gogrepo.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-
from __future__ import print_function
from __future__ import division
from __future__ import unicode_literals

__appname__ = 'gogrepo.py'
__author__ = 'eddie3'
__version__ = '0.3a'
__url__ = 'https://github.com/eddie3/gogrepo'

# imports
import os
import sys
import threading
import logging
import contextlib
import json
import html5lib
import pprint
import time
import zipfile
import hashlib
import getpass
import argparse
import codecs
import io
import datetime
import shutil
import socket
import xml.etree.ElementTree

# python 2 / 3 imports
try:
    # python 2
    from Queue import Queue
    import cookielib as cookiejar
    from httplib import BadStatusLine
    from urlparse import urlparse
    from urllib import urlencode
    from urllib2 import HTTPError, URLError, HTTPCookieProcessor, build_opener, Request
    from itertools import izip_longest as zip_longest
    from StringIO import StringIO
except ImportError:
    # python 3
    from queue import Queue
    import http.cookiejar as cookiejar
    from http.client import BadStatusLine
    from urllib.parse import urlparse, urlencode
    from urllib.request import HTTPCookieProcessor, HTTPError, URLError, build_opener, Request
    from itertools import zip_longest
    from io import StringIO

# python 2 / 3 renames
try: input = raw_input
except NameError: pass

# optional imports
try:
    from html2text import html2text
except ImportError:
    def html2text(x): return x

# lib mods
cookiejar.MozillaCookieJar.magic_re = r'.*'  # bypass the hardcoded "Netscape HTTP Cookie File" check

# configure logging
logFormatter = logging.Formatter("%(asctime)s | %(message)s", datefmt='%H:%M:%S')
rootLogger = logging.getLogger('ws')
rootLogger.setLevel(logging.DEBUG)
consoleHandler = logging.StreamHandler(sys.stdout)
consoleHandler.setFormatter(logFormatter)
rootLogger.addHandler(consoleHandler)

# logging aliases
info = rootLogger.info
warn = rootLogger.warning
debug = rootLogger.debug
error = rootLogger.error
log_exception = rootLogger.exception

# filepath constants
GAME_STORAGE_DIR = r'.'
COOKIES_FILENAME = r'gog-cookies.dat'
MANIFEST_FILENAME = r'gog-manifest.dat'
SERIAL_FILENAME = r'!serial.txt'
INFO_FILENAME = r'!info.txt'

# global web utilities
global_cookies = cookiejar.LWPCookieJar(COOKIES_FILENAME)
cookieproc = HTTPCookieProcessor(global_cookies)
opener = build_opener(cookieproc)
treebuilder = html5lib.treebuilders.getTreeBuilder('etree')
parser = html5lib.HTMLParser(tree=treebuilder, namespaceHTMLElements=False)

# GOG URLs
GOG_HOME_URL = r'https://www.gog.com'
GOG_ACCOUNT_URL = r'https://www.gog.com/account'
GOG_LOGIN_URL = r'https://login.gog.com/login_check'

# GOG Constants
GOG_MEDIA_TYPE_GAME  = '1'
GOG_MEDIA_TYPE_MOVIE = '2'

# HTTP request settings
HTTP_FETCH_DELAY = 1   # in seconds
HTTP_RETRY_DELAY = 5   # in seconds
HTTP_RETRY_COUNT = 3
HTTP_GAME_DOWNLOADER_THREADS = 4
HTTP_PERM_ERRORCODES = (404, 403, 503)

# Save manifest data for these os and lang combinations
DEFAULT_OS_LIST = ['windows']
DEFAULT_LANG_LIST = ['en']

# These file types don't have md5 data from GOG
SKIP_MD5_FILE_EXT = ['.txt', '.zip']

# Language table that maps two letter language to their unicode gogapi json name
LANG_TABLE = {'en': u'English',   # English
              'bl': u'\u0431\u044a\u043b\u0433\u0430\u0440\u0441\u043a\u0438',  # Bulgarian
              'ru': u'\u0440\u0443\u0441\u0441\u043a\u0438\u0439',              # Russian
              'gk': u'\u0395\u03bb\u03bb\u03b7\u03bd\u03b9\u03ba\u03ac',        # Greek
              'sb': u'\u0421\u0440\u043f\u0441\u043a\u0430',                    # Serbian
              'ar': u'\u0627\u0644\u0639\u0631\u0628\u064a\u0629',              # Arabic
              'br': u'Portugu\xeas do Brasil',  # Brazilian Portuguese
              'jp': u'\u65e5\u672c\u8a9e',      # Japanese
              'ko': u'\ud55c\uad6d\uc5b4',      # Korean
              'fr': u'fran\xe7ais',             # French
              'cn': u'\u4e2d\u6587',            # Chinese
              'cz': u'\u010desk\xfd',           # Czech
              'hu': u'magyar',                  # Hungarian
              'pt': u'portugu\xeas',            # Portuguese
              'tr': u'T\xfcrk\xe7e',            # Turkish
              'sk': u'slovensk\xfd',            # Slovak
              'nl': u'nederlands',              # Dutch
              'ro': u'rom\xe2n\u0103',          # Romanian
              'es': u'espa\xf1ol',      # Spanish
              'pl': u'polski',          # Polish
              'it': u'italiano',        # Italian
              'de': u'Deutsch',         # German
              'da': u'Dansk',           # Danish
              'sv': u'svenska',         # Swedish
              'fi': u'Suomi',           # Finnish
              'no': u'norsk',           # Norsk
              }

VALID_OS_TYPES = ['windows', 'linux', 'mac']
VALID_LANG_TYPES = list(LANG_TABLE.keys())

ORPHAN_DIR_NAME = '!orphaned'
ORPHAN_DIR_EXCLUDE_LIST = [ORPHAN_DIR_NAME, '!misc']
ORPHAN_FILE_EXCLUDE_LIST = [INFO_FILENAME, SERIAL_FILENAME]

def request(url, args=None, byte_range=None, retries=HTTP_RETRY_COUNT, delay=HTTP_FETCH_DELAY):
    """Performs web request to url with optional retries, delay, and byte range.
    """
    _retry = False
    time.sleep(delay)

    try:
        if args is not None:
            enc_args = urlencode(args)
            enc_args = enc_args.encode('ascii') # needed for Python 3
        else:
            enc_args = None
        req = Request(url, data=enc_args)
        if byte_range is not None:
            req.add_header('Range', 'bytes=%d-%d' % byte_range)
        page = opener.open(req)
    except (HTTPError, URLError, socket.error, BadStatusLine) as e:
        if isinstance(e, HTTPError):
            if e.code in HTTP_PERM_ERRORCODES:  # do not retry these HTTP codes
                warn('request failed: %s.  will not retry.', e)
                raise
        if retries > 0:
            _retry = True
        else:
            raise

        if _retry:
            warn('request failed: %s (%d retries left) -- will retry in %ds...' % (e, retries, HTTP_RETRY_DELAY))
            return request(url=url, args=args, byte_range=byte_range, retries=retries-1, delay=HTTP_RETRY_DELAY)

    return contextlib.closing(page)


# --------------------------
# Helper types and functions
# --------------------------
class AttrDict(dict):
    def __init__(self, **kw):
        self.update(kw)

    def __getattr__(self, key):
        return self[key]

    def __setattr__(self, key, val):
        self[key] = val

class ConditionalWriter(object):
    """File writer that only updates file on disk if contents chanaged"""

    def __init__(self, filename):
        self._buffer = None
        self._filename = filename

    def __enter__(self):
        self._buffer = tmp = StringIO()
        return tmp

    def __exit__(self, _exc_type, _exc_value, _traceback):
        tmp = self._buffer
        if tmp:
            pos = tmp.tell()
            tmp.seek(0)

            file_changed = not os.path.exists(self._filename)
            if not file_changed:
                with codecs.open(self._filename, 'r', 'utf-8') as orig:
                    for (new_chunk, old_chunk) in zip_longest(tmp, orig):
                        if new_chunk != old_chunk:
                            file_changed = True
                            break

            if file_changed:
                with codecs.open(self._filename, 'w', 'utf-8') as overwrite:
                    tmp.seek(0)
                    shutil.copyfileobj(tmp, overwrite)

def load_cookies():
    # try to load as default lwp format
    try:
        global_cookies.load()
        return
    except IOError:
        pass

    # try to import as mozilla 'cookies.txt' format
    try:
        tmp_jar = cookiejar.MozillaCookieJar(global_cookies.filename)
        tmp_jar.load()
        for c in tmp_jar:
            global_cookies.set_cookie(c)
        global_cookies.save()
        return
    except IOError:
        pass

    error('failed to load cookies, did you login first?')
    raise SystemExit(1)


def load_manifest(filepath=MANIFEST_FILENAME):
    info('loading local manifest...')
    try:
        with codecs.open(MANIFEST_FILENAME, 'rU', 'utf-8') as r:
            ad = r.read().replace('{', 'AttrDict(**{').replace('}', '})')
        return eval(ad)
    except IOError:
        return []


def save_manifest(items):
    info('saving manifest...')
    with codecs.open(MANIFEST_FILENAME, 'w', 'utf-8') as w:
        print('# {} games'.format(len(items)), file=w)
        pprint.pprint(items, width=123, stream=w)


def open_notrunc(name, bufsize=4*1024):
    flags = os.O_WRONLY | os.O_CREAT
    if hasattr(os, "O_BINARY"):
        flags |= os.O_BINARY  # windows
    fd = os.open(name, flags, 0o666)
    return os.fdopen(fd, 'wb', bufsize)


def hashfile(afile, blocksize=65536):
    afile = open(afile, 'rb')
    hasher = hashlib.md5()
    buf = afile.read(blocksize)
    while len(buf) > 0:
        hasher.update(buf)
        buf = afile.read(blocksize)
    return hasher.hexdigest()


def test_zipfile(filename):
    """Opens filename and tests the file for ZIP integrity.  Returns True if
    zipfile passes the integrity test, False otherwise.
    """
    try:
        with zipfile.ZipFile(filename, 'r') as f:
            if f.testzip() is None:
                return True
    except zipfile.BadZipfile:
        return False

    return False


def pretty_size(n):
    for unit in ['B', 'KB', 'MB', 'GB', 'TB']:
        if n < 1024 or unit == 'TB':
            break
        n = n / 1024  # start at KB

    if unit == 'B':
        return "{0}{1}".format(n, unit)
    else:
        return "{0:.2f}{1}".format(n, unit)


def get_total_size(dir):
    total = 0
    for (root, dirnames, filenames) in os.walk(dir):
        for f in filenames:
            total += os.path.getsize(os.path.join(root, f))
    return total


