wesnoth/data/tools/wmlvalidator

#!/usr/bin/env python
"""
wmltest -- tool to validate the syntax and semantics of WML.

Use --help to see usage.
"""
#TODO:
#-define verbosity levels better

import wesnoth.wmldata as wmldata
import wesnoth.wmlparser as wmlparser
import wesnoth.wmlgrammar as wmlgrammar
import re

def print_indent(string, depth, char=' '):
    print "%s%s" % (depth * char, string)

class Validator:
    """
    The class that takes a wmlgrammar object to validate wml trees with
    """
    def __init__(self, schema, verbosity=0):
        self.schema = wmlgrammar.Grammar(schema)
        self.verbosity = verbosity

    def validate(self, node, depth=0, name=None):
        """
        Validate the given DataSub node.
        depth indicates how deep we've recursed into the tree,
        name is a mechanism for overwriting the node name to look up in the schema, used for overloaded names.
        """
        if not name or name == node.name:
            name = node.name
            verbosename = name
        else:
            verbosename = "%s (%s)" % (node.name, name)

        if self.verbosity > 1:
            print_indent(node.name, depth)

        try:
            schema = self.schema.get_element(name)
        except KeyError:
            print "No valid schema found for %s" % verbosename
            return


        # TODO: the blocks below probably need to be rewritten

        # Validate the attributes
        for attribute in schema.get_attributes():
            matches = node.get_texts(attribute.name)
            nummatches = len(matches)
            if attribute.freq == wmlgrammar.REQUIRED and nummatches != 1:
                print "(%s:%d) Attribute '[%s] %s' should appear exactly once, not %d times" % (node.file, node.line, verbosename, attribute.name, nummatches)
            elif attribute.freq == wmlgrammar.OPTIONAL and nummatches > 1:
                print "(%s:%d) Attribute '[%s] %s' should appear at most once, not %d times" % (node.file, node.line, verbosename, attribute.name, nummatches)
            elif attribute.freq == wmlgrammar.FORBIDDEN and nummatches > 0:
                print "(%s:%d) Attribute '[%s] %s' should not appear. It appears %d times" % (node.file, node.line, verbosename, attribute.name, nummatches)
            for match in matches:
                if not attribute.validate(match.data):
                    print "(%s:%d) Attribute '[%s] %s's value should be %s, found: %s" % (node.file, node.line, verbosename, attribute.name, attribute.type, match.data)
                node.remove(match) # Get rid of these so we can see what's left
        for attribute in node.get_all_text():
            print "(%s:%d) Attribute '[%s] %s' found, which has no meaning there" % (node.file, node.line, verbosename, attribute.name)

        # Validate the elements
        for element in schema.get_elements():
            matches = node.get_subs(element.name)
            nummatches = len(matches)
            if element.freq == wmlgrammar.REQUIRED and nummatches != 1:
                print "(%s:%d) Element '[%s] [%s]' should appear exactly once, not %d times" % (node.file, node.line, verbosename, element.name, nummatches)
            elif element.freq == wmlgrammar.OPTIONAL and nummatches > 1:
                print "(%s:%d) Element '[%s] [%s]' should appear at most once, not %d times" % (node.file, node.line, verbosename, element.name, nummatches)
            for match in matches:
                self.validate(match, depth+1, element.subname)
                node.remove(match)

        for element in node.get_all_subs():
            print "(%s:%d) Element '[%s] [%s]' found, which has no meaning there" % (node.file, node.line, verbosename, element.name)
            # Do we want to do this?
            if False:
                print "Attempting to validate [%s] anyway" % element.name
                self.validate(element, depth+1)

if __name__ == '__main__':
    import argparse, subprocess, os, codecs, sys

    # Ugly hack to force the output of UTF-8.
    # This prevents us from crashing when we're being verbose
    #  and encounter a non-ascii character.
    sys.stdout = codecs.getwriter('utf-8')(sys.stdout)

    ap = argparse.ArgumentParser("Usage: %(prog)s [options]")
    ap.add_argument("-p", "--path",
        help = "Specify Wesnoth's data dir",
        dest = "path")
    ap.add_argument("-u", "--userpath",
        help = "Specify user data dir",
        dest = "userpath")
    ap.add_argument("-s", "--schema",
        help = "Specify WML schema",
        dest = "schema")
    ap.add_argument("-v", "--verbose",
        action = "count",
        dest = "verbose",
        help = "Increase verbosity, 4 is the maximum.")
    ap.add_argument("-D", "--define",
        action = "append",
        dest = "defines",
        default = [],
        help = "Define (empty) preprocessor macros, for campaign/multiplayer inclusion.")
    ap.add_argument("filename",
        nargs = "*",
        help = "Files to validate")
    args = ap.parse_args()

    if not args.path:
        try:
            p = subprocess.Popen(["wesnoth", "--path"], stdout = subprocess.PIPE)
            path = p.stdout.read().strip()
            args.path = os.path.join(path, "data")
            sys.stderr.write("No Wesnoth path given.\nAutomatically found '%s'\n" % (args.path, ) )
        except OSError:
            args.path = '.'
            sys.stderr.write("Could not determine Wesnoth path.\nAssuming '%s'\n" % (args.path, ) )

    if len(args.filename) < 1:
        args.filename.append(os.path.join(args.path, '_main.cfg'))

    if args.verbose > 1:
        print "Args: %s\n"% (args, )

    if not args.schema:
        args.schema = os.path.join(args.path, 'schema.cfg')

    # Parse the schema
    parser = wmlparser.Parser(args.path)

    if args.verbose > 3:
        parser.verbose = True
    parser.parse_file(args.schema)

    schema = wmldata.DataSub("schema")
    parser.parse_top(schema)

    # Construct the validator
    validator = Validator(schema, args.verbose)

    # Parse the WML
    parser = wmlparser.Parser(args.path, args.userpath)

    if args.verbose > 3:
        parser.verbose = True

    if args.userpath:
        parser.parse_text("{~add-ons}")
    for file in args.filename:
        parser.parse_file(file)
    for macro in args.defines:
        parser.parse_text("#define %s \n#enddef" % (macro, ) )

    data = wmldata.DataSub("root")
    parser.parse_top(data)

    # Validate
    validator.validate(data)

# vim: tabstop=4: shiftwidth=4: expandtab: softtabstop=4: autoindent: