cmlenz@1: # -*- coding: utf-8 -*-
cmlenz@1: #
cmlenz@1: # Copyright (C) 2006 Christopher Lenz
cmlenz@1: # All rights reserved.
cmlenz@1: #
cmlenz@1: # This software is licensed as described in the file COPYING, which
cmlenz@1: # you should have received as part of this distribution. The terms
cmlenz@1: # are also available at http://trac.edgewall.com/license.html.
cmlenz@1: #
cmlenz@1: # This software consists of voluntary contributions made by many
cmlenz@1: # individuals. For the exact contribution history, see the revision
cmlenz@1: # history and logs, available at http://projects.edgewall.com/trac/.
cmlenz@1: 
cmlenz@1: """This module provides different kinds of serialization methods for XML event
cmlenz@1: streams.
cmlenz@1: """
cmlenz@1: 
cmlenz@1: try:
cmlenz@1:     frozenset
cmlenz@1: except NameError:
cmlenz@1:     from sets import ImmutableSet as frozenset
cmlenz@1: 
cmlenz@18: from markup.core import Markup, Namespace, QName, Stream
cmlenz@1: from markup.filters import WhitespaceFilter
cmlenz@1: 
cmlenz@1: __all__ = ['Serializer', 'XMLSerializer', 'HTMLSerializer']
cmlenz@1: 
cmlenz@1: 
cmlenz@1: class Serializer(object):
cmlenz@1:     """Base class for serializers."""
cmlenz@1: 
cmlenz@1:     def serialize(self, stream):
cmlenz@1:         raise NotImplementedError
cmlenz@1: 
cmlenz@1: 
cmlenz@1: class XMLSerializer(Serializer):
cmlenz@1:     """Produces XML text from an event stream.
cmlenz@1:     
cmlenz@1:     >>> from markup.builder import tag
cmlenz@1:     >>> elem = tag.DIV(tag.A(href='foo'), tag.BR, tag.HR(noshade=True))
cmlenz@1:     >>> print ''.join(XMLSerializer().serialize(elem.generate()))
cmlenz@1:     <div><a href="foo"/><br/><hr noshade="True"/></div>
cmlenz@1:     """
cmlenz@1: 
cmlenz@1:     def serialize(self, stream):
cmlenz@1:         ns_attrib = []
cmlenz@1:         ns_mapping = {}
cmlenz@1: 
cmlenz@1:         stream = PushbackIterator(stream)
cmlenz@1:         for kind, data, pos in stream:
cmlenz@1: 
cmlenz@1:             if kind is Stream.DOCTYPE:
cmlenz@1:                 # FIXME: what if there's no system or public ID in the input?
cmlenz@1:                 yield Markup('<!DOCTYPE %s "%s" "%s">\n' % data)
cmlenz@1: 
cmlenz@1:             elif kind is Stream.START_NS:
cmlenz@1:                 prefix, uri = data
cmlenz@1:                 if uri not in ns_mapping:
cmlenz@1:                     ns_mapping[uri] = prefix
cmlenz@1:                     if not prefix:
cmlenz@1:                         ns_attrib.append((QName('xmlns'), uri))
cmlenz@1:                     else:
cmlenz@1:                         ns_attrib.append((QName('xmlns:%s' % prefix), uri))
cmlenz@1: 
cmlenz@1:             elif kind is Stream.START:
cmlenz@1:                 tag, attrib = data
cmlenz@1: 
cmlenz@1:                 tagname = tag.localname
cmlenz@1:                 if tag.namespace:
cmlenz@1:                     try:
cmlenz@1:                         prefix = ns_mapping[tag.namespace]
cmlenz@1:                         if prefix:
cmlenz@1:                             tagname = prefix + ':' + tag.localname
cmlenz@1:                     except KeyError:
cmlenz@1:                         ns_attrib.append((QName('xmlns'), tag.namespace))
cmlenz@1:                 buf = ['<', tagname]
cmlenz@1: 
cmlenz@1:                 if ns_attrib:
cmlenz@1:                     attrib.extend(ns_attrib)
cmlenz@1:                     ns_attrib = []
cmlenz@1:                 for attr, value in attrib:
cmlenz@1:                     attrname = attr.localname
cmlenz@1:                     if attr.namespace:
cmlenz@1:                         try:
cmlenz@1:                             prefix = ns_mapping[attr.namespace]
cmlenz@1:                         except KeyError:
cmlenz@1:                             # FIXME: synthesize a prefix for the attribute?
cmlenz@1:                             prefix = ''
cmlenz@1:                         if prefix:
cmlenz@1:                             attrname = prefix + ':' + attrname
cmlenz@1:                     buf.append(' %s="%s"' % (attrname, Markup.escape(value)))
cmlenz@1: 
cmlenz@1:                 kind, data, pos = stream.next()
cmlenz@1:                 if kind is Stream.END:
cmlenz@1:                     buf.append('/>')
cmlenz@1:                 else:
cmlenz@1:                     buf.append('>')
cmlenz@1:                     stream.pushback((kind, data, pos))
cmlenz@1: 
cmlenz@1:                 yield Markup(''.join(buf))
cmlenz@1: 
cmlenz@1:             elif kind is Stream.END:
cmlenz@1:                 tag = data
cmlenz@1:                 tagname = tag.localname
cmlenz@1:                 if tag.namespace:
cmlenz@19:                     try:
cmlenz@19:                         prefix = ns_mapping[tag.namespace]
cmlenz@19:                         if prefix:
cmlenz@19:                             tagname = prefix + ':' + tag.localname
cmlenz@19:                     except KeyError:
cmlenz@19:                         pass
cmlenz@1:                 yield Markup('</%s>' % tagname)
cmlenz@1: 
cmlenz@1:             elif kind is Stream.TEXT:
cmlenz@1:                 yield Markup.escape(data, quotes=False)
cmlenz@1: 
cmlenz@1: 
cmlenz@1: class HTMLSerializer(Serializer):
cmlenz@1:     """Produces HTML text from an event stream.