def item_checkdb(search_id, gamesdb):
    for i in range(len(gamesdb)):
        if search_id == gamesdb[i].id:
            return i
    return None


def handle_game_updates(olditem, newitem):
    if newitem.has_updates:
        info('  -> gog flagged this game as updated')

    if olditem.title != newitem.title:
        info('  -> title has changed "{}" -> "{}"'.format(olditem.title, newitem.title))
        # TODO: rename the game directory

    if olditem.long_title != newitem.long_title:
        try:
            info('  -> long title has change "{}" -> "{}"'.format(olditem.long_title, newitem.long_title))
        except UnicodeEncodeError:
            pass

    if olditem.changelog != newitem.changelog and newitem.changelog not in [None, '']:
        info('  -> changelog was updated')

    if olditem.serial != newitem.serial:
        info('  -> serial key has changed')


def fetch_file_info(d, fetch_md5):
    # fetch file name/size
    with request(d.href, byte_range=(0, 0)) as page:
        d.name = urlparse(page.geturl()).path.split('/')[-1]
        d.size = int(page.headers['Content-Range'].split('/')[-1])

        # fetch file md5
        if fetch_md5:
            if os.path.splitext(page.geturl())[1].lower() not in SKIP_MD5_FILE_EXT:
                tmp_md5_url = page.geturl().replace('?', '.xml?')
                try:
                    with request(tmp_md5_url) as page:
                        shelf_etree = xml.etree.ElementTree.parse(page).getroot()
                        d.md5 = shelf_etree.attrib['md5']
                except HTTPError as e:
                    if e.code == 404:
                        warn("no md5 data found for {}".format(d.name))
                    else:
                        raise
                except xml.etree.ElementTree.ParseError:
                    warn('xml parsing error occurred trying to get md5 data for {}'.format(d.name))


def filter_downloads(out_list, downloads_list, lang_list, os_list):
    """filters any downloads information against matching lang and os, translates
    them, and extends them into out_list
    """
    filtered_downloads = []
    downloads_dict = dict(downloads_list)

    # hold list of valid languages languages as known by gogapi json stuff
    valid_langs = []
    for lang in lang_list:
        valid_langs.append(LANG_TABLE[lang])

    # check if lang/os combo passes the specified filter
    for lang in downloads_dict:
        if lang in valid_langs:
            for os_type in downloads_dict[lang]:
                if os_type in os_list:
                    for download in downloads_dict[lang][os_type]:
                        # passed the filter, create the entry
                        d = AttrDict(desc=download['name'],
                                     os_type=os_type,
                                     lang=lang,
                                     version=download['version'],
                                     href=GOG_HOME_URL + download['manualUrl'],
                                     md5=None,
                                     name=None,
                                     size=None
                                     )
                        try:
                            fetch_file_info(d, True)
                        except HTTPError:
                            warn("failed to fetch %s" % d.href)
                        filtered_downloads.append(d)

    out_list.extend(filtered_downloads)


def filter_extras(out_list, extras_list):
    """filters and translates extras information and adds them into out_list
    """
    filtered_extras = []

    for extra in extras_list:
        d = AttrDict(desc=extra['name'],
                     os_type='extra',
                     lang='',
                     version=None,
                     href=GOG_HOME_URL + extra['manualUrl'],
                     md5=None,
                     name=None,
                     size=None,
                     )
        try:
            fetch_file_info(d, False)
        except HTTPError:
            warn("failed to fetch %s" % d.href)
        filtered_extras.append(d)

    out_list.extend(filtered_extras)


def filter_dlcs(item, dlc_list, lang_list, os_list):
    """filters any downloads/extras information against matching lang and os, translates
    them, and adds them to the item downloads/extras

    dlcs can contain dlcs in a recursive fashion, and oddly GOG does do this for some titles.
    """
    for dlc_dict in dlc_list:
        filter_downloads(item.downloads, dlc_dict['downloads'], lang_list, os_list)
        filter_extras(item.extras, dlc_dict['extras'])
        filter_dlcs(item, dlc_dict['dlcs'], lang_list, os_list)  # recursive


def process_argv(argv):
    p1 = argparse.ArgumentParser(description='%s (%s)' % (__appname__, __url__), add_help=False)
    sp1 = p1.add_subparsers(help='commands', dest='cmd', title='commands')

    g1 = sp1.add_parser('login', help='Login to GOG and save a local copy of your authenticated cookie')
    g1.add_argument('username', action='store', help='GOG username/email', nargs='?', default=None)
    g1.add_argument('password', action='store', help='GOG password', nargs='?', default=None)

    g1 = sp1.add_parser('update', help='Update locally saved game manifest from GOG server')
    g1.add_argument('-os', action='store', help='operating system(s)', nargs='*', default=DEFAULT_OS_LIST)
    g1.add_argument('-lang', action='store', help='game language(s)', nargs='*', default=DEFAULT_LANG_LIST)
    g2 = g1.add_mutually_exclusive_group()  # below are mutually exclusive
    g2.add_argument('-skipknown', action='store_true', help='skip games already known by manifest')
    g2.add_argument('-updateonly', action='store_true', help='only games marked with the update tag')
    g2.add_argument('-id', action='store', help='id/dirname of a specific game to update')

