.attrib: del self.attrib['value'] value = property(_value__get, _value__set, _value__del, doc=_value__get.__doc__) def _type__get(self): """ Return the type of this element (using the type attribute). """ return self.get('type', 'text').lower() def _type__set(self, value): self.set('type', value) type = property(_type__get, _type__set, doc=_type__get.__doc__) def checkable(self): """ Boolean: can this element be checked? """ return self.type in ['checkbox', 'radio'] checkable = property(checkable, doc=checkable.__doc__) def _checked__get(self): """ Boolean attribute to get/set the presence of the ``checked`` attribute. You can only use this on checkable input types. """ if not self.checkable: raise AttributeError('Not a checkable input type') return 'checked' in self.attrib def _checked__set(self, value): if not self.checkable: raise AttributeError('Not a checkable input type') if value: self.set('checked', '') else: if 'checked' in self.attrib: del self.attrib['checked'] checked = property(_checked__get, _checked__set, doc=_checked__get.__doc__) HtmlElementClassLookup._default_element_classes['input'] = InputElement class LabelElement(HtmlElement): """ Represents a ``<label>`` element. Label elements are linked to other elements with their ``for`` attribute. You can access this element with ``label.for_element``. """ def _for_element__get(self): """ Get/set the element this label points to. Return None if it can't be found. """ id = self.get('for') if not id: return None return self.body.get_element_by_id(id) def _for_element__set(self, other): id = other.get('id') if not id: raise TypeError( "Element %r has no id attribute" % other) self.set('for', id) def _for_element__del(self): if 'id' in self.attrib: del self.attrib['id'] for_element = property(_for_element__get, _for_element__set, _for_element__del, doc=_for_element__get.__doc__) HtmlElementClassLookup._default_element_classes['label'] = LabelElement ############################################################ ## Serialization ############################################################ def html_to_xhtml(html): """Convert all tags in an HTML tree to XHTML by moving them to the XHTML namespace. """ try: html = html.getroot() except AttributeError: pass prefix = "{%s}" % XHTML_NAMESPACE for el in html.iter(): tag = el.tag if isinstance(tag, basestring): if tag[0] != '{': el.tag = prefix + tag def xhtml_to_html(xhtml): """Convert all tags in an XHTML tree to HTML by removing their XHTML namespace. """ try: xhtml = xhtml.getroot() except AttributeError: pass prefix = "{%s}" % XHTML_NAMESPACE prefix_len = len(prefix) for el in xhtml.iter(prefix + "*"): el.tag = el.tag[prefix_len:] # This isn't a general match, but it's a match for what libxml2 # specifically serialises: __str_replace_meta_content_type = re.compile( r'<meta http-equiv="Content-Type"[^>]*>').sub __bytes_replace_meta_content_type = re.compile( r'<meta http-equiv="Content-Type"[^>]*>'.encode('ASCII')).sub def tostring(doc, pretty_print=False, include_meta_content_type=False, encoding=None, method="html", with_tail=True, doctype=None): """Return an HTML string representation of the document. Note: if include_meta_content_type is true this will create a ``<meta http-equiv="Content-Type" ...>`` tag in the head; regardless of the value of include_meta_content_type any existing ``<meta http-equiv="Content-Type" ...>`` tag will be removed The ``encoding`` argument controls the output encoding (defauts to ASCII, with &#...; character references for any characters outside of ASCII). Note that you can pass the name ``'unicode'`` as ``encoding`` argument to serialise to a unicode string. The ``method`` argument defines the output method. It defaults to 'html', but can also be 'xml' for xhtml output, or 'text' to serialise to plain text without markup. To leave out the tail text of the top-level element that is being serialised, pass ``with_tail=False``. The ``doctype`` option allows passing in a plain string that will be serialised before the XML tree. Note that passing in non well-formed content here will make the XML output non well-formed. Also, an existing doctype in the document tree will not be removed when serialising an ElementTree instance. Example:: >>> from lxml import html >>> root = html.fragment_fromstring('<p>Hello<br>world!</p>') >>> html.tostring(root) b'<p>Hello<br>world!</p>' >>> html.tostring(root, method='html') b'<p>Hello<br>world!</p>' >>> html.tostring(root, method='xml') b'<p>Hello<br/>world!</p>' >>> html.tostring(root, method='text') b'Helloworld!' >>> html.tostring(root, method='text', encoding=unicode) u'Helloworld!' >>> root = html.fragment_fromstring('<div><p>Hello<br>world!</p>TAIL</div>') >>> html.tostring(root[0], method='text', encoding=unicode) u'Helloworld!TAIL' >>> html.tostring(root[0], method='text', encoding=unicode, with_tail=False) u'Helloworld!' >>> doc = html.document_fromstring('<p>Hello<br>world!</p>') >>> html.tostring(doc, method='html', encoding=unicode) u'<html><body><p>Hello<br>world!</p></body></html>' >>> print(html.tostring(doc, method='html', encoding=unicode, ... doctype='<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"' ... ' "http://www.w3.org/TR/html4/strict.dtd">')) <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd"> <html><body><p>Hello<br>world!</p></body></html> """ html = etree.tostring(doc, method=method, pretty_print=pretty_print, encoding=encoding, with_tail=with_tail, doctype=doctype) if method == 'html' and not include_meta_content_type: if isinstance(html, str): html = __str_replace_meta_content_type('', html) else: html = __bytes_replace_meta_content_type(bytes(), html) return html tostring.__doc__ = __fix_docstring(tostring.__doc__) def open_in_browser(doc, encoding=None): """ Open the HTML document in a web browser, saving it to a temporary file to open it. Note that this does not delete the file after use. This is mainly meant for debugging. """ import os import webbrowser import tempfile if not isinstance(doc, etree._ElementTree): doc = etree.ElementTree(doc) handle, fn = tempfile.mkstemp(suffix='.html') f = os.fdopen(handle, 'wb') try: doc.write(f, method="html", encoding=encoding or doc.docinfo.encoding or "UTF-8") finally: # we leak the file itself here, but we should at least close it f.close() url = 'file://' + fn.replace(os.path.sep, '/') print(url) webbrowser.open(url) ################################################################################ # configure Element class lookup ################################################################################ class HTMLParser(etree.HTMLParser): """An HTML parser that is configured to return lxml.html Element objects. """ def __init__(self, **kwargs): super(HTMLParser, self).__init__(**kwargs) self.set_element_class_lookup(HtmlElementClassLookup()) class XHTMLParser(etree.XMLParser): """An XML parser that is configured to return lxml.html Element objects. Note that this parser is not really XHTML aware unless you let it load a DTD that declares the HTML entities. To do this, make sure you have the XHTML DTDs installed in your catalogs, and create the parser like this:: >>> parser = XHTMLParser(load_dtd=True) If you additionally want to validate the document, use this:: >>> parser = XHTMLParser(dtd_validation=True) For catalog support, see http://www.xmlsoft.org/catalog.html. """ def __init__(self, **kwargs): super(XHTMLParser, self).__init__(**kwargs) self.set_element_class_lookup(HtmlElementClassLookup()) def Element(*args, **kw): """Create a new HTML Element. This can also be used for XHTML documents. """ v = html_parser.makeelement(*args, **kw) return v html_parser = HTMLParser() xhtml_parser = XHTMLParser()