xmarks_recover.py

# Download the xmarks "View Bookmarks Revision" from their webpage for all revisions, then run this to audit changes.

import glob, re

class BookmarkSeparator(object):
	name = '<<SEPARATOR>>'
	add_date = 0
	last_modified = 0
	description = '<<SEPARATOR>>'

	def printout(self, pathprefix):
		print('{}/<<SEPARATOR>>'.format(pathprefix))

	def iter_path(self):
		return iter('')

class Bookmark(object):
	def __init__(self, href, add_date, last_modified, name, shortcut_url):
		self.href = href
		self.add_date = add_date
		self.last_modified = last_modified
		self.name = name
		self.description = ''
		self.shortcut_url = shortcut_url

	def printout(self, pathprefix):
		print('{}/{}'.format(pathprefix, self.name))

	def iter_path(self):
		yield (self.name, self)

class BookmarkFolder(object):
	def __init__(self, depth, parseiter, add_date, last_modified, name, personal_toolbar_folder):
		self.add_date = int((add_date if add_date else '0'),10)
		self.last_modified = int((last_modified if last_modified else '0'),10)
		self.name = name
		self.description = ''
		self.personal_toolbar_folder = personal_toolbar_folder
		if self.personal_toolbar_folder:
			self.name = '<<TOOLBAR>>'

		self.contents = []

		if parseiter is None:
			return

		debug_previous_line = (None, None)
		prev_entity = self
		prev_is_description = False
		for line in parseiter:
			line = line.rstrip('\n')
			debug_previous_line = (line, debug_previous_line[0])

			try:
				if line.startswith('<DD>'):
					prev_entity.description = line[4:]
					prev_is_description = True
					continue
				elif prev_is_description and not line.lstrip(' ').startswith('<'):
					prev_entity.description += '\n' + line
					continue
				else:
					prev_is_description = False

				if not line.strip(' '):
					continue

				if line == '    '*(depth-1) + '<DL><p>':
					continue # our start tags

				if not line.startswith('    '*depth + '<'):
					assert line == '    '*(depth-1) + '</DL><p>' # our end tags
					return

				line = line.lstrip(' ')
				if line.startswith('<DT><H3 '):
					m = re.match('^<DT><H3 ADD_DATE="([0-9]+)" LAST_MODIFIED="([0-9]*)"( PERSONAL_TOOLBAR_FOLDER="true")?>(.+)</H3>$', line)
					assert m is not None
					prev_entity = BookmarkFolder(depth+1, parseiter, add_date=m.group(1), last_modified=m.group(2), name=m.group(4), personal_toolbar_folder=(m.group(3) is not None))
					self.contents.append(prev_entity)
					continue

				elif line == '<HR>':
					prev_entity = BookmarkSeparator()
					self.contents.append(prev_entity)

				elif line.startswith('<DT><A HREF='):
					m = re.match('^<DT><A HREF="([^"]+)" ADD_DATE="([0-9]*)" LAST_MODIFIED="([0-9]*)"(?: SHORTCUTURL="([^"]+)")?>(.+)</A>$', line)
					assert m is not None
					prev_entity = Bookmark(href=m.group(1), add_date=m.group(2), last_modified=m.group(3), name=m.group(5), shortcut_url=m.group(4))
					self.contents.append(prev_entity)
					continue

				else:
					raise RuntimeError('Unable to parse line: {}'.format(line))
			except:
				print('Prev: {}'.format(debug_previous_line[1]))
				print('Line: {}'.format(line))
				raise

	def printout(self, pathprefix=None):
		name = '{}/{}'.format(pathprefix, self.name)
		if pathprefix is None:
			name = self.name
		print('{}/'.format(name))
		for f in self.contents:
			f.printout(name)

	def iter_path(self):
		for child in self.contents:
			for path, obj in child.iter_path():
				yield ('{}/{}'.format(self.name, path), obj)


def parse_data(htmlfile):
	with open(htmlfile,'r') as f:
		parseiter = iter(f)
		try:
			while parseiter.next().rstrip('\n'):
				pass
		except StopIteration:
			return BookmarkFolder(1, None, add_date='', last_modified='', name='', personal_toolbar_folder=False)
		else:
			return BookmarkFolder(1, parseiter, add_date='', last_modified='', name='', personal_toolbar_folder=False)

revisions = []
for htmlfile in sorted(glob.glob('*.html')):
	try:
		revisions.append((int(htmlfile.replace('.html',''),10), parse_data(htmlfile)))
	except:
		print htmlfile
		raise

recent = (None, None)
for revid, rev in revisions:
	paths = dict(rev.iter_path())
	recent = (paths, recent[0])
	if recent[1] is None:
		continue

	rev_changes = []

	old_hrefs = set(map(lambda x: x.href, recent[1].values()))
	new_hrefs = set(map(lambda x: x.href, recent[0].values()))

	all_paths = set()
	all_paths.update(recent[0].keys())
	all_paths.update(recent[1].keys())
	for path in sorted(all_paths):
		if path in recent[1] and recent[1][path].href in new_hrefs:
			continue # Not a deletion or insertion even if inconsistent
		elif path in recent[0] and recent[0][path].href in old_hrefs:
			continue # Not a deletion or insertion even if inconsistent

		if path not in recent[0]:
			rev_changes.append('Removed:  {}'.format(path))
		if path not in recent[1]:
			rev_changes.append('Added:    {}'.format(path))
		if path in recent[1] and path in recent[0] and recent[1][path].href != recent[0][path].href:
			rev_changes.append('Edited:   {}'.format(path))

	if rev_changes:
		print('*** REVISION #{} ***'.format(revid))
		print('')
		print('\n'.join(rev_changes))
		print('')