    g1 = sp1.add_parser('download', help='Download all your GOG games and extra files')
    g1.add_argument('savedir', action='store', help='directory to save downloads to', nargs='?', default='.')
    g1.add_argument('-dryrun', action='store_true', help='display, but skip downloading of any files')
    g1.add_argument('-skipextras', action='store_true', help='skip downloading of any GOG extra files')
    g1.add_argument('-skipgames', action='store_true', help='skip downloading of any GOG game files')
    g1.add_argument('-id', action='store', help='id of the game in the manifest to download')
    g1.add_argument('-wait', action='store', type=float,
                    help='wait this long in hours before starting', default=0.0)  # sleep in hr
    g1.add_argument('-skipids', action='store', help='id[s] of the game[s] in the manifest to NOT download')

    g1 = sp1.add_parser('import', help='Import files with any matching MD5 checksums found in manifest')
    g1.add_argument('src_dir', action='store', help='source directory to import games from')
    g1.add_argument('dest_dir', action='store', help='directory to copy and name imported files to')

    g1 = sp1.add_parser('backup', help='Perform an incremental backup to specified directory')
    g1.add_argument('src_dir', action='store', help='source directory containing gog items')
    g1.add_argument('dest_dir', action='store', help='destination directory to backup files to')

    g1 = sp1.add_parser('verify', help='Scan your downloaded GOG files and verify their size, MD5, and zip integrity')
    g1.add_argument('gamedir', action='store', help='directory containing games to verify', nargs='?', default='.')
    g1.add_argument('-id', action='store', help='id of a specific game to verify')
    g1.add_argument('-skipmd5', action='store_true', help='do not perform MD5 check')
    g1.add_argument('-skipsize', action='store_true', help='do not perform size check')
    g1.add_argument('-skipzip', action='store_true', help='do not perform zip integrity check')
    g1.add_argument('-delete', action='store_true', help='delete any files which fail integrity test')

    g1 = sp1.add_parser('clean', help='Clean your games directory of files not known by manifest')
    g1.add_argument('cleandir', action='store', help='root directory containing gog games to be cleaned')
    g1.add_argument('-dryrun', action='store_true', help='do not move files, only display what would be cleaned')

    g1 = p1.add_argument_group('other')
    g1.add_argument('-h', '--help', action='help', help='show help message and exit')
    g1.add_argument('-v', '--version', action='version', help='show version number and exit',
                    version="%s (version %s)" % (__appname__, __version__))

    # parse the given argv.  raises SystemExit on error
    args = p1.parse_args(argv[1:])

    if args.cmd == 'update':
        for lang in args.lang:  # validate the language
            if lang not in VALID_LANG_TYPES:
                error('error: specified language "%s" is not one of the valid languages %s' % (lang, VALID_LANG_TYPES))
                raise SystemExit(1)

        for os_type in args.os:  # validate the os type
            if os_type not in VALID_OS_TYPES:
                error('error: specified os "%s" is not one of the valid os types %s' % (os_type, VALID_OS_TYPES))
                raise SystemExit(1)

    return args


# --------
# Commands
# --------
def cmd_login(user, passwd):
    """Attempts to log into GOG and saves the resulting cookiejar to disk.
    """
    login_data = {'user': user,
                  'passwd': passwd,
                  'auth_url': None,
                  'login_token': None,
                  'two_step_url': None,
                  'two_step_token': None,
                  'two_step_security_code': None,
                  'login_success': False,
                  }

    global_cookies.clear()  # reset cookiejar

    # prompt for login/password if needed
    if login_data['user'] is None:
        login_data['user'] = input("Username: ")
    if login_data['passwd'] is None:
        login_data['passwd'] = getpass.getpass()

    info("attempting gog login as '{}' ...".format(login_data['user']))

    # fetch the auth url
    with request(GOG_HOME_URL, delay=0) as page:
        etree = html5lib.parse(page, namespaceHTMLElements=False)
        for elm in etree.findall('.//script'):
            if elm.text is not None and 'GalaxyAccounts' in elm.text:
                login_data['auth_url'] = elm.text.split("'")[3]
                break

    # fetch the login token
    with request(login_data['auth_url'], delay=0) as page:
        etree = html5lib.parse(page, namespaceHTMLElements=False)
        # Bail if we find a request for a reCAPTCHA
        if len(etree.findall('.//div[@class="g-recaptcha"]')) > 0:
            error("cannot continue, gog is asking for a reCAPTCHA :(  try again in a few minutes.")
            return
        for elm in etree.findall('.//input'):
            if elm.attrib['id'] == 'login__token':
                login_data['login_token'] = elm.attrib['value']
                break

    # perform login and capture two-step token if required
    with request(GOG_LOGIN_URL, delay=0, args={'login[username]': login_data['user'],
                                               'login[password]': login_data['passwd'],
                                               'login[login]': '',
                                               'login[_token]': login_data['login_token']}) as page:
        etree = html5lib.parse(page, namespaceHTMLElements=False)
        if 'two_step' in page.geturl():
            login_data['two_step_url'] = page.geturl()
            for elm in etree.findall('.//input'):
                if elm.attrib['id'] == 'second_step_authentication__token':
                    login_data['two_step_token'] = elm.attrib['value']
                    break
        elif 'on_login_success' in page.geturl():
            login_data['login_success'] = True

    # perform two-step if needed
    if login_data['two_step_url'] is not None:
        login_data['two_step_security_code'] = input("enter two-step security code: ")