cmlenz@1:     
cmlenz@1:     >>> from markup.builder import tag
cmlenz@1:     >>> elem = tag.DIV(tag.A(href='foo'), tag.BR, tag.HR(noshade=True))
cmlenz@1:     >>> print ''.join(HTMLSerializer().serialize(elem.generate()))
cmlenz@1:     <div><a href="foo"></a><br><hr noshade></div>
cmlenz@1:     """
cmlenz@1: 
cmlenz@18:     NAMESPACE = Namespace('http://www.w3.org/1999/xhtml')
cmlenz@1: 
cmlenz@1:     _EMPTY_ELEMS = frozenset(['area', 'base', 'basefont', 'br', 'col', 'frame',
cmlenz@1:                               'hr', 'img', 'input', 'isindex', 'link', 'meta',
cmlenz@1:                               'param'])
cmlenz@1:     _BOOLEAN_ATTRS = frozenset(['selected', 'checked', 'compact', 'declare',
cmlenz@1:                                 'defer', 'disabled', 'ismap', 'multiple',
cmlenz@1:                                 'nohref', 'noresize', 'noshade', 'nowrap'])
cmlenz@1: 
cmlenz@1:     def serialize(self, stream):
cmlenz@1:         ns_mapping = {}
cmlenz@1: 
cmlenz@1:         stream = PushbackIterator(stream)
cmlenz@1:         for kind, data, pos in stream:
cmlenz@1: 
cmlenz@1:             if kind is Stream.DOCTYPE:
cmlenz@1:                 yield Markup('<!DOCTYPE %s "%s" "%s">\n' % data)
cmlenz@1: 
cmlenz@1:             elif kind is Stream.START_NS:
cmlenz@1:                 prefix, uri = data
cmlenz@1:                 if uri not in ns_mapping:
cmlenz@1:                     ns_mapping[uri] = prefix
cmlenz@1: 
cmlenz@1:             elif kind is Stream.START:
cmlenz@1:                 tag, attrib = data
cmlenz@18:                 if tag.namespace and tag not in self.NAMESPACE:
cmlenz@1:                     continue # not in the HTML namespace, so don't emit
cmlenz@1:                 buf = ['<', tag.localname]
cmlenz@1:                 for attr, value in attrib:
cmlenz@18:                     if attr.namespace and attr not in self.NAMESPACE:
cmlenz@1:                         continue # not in the HTML namespace, so don't emit
cmlenz@1:                     if attr.localname in self._BOOLEAN_ATTRS:
cmlenz@1:                         if value:
cmlenz@1:                             buf.append(' %s' % attr.localname)
cmlenz@1:                     else:
cmlenz@1:                         buf.append(' %s="%s"' % (attr.localname,
cmlenz@1:                                                  Markup.escape(value)))
cmlenz@1: 
cmlenz@1:                 if tag.localname in self._EMPTY_ELEMS:
cmlenz@1:                     kind, data, pos = stream.next()
cmlenz@1:                     if kind is not Stream.END:
cmlenz@1:                         stream.pushback((kind, data, pos))
cmlenz@1: 
cmlenz@1:                 yield Markup(''.join(buf + ['>']))
cmlenz@1: 
cmlenz@1:             elif kind is Stream.END:
cmlenz@1:                 tag = data
cmlenz@18:                 if tag.namespace and tag not in self.NAMESPACE:
cmlenz@1:                     continue # not in the HTML namespace, so don't emit
cmlenz@1:                 yield Markup('</%s>' % tag.localname)
cmlenz@1: 
cmlenz@1:             elif kind is Stream.TEXT:
cmlenz@1:                 yield Markup.escape(data, quotes=False)
cmlenz@1: 
cmlenz@1: 
cmlenz@1: class PushbackIterator(object):
cmlenz@1:     """A simple wrapper for iterators that allows pushing items back on the
cmlenz@1:     queue via the `pushback()` method.
cmlenz@1:     
cmlenz@1:     That can effectively be used to peek at the next item."""
cmlenz@1:     __slots__ = ['iterable', 'buf']
cmlenz@1: 
cmlenz@1:     def __init__(self, iterable):
cmlenz@1:         self.iterable = iter(iterable)
cmlenz@1:         self.buf = []
cmlenz@1: 
cmlenz@1:     def __iter__(self):
cmlenz@1:         return self
cmlenz@1: 
cmlenz@1:     def next(self):
cmlenz@1:         if self.buf:
cmlenz@1:             return self.buf.pop(0)
cmlenz@1:         return self.iterable.next()
cmlenz@1: 
cmlenz@1:     def pushback(self, item):
cmlenz@1:         self.buf.append(item)