main.py

from formatter import Formatter
from html_parser import Parser
import sys
from distutils import dir_util
import os

def main(argv):
	for arg in argv:
		main_one_arg(arg)

def main_one_arg(directory):
	if not os.path.exists(directory):
		raise OSError("Directory " + str(directory) + " does not exist.")
	destination = directory + "-protoxml"
	if not os.path.exists(destination):
		os.makedirs(destination)

	for filename in os.listdir(directory):
		parse_and_format(directory, filename, destination)

def parse_and_format(dir_path, filename, dest, level=1):
	assert(os.path.exists(os.path.join(dir_path, filename)))
	path_to_file = os.path.join(dir_path, filename)
	path_to_dst = os.path.join(dest, filename)

	if os.path.isdir(path_to_file):
		if filename == "images":
			dir_util.copy_tree(path_to_file, path_to_dst)
		else:
			if not os.path.exists(path_to_dst):
				os.makedirs(path_to_dst)
			for sub_file in os.listdir(path_to_file):
				parse_and_format(path_to_file, sub_file, path_to_dst, level+1)
	elif os.path.isfile(path_to_file):
		parse_format_and_write_file(path_to_file, os.path.join(dest,
			filename.replace(".html", ".protoxml")), level)
	else:
		print "Unhandled file type (probably symlink) with file " + str(path_to_file)

def parse_format_and_write_file(filename, dest, level):
	has_been_parsed = False
	parser = None
	formatter = None
	try:
		parser = Parser(filename)
		parser.parse()
		has_been_parsed = True
	except Exception as e:
		print "Error with file while parsing: " + str(filename) + " (this file might not be HTML valid enough)"
		print "Error message: " + str(e)
	if has_been_parsed:
		try:
			formatter = Formatter(parser.get_data(), level)
			formatter.format()
			formatter.write_to_file(dest)
		except Exception as e:
			print "Error with file while fomatting: " + str(filename)


if __name__=="__main__":
	main(sys.argv[1:]) # don't get the first arg (it's the command name)