        # Send the security code back to GOG
        with request(login_data['two_step_url'], delay=0,
                     args={'second_step_authentication[token][letter_1]': login_data['two_step_security_code'][0],
                           'second_step_authentication[token][letter_2]': login_data['two_step_security_code'][1],
                           'second_step_authentication[token][letter_3]': login_data['two_step_security_code'][2],
                           'second_step_authentication[token][letter_4]': login_data['two_step_security_code'][3],
                           'second_step_authentication[send]': "",
                           'second_step_authentication[_token]': login_data['two_step_token']}) as page:
            if 'on_login_success' in page.geturl():
                login_data['login_success'] = True

    # save cookies on success
    if login_data['login_success']:
        info('login successful!')
        global_cookies.save()
    else:
        error('login failed, verify your username/password and try again.')


def cmd_update(os_list, lang_list, skipknown, updateonly, id):
    media_type = GOG_MEDIA_TYPE_GAME
    items = []
    known_ids = []
    i = 0

    load_cookies()

    gamesdb = load_manifest()

    api_url  = GOG_ACCOUNT_URL
    api_url += "/getFilteredProducts"

    # Make convenient list of known ids
    if skipknown:
        for item in gamesdb:
            known_ids.append(item.id)

    # Fetch shelf data
    done = False
    while not done:
        i += 1  # starts at page 1
        if i == 1:
            info('fetching game product data (page %d)...' % i)
        else:
            info('fetching game product data (page %d / %d)...' % (i, json_data['totalPages']))
        with request(api_url, args={'mediaType': media_type,
                                    'sortBy': 'title',  # sort order
                                    'page': str(i)}, delay=0) as data_request:
            reader = codecs.getreader("utf-8")
            try:
                json_data = json.load(reader(data_request))
            except ValueError:
                error('failed to load product data (are you still logged in?)')
                raise SystemExit(1)

            # Parse out the interesting fields and add to items dict
            for item_json_data in json_data['products']:
                # skip games marked as hidden
                if item_json_data.get('isHidden', False) is True:
                    continue

                item = AttrDict()
                item.id = item_json_data['id']
                item.title = item_json_data['slug']
                item.long_title = item_json_data['title']
                item.genre = item_json_data['category']
                item.image_url = item_json_data['image']
                item.store_url = item_json_data['url']
                item.media_type = media_type
                item.rating = item_json_data['rating']
                item.has_updates = bool(item_json_data['updates']) or bool(item_json_data['isNew'])

                if id:
                    if item.title == id or str(item.id) == id:  # support by game title or gog id
                        info('found "{}" in product data!'.format(item.title))
                        items.append(item)
                        done = True
                elif updateonly:
                    if item.has_updates:
                        items.append(item)
                elif skipknown:
                    if item.id not in known_ids:
                        items.append(item)
                else:
                    items.append(item)

            if i >= json_data['totalPages']:
                done = True

    # bail if there's nothing to do
    if len(items) == 0:
        if id:
            warn('game id "{}" was not found in your product data'.format(id))
        elif updateonly:
            warn('no new game updates found.')
        elif skipknown:
            warn('no new games found.')
        else:
            warn('nothing to do')
        return

    items_count = len(items)
    print_padding = len(str(items_count))
    if not id and not updateonly and not skipknown:
        info('found %d games !!%s' % (items_count, '!'*int(items_count/100)))  # teehee

    # fetch item details
    i = 0
    for item in sorted(items, key=lambda item: item.title):
        api_url  = GOG_ACCOUNT_URL
        api_url += "/gameDetails/{}.json".format(item.id)

        i += 1
        info("(%*d / %d) fetching game details for %s..." % (print_padding, i, items_count, item.title))

        try:
            with request(api_url) as data_request:
                reader = codecs.getreader("utf-8")
                item_json_data = json.load(reader(data_request))

                item.bg_url = item_json_data['backgroundImage']
                item.serial = item_json_data['cdKey']
                item.forum_url = item_json_data['forumLink']
                item.changelog = item_json_data['changelog']
                item.release_timestamp = item_json_data['releaseTimestamp']
                item.gog_messages = item_json_data['messages']
                item.downloads = []
                item.extras = []

                # parse json data for downloads/extras/dlcs
                filter_downloads(item.downloads, item_json_data['downloads'], lang_list, os_list)
                filter_extras(item.extras, item_json_data['extras'])
                filter_dlcs(item, item_json_data['dlcs'], lang_list, os_list)

                # update gamesdb with new item
                item_idx = item_checkdb(item.id, gamesdb)
                if item_idx is not None:
                    handle_game_updates(gamesdb[item_idx], item)
                    gamesdb[item_idx] = item
                else:
                    gamesdb.append(item)

        except Exception:
            log_exception('error')

    # save the manifest to disk
    save_manifest(gamesdb)


def cmd_import(src_dir, dest_dir):
    """Recursively finds all files within root_dir and compares their MD5 values
    against known md5 values from the manifest.  If a match is found, the file will be copied
    into the game storage dir.
    """
    gamesdb = load_manifest()

    info("collecting md5 data out of the manifest")
    md5_info = {}  # holds tuples of (title, filename) with md5 as key

    for game in gamesdb:
        for game_item in game.downloads:
            if game_item.md5 is not None:
                md5_info[game_item.md5] = (game.title, game_item.name)

    info("searching for files within '%s'" % src_dir)
    file_list = []
    for (root, dirnames, filenames) in os.walk(src_dir):
        for f in filenames:
            if os.path.splitext(f)[1].lower() not in SKIP_MD5_FILE_EXT:
                file_list.append(os.path.join(root, f))

