augment_data function in dictization_functions.py. I found the ##fill junk and extras portion confusing

dictization_functions.py

def augment_data(data, schema):
    '''add missing, extras and junk data'''
    flattented_schema = flatten_schema(schema)
    key_combinations = get_all_key_combinations(data, flattented_schema)

    full_schema = make_full_schema(data, schema)

    new_data = copy.copy(data)

    ## fill junk and extras

    for key, value in new_data.items():
        if key in full_schema:
            continue

        ## check if any thing naugthy is placed against subschemas
        initial_tuple = key[::2]
        if initial_tuple in [initial_key[:len(initial_tuple)]
                             for initial_key in flattented_schema]:
            if data[key] <> []:
                raise DataError('Only lists of dicts can be placed against '
                                'subschema %s, not %s' % (key,type(data[key])))

        if key[:-1] in key_combinations:
            extras_key = key[:-1] + ('__extras',)
            extras = new_data.get(extras_key, {})
            extras[key[-1]] = value
            new_data[extras_key] = extras
        else:
            junk = new_data.get(("__junk",), {})
            junk[key] = value
            new_data[("__junk",)] = junk
        new_data.pop(key)

    ## add missing

    for key, value in full_schema.items():
        if key not in new_data and not key[-1].startswith("__"):
            new_data[key] = missing

    return new_data