# XML parsing

import re

import xml.dom.minidom as minidom

import xml.parsers.expat as expat

import exc

import streams

def parse_str(str_): return minidom.parseString(str_).documentElement

class ConsecXmlInputStream(streams.TracedStream):

    '''Wraps an input stream, inserting EOF after each consecutive XML document

    '''

    def __init__(self, stream):

        self.eof = False

        self.root_close_tag = None

        def trace(line):

            if self.root_close_tag == None: # before XML document

                match = re.match(r'^<(?!\?)([^\s>]+)', line)

                if match: self.root_close_tag = '</'+match.group(1)+'>'

            else: # inside XML document

                if line.rstrip().endswith(self.root_close_tag):

                    self.root_close_tag = None # now outside XML document

                    self.eof = True # next read will produce EOF

        streams.TracedStream.__init__(self, trace, stream)

    def readline(self):

        if self.eof:

            self.eof = False

            return '' # don't read from underlying stream

        else: return streams.TracedStream.readline(self)

    def read(self, n): return self.readline() # forward all reads to readline()

def parse_next(stream, on_error=exc.raise_):

    '''Parses the next document in a consecutive sequence of XML documents.

    @return The next document's root, or None if no more documents available

    '''

    try: return minidom.parse(ConsecXmlInputStream(stream)).documentElement

    except expat.ExpatError, e:

        # Save in case another exception raised, overwriting sys.exc_info()

        exc.add_traceback(e)

        if str(e).startswith('no element found:'): return None # no more docs

        else: # parser error, which corrupts the rest of the current document

            on_error(e)

            streams.consume(stream) # advance to next XML document

def docs_iter(stream, *args, **kw_args):

    return iter(lambda: parse_next(stream, *args, **kw_args), None)