    info("comparing md5 file hashes")
    for f in file_list:
        fname = os.path.basename(f)
        info("calculating md5 for '%s'" % fname)
        h = hashfile(f)
        if h in md5_info:
            title, fname = md5_info[h]
            src_dir = os.path.join(dest_dir, title)
            dest_file = os.path.join(src_dir, fname)
            info('found a match! [%s] -> %s' % (h, fname))
            if os.path.isfile(dest_file):
                if h == hashfile(dest_file):
                    info('destination file already exists with the same md5 value.  skipping copy.')
                    continue
            info("copying to %s..." % dest_file)
            if not os.path.isdir(src_dir):
                os.makedirs(src_dir)
            shutil.copy(f, dest_file)


def cmd_download(savedir, skipextras, skipgames, skipids, dryrun, id):
    sizes, rates, errors = {}, {}, {}
    work = Queue()  # build a list of work items

    load_cookies()

    items = load_manifest()
    work_dict = dict()

    # util
    def megs(b):
        return '%.1fMB' % (b / float(1024**2))
    def gigs(b):
        return '%.2fGB' % (b / float(1024**3))

    if id:
        id_found = False
        for item in items:
            if item.title == id:
                items = [item]
                id_found = True
                break
        if not id_found:
            error('no game with id "{}" was found.'.format(id))
            exit(1)

    if skipids:
        info("skipping games with id[s]: {%s}" % skipids)
        ignore_list = skipids.split(",")
        items[:] = [item for item in items if item.title not in ignore_list]

    # Find all items to be downloaded and push into work queue
    for item in sorted(items, key=lambda g: g.title):
        info("{%s}" % item.title)
        item_homedir = os.path.join(savedir, item.title)
        if not dryrun:
            if not os.path.isdir(item_homedir):
                os.makedirs(item_homedir)

        if skipextras:
            item.extras = []

        if skipgames:
            item.downloads = []

        # Generate and save a game info text file
        if not dryrun:
            with ConditionalWriter(os.path.join(item_homedir, INFO_FILENAME)) as fd_info:
                fd_info.write(u'{0}-- {1} --{0}{0}'.format(os.linesep, item.long_title))
                fd_info.write(u'title.......... {}{}'.format(item.title, os.linesep))
                if item.genre:
                    fd_info.write(u'genre.......... {}{}'.format(item.genre, os.linesep))
                fd_info.write(u'game id........ {}{}'.format(item.id, os.linesep))
                fd_info.write(u'url............ {}{}'.format(GOG_HOME_URL + item.store_url, os.linesep))
                if item.rating > 0:
                    fd_info.write(u'user rating.... {}%{}'.format(item.rating * 2, os.linesep))
                if item.release_timestamp > 0:
                    rel_date = datetime.datetime.fromtimestamp(item.release_timestamp).strftime('%B %d, %Y')
                    fd_info.write(u'release date... {}{}'.format(rel_date, os.linesep))
                if hasattr(item, 'gog_messages') and item.gog_messages:
                    fd_info.write(u'{0}gog messages...:{0}'.format(os.linesep))
                    for gog_msg in item.gog_messages:
                        fd_info.write(u'{0}{1}{0}'.format(os.linesep, html2text(gog_msg).strip()))
                fd_info.write(u'{0}game items.....:{0}{0}'.format(os.linesep))
                for game_item in item.downloads:
                    fd_info.write(u'    [{}] -- {}{}'.format(game_item.name, game_item.desc, os.linesep))
                    if game_item.version:
                        fd_info.write(u'        version: {}{}'.format(game_item.version, os.linesep))
                if len(item.extras) > 0:
                    fd_info.write(u'{0}extras.........:{0}{0}'.format(os.linesep))
                    for game_item in item.extras:
                        fd_info.write(u'    [{}] -- {}{}'.format(game_item.name, game_item.desc, os.linesep))
                if item.changelog:
                    fd_info.write(u'{0}changelog......:{0}{0}'.format(os.linesep))
                    fd_info.write(html2text(item.changelog).strip())
                    fd_info.write(os.linesep)
        # Generate and save a game serial text file
        if not dryrun:
            if item.serial != '':
                with ConditionalWriter(os.path.join(item_homedir, SERIAL_FILENAME)) as fd_serial:
                    item.serial = item.serial.replace(u'<span>', '')
                    item.serial = item.serial.replace(u'</span>', os.linesep)
                    fd_serial.write(item.serial)

        # Populate queue with all files to be downloaded
        for game_item in item.downloads + item.extras:
            if game_item.name is None:
                continue  # no game name, usually due to 404 during file fetch
            dest_file = os.path.join(item_homedir, game_item.name)

            if os.path.isfile(dest_file):
                if game_item.size is None:
                    warn('     unknown    %s has no size info.  skipping')
                    continue
                elif game_item.size != os.path.getsize(dest_file):
                    warn('     fail       %s has incorrect size.' % game_item.name)
                else:
                    info('     pass       %s' % game_item.name)
                    continue  # move on to next game item

            info('     download   %s' % game_item.name)
            sizes[dest_file] = game_item.size

            work_dict[dest_file] = (game_item.href, game_item.size, 0, game_item.size-1, dest_file)

    for work_item in work_dict:
        work.put(work_dict[work_item])

    if dryrun:
        info("{} left to download".format(gigs(sum(sizes.values()))))
        return  # bail, as below just kicks off the actual downloading

    info('-'*60)

