xml-lint

#!/usr/bin/python

import os.path
import re
import sys
from subprocess import check_output, check_call, CalledProcessError

misc_pattern = re.compile('<resource\s*url="(.+?)"\s*location="(.+?)"\s+\/>')
xsd_pattern = re.compile('xsi:noNamespaceSchemaLocation="(.+?)"')

def get_mappings(misc_file, base_dir):
    mapping = {}
    for line in open(misc_file):
        result = misc_pattern.search(line)
        if result is not None:
            mapping[result.group(1)] = result.group(2).replace('$PROJECT_DIR$', base_dir)
    return mapping

def get_xml_files(search_dir):
    try:
        return check_output(['find', search_dir, '-name', '*.xml']).splitlines()
    except CalledProcessError:
        print("WARNING: No XML files found")
        return []

def lint_only(file):
    try:
        check_call(['xmllint', '--noout', file])
    except CalledProcessError as err:
        exit(err.returncode)

def lint_with_xsd(file, xsd):
    try:
        check_call(['xmllint', '--noout', '--schema', xsd, file])
    except CalledProcessError as err:
        exit(err.returncode)

def search_file_for_xsd(file):
    handle = open(file)
    for line in handle:
        match = xsd_pattern.search(line)
        if match is not None:
            handle.close()
            return match.group(1)
    handle.close()
    return None

def validate_file(file, mapping):
    print("validating file {}".format(file))
    xsd = search_file_for_xsd(file)
    if xsd is not None:
        if xsd in mapping:
            lint_with_xsd(file, mapping[xsd])
        else:
            print('WARNING: Unable to map XSD to path: {}'.format(xsd))
    else:
        print('WARNING: Unable to find XSD for file: {}'.format(file))
    lint_only(file)

def main():
    if len(sys.argv) < 4:
        print("Usage: xml-lint <misc.xml> <app_base_dir> <search_dir>")
        exit(1)

    misc_file = sys.argv[1]
    base_dir = os.path.abspath(sys.argv[2])
    search_dir = os.path.abspath(sys.argv[3])

    if not os.path.isfile(misc_file):
        print("{} is not a file".format(misc_file))
        exit(1)

    mapping = get_mappings(misc_file, base_dir)
    print("loaded {} XSD mapping(s)".format(len(mapping)))

    xml_files = get_xml_files(search_dir)
    print("found {} XML file(s)".format(len(xml_files)))

    print("")
    for file in xml_files:
        validate_file(file, mapping)
        print("")

if __name__ == '__main__':
    main()