    # work item I/O loop
    def ioloop(tid, path, page, out):
        sz, t0 = True, time.time()
        while sz:
            buf = page.read(4*1024)
            t = time.time()
            out.write(buf)
            sz, dt, t0 = len(buf), t - t0, t
            with lock:
                sizes[path] -= sz
                rates.setdefault(path, []).append((tid, (sz, dt)))

    # downloader worker thread main loop
    def worker():
        tid = threading.current_thread().ident
        while not work.empty():
            (href, sz, start, end, path) = work.get()
            try:
                dest_dir = os.path.dirname(path)
                with lock:
                    if not os.path.isdir(dest_dir):
                        os.makedirs(dest_dir)
                    if os.path.exists(path) and os.path.getsize(path) > sz:  # if needed, truncate file if ours is larger than expected size
                        with open_notrunc(path) as f:
                            f.truncate(sz)
                with open_notrunc(path) as out:
                    out.seek(start)
                    se = start, end
                    try:
                        with request(href, byte_range=se) as page:
                            hdr = page.headers['Content-Range'].split()[-1]
                            if hdr != '%d-%d/%d' % (start, end, sz):
                                with lock:
                                    error("chunk request has unexpected Content-Range. "
                                          "expected '%d-%d/%d' received '%s'. skipping."
                                          % (start, end, sz, hdr))
                            else:
                                assert out.tell() == start
                                ioloop(tid, path, page, out)
                                assert out.tell() == end + 1
                    except HTTPError as e:
                        error("failed to download %s, byte_range=%s" % (os.path.basename(path), str(se)))
            except IOError as e:
                with lock:
                    print('!', path, file=sys.stderr)
                    errors.setdefault(path, []).append(e)
            work.task_done()

    # detailed progress report
    def progress():
        with lock:
            left = sum(sizes.values())
            for path, flowrates in sorted(rates.items()):
                flows = {}
                for tid, (sz, t) in flowrates:
                    szs, ts = flows.get(tid, (0, 0))
                    flows[tid] = sz + szs, t + ts
                bps = sum(szs/ts for szs, ts in list(flows.values()) if ts > 0)
                info('%10s %8.1fMB/s %2dx  %s' % \
                    (megs(sizes[path]), bps / 1024.0**2, len(flows), "%s/%s" % (os.path.basename(os.path.split(path)[0]), os.path.split(path)[1])))
            if len(rates) != 0:  # only update if there's change
                info('%s remaining' % gigs(left))
            rates.clear()

    # process work items with a thread pool
    lock = threading.Lock()
    pool = []
    for i in range(HTTP_GAME_DOWNLOADER_THREADS):
        t = threading.Thread(target=worker)
        t.daemon = True
        t.start()
        pool.append(t)
    try:
        while any(t.is_alive() for t in pool):
            progress()
            time.sleep(1)
    except KeyboardInterrupt:
        raise
    except:
        with lock:
            log_exception('')
        raise


def cmd_backup(src_dir, dest_dir):
    gamesdb = load_manifest()

    info('finding all known files in the manifest')
    for game in sorted(gamesdb, key=lambda g: g.title):
        touched = False
        for itm in game.downloads + game.extras:
            if itm.name is None:
                continue

            src_game_dir = os.path.join(src_dir, game.title)
            src_file = os.path.join(src_game_dir, itm.name)
            dest_game_dir = os.path.join(dest_dir, game.title)
            dest_file = os.path.join(dest_game_dir, itm.name)

            if os.path.isfile(src_file):
                if itm.size != os.path.getsize(src_file):
                    warn('source file %s has unexpected size. skipping.' % src_file)
                    continue
                if not os.path.isdir(dest_game_dir):
                    os.makedirs(dest_game_dir)
                if not os.path.exists(dest_file) or itm.size != os.path.getsize(dest_file):
                    info('copying to %s...' % dest_file)
                    shutil.copy(src_file, dest_file)
                    touched = True

        # backup the info and serial files too
        if touched and os.path.isdir(dest_game_dir):
            for extra_file in [INFO_FILENAME, SERIAL_FILENAME]:
                if os.path.exists(os.path.join(src_game_dir, extra_file)):
                    shutil.copy(os.path.join(src_game_dir, extra_file), dest_game_dir)


def cmd_verify(gamedir, check_md5, check_filesize, check_zips, delete_on_fail, id):
    """Verifies all game files match manifest with any available md5 & file size info
    """
    item_count = 0
    missing_cnt = 0
    bad_md5_cnt = 0
    bad_size_cnt = 0
    bad_zip_cnt = 0
    del_file_cnt = 0

    items = load_manifest()

    # filter items based on id
    if id:
        games_to_check = []
        for game in sorted(items, key=lambda g: g.title):
            if game.title == id or str(game.id) == id:
                games_to_check.append(game)
        if len(games_to_check) == 0:
            warn('no known files with id "{}"'.format(id))
            return
        info('verifying known files with id "{}"'.format(id))
    else:
        info('verifying all known files in the manifest')
        games_to_check = sorted(items, key=lambda g: g.title)

    for game in games_to_check:
        for itm in game.downloads + game.extras:
            if itm.name is None:
                warn('no known filename for "%s (%s)"' % (game.title, itm.desc))
                continue

            item_count += 1

            itm_dirpath = os.path.join(game.title, itm.name)
            itm_file = os.path.join(gamedir, game.title, itm.name)

            if os.path.isfile(itm_file):
                info('verifying %s...' % itm_dirpath)

                fail = False
                if check_md5 and itm.md5 is not None:
                    if itm.md5 != hashfile(itm_file):
                        info('mismatched md5 for %s' % itm_dirpath)
                        bad_md5_cnt += 1
                        fail = True
                if check_filesize and itm.size is not None:
                    if itm.size != os.path.getsize(itm_file):
                        info('mismatched file size for %s' % itm_dirpath)
                        bad_size_cnt += 1
                        fail = True
                if check_zips and itm.name.lower().endswith('.zip'):
                    if not test_zipfile(itm_file):
                        info('zip test failed for %s' % itm_dirpath)
                        bad_zip_cnt += 1
                if delete_on_fail and fail:
                    info('deleting %s' % itm_dirpath)
                    os.remove(itm_file)
                    del_file_cnt += 1
            else:
                info('missing file %s' % itm_dirpath)
                missing_cnt += 1

    info('')
    info('--totals------------')
    info('known items......... %d' % item_count)
    info('have items.......... %d' % (item_count - missing_cnt - del_file_cnt))
    info('missing items....... %d' % (missing_cnt + del_file_cnt))
    if check_md5:
        info('md5 mismatches...... %d' % bad_md5_cnt)
    if check_filesize:
        info('size mismatches..... %d' % bad_size_cnt)
    if check_zips:
        info('zipfile failures.... %d' % bad_zip_cnt)
    if delete_on_fail:
        info('deleted items....... %d' % del_file_cnt)


def cmd_clean(cleandir, dryrun):
    items = load_manifest()
    items_by_title = {}
    total_size = 0  # in bytes
    have_cleaned = False

    # make convenient dict with title/dirname as key
    for item in items:
        items_by_title[item.title] = item

    # create orphan root dir
    orphan_root_dir = os.path.join(cleandir, ORPHAN_DIR_NAME)
    if not os.path.isdir(orphan_root_dir):
        if not dryrun:
            os.makedirs(orphan_root_dir)

    info("scanning local directories within '{}'...".format(cleandir))
    for cur_dir in sorted(os.listdir(cleandir)):
        cur_fulldir = os.path.join(cleandir, cur_dir)
        if os.path.isdir(cur_fulldir) and cur_dir not in ORPHAN_DIR_EXCLUDE_LIST:
            if cur_dir not in items_by_title:
                info("orphaning dir  '{}'".format(cur_dir))
                have_cleaned = True
                total_size += get_total_size(cur_fulldir)
                if not dryrun:
                    shutil.move(cur_fulldir, orphan_root_dir)
            else:
                # dir is valid game folder, check its files
                expected_filenames = []
                for game_item in items_by_title[cur_dir].downloads + items_by_title[cur_dir].extras:
                    expected_filenames.append(game_item.name)
                for cur_dir_file in os.listdir(cur_fulldir):
                    if os.path.isdir(os.path.join(cleandir, cur_dir, cur_dir_file)):
                        continue  # leave subdirs alone
                    if cur_dir_file not in expected_filenames and cur_dir_file not in ORPHAN_FILE_EXCLUDE_LIST:
                        info("orphaning file '{}'".format(os.path.join(cur_dir, cur_dir_file)))
                        have_cleaned = True
                        dest_dir = os.path.join(orphan_root_dir, cur_dir)
                        if not os.path.isdir(dest_dir):
                            if not dryrun:
                                os.makedirs(dest_dir)
                        file_to_move = os.path.join(cleandir, cur_dir, cur_dir_file)
                        total_size += os.path.getsize(file_to_move)
                        if not dryrun:
                            shutil.move(file_to_move, dest_dir)

    if have_cleaned:
        info('')
        info('total size of newly orphaned files: {}'.format(pretty_size(total_size)))
        if not dryrun:
            info('orphaned items moved to: {}'.format(orphan_root_dir))
    else:
        info('nothing to clean. nice and tidy!')


def main(args):
    stime = datetime.datetime.now()

    if args.cmd == 'login':
        cmd_login(args.username, args.password)
        return  # no need to see time stats
    elif args.cmd == 'update':
        cmd_update(args.os, args.lang, args.skipknown, args.updateonly, args.id)
    elif args.cmd == 'download':
        if args.wait > 0.0:
            info('sleeping for %.2fhr...' % args.wait)
            time.sleep(args.wait * 60 * 60)
        cmd_download(args.savedir, args.skipextras, args.skipgames, args.skipids, args.dryrun, args.id)
    elif args.cmd == 'import':
        cmd_import(args.src_dir, args.dest_dir)
    elif args.cmd == 'verify':
        check_md5 = not args.skipmd5
        check_filesize = not args.skipsize
        check_zips = not args.skipzip
        cmd_verify(args.gamedir, check_md5, check_filesize, check_zips, args.delete, args.id)
    elif args.cmd == 'backup':
        cmd_backup(args.src_dir, args.dest_dir)
    elif args.cmd == 'clean':
        cmd_clean(args.cleandir, args.dryrun)

    etime = datetime.datetime.now()
    info('--')
    info('total time: %s' % (etime - stime))


if __name__ == "__main__":
    try:
        main(process_argv(sys.argv))
        info('exiting...')
    except KeyboardInterrupt:
        info('exiting...')
        sys.exit(1)
    except SystemExit:
        raise
    except:
        log_exception('fatal...')
        sys.exit(1)