wgets

#!/usr/bin/env python2
# -*- coding: utf-8 -*-
from __future__ import print_function

import itertools as it, operator as op, functools as ft
import os, sys, types, urlparse, urllib, subprocess, hashlib


def print_usage(extras=None, exit_with=1):
	p = ft.partial(print, file=sys.stderr)
	cmd = os.path.basename(sys.argv[0])
	example_args = 'http://os.archlinuxarm.org/os/'\
		'ArchLinuxARM-sun4i-latest.tar.gz', 'cea5d785df19151806aa5ac3a917e41c'
	p('Usage: {} [ wget_opts ] url checksum'.format(cmd))
	p('Example: {} -c {} {}'.format(cmd, *example_args))
	if extras is not None:
		if isinstance(extras, types.StringTypes): extras = [extras]
		for line in extras: p(line)
	if exit_with is not None: sys.exit(exit_with)


class InputError(Exception):

	def print_usage(self):
		if not self.args: print_usage()
		msg = self.args[0].format(*self.args[1:])
		print_usage('ERROR: {}'.format(msg))


def process_cli(args=None):
	if args is None: args = sys.argv[1:]
	args = args or ['-h']

	for arg in args:
		if arg.startswith('-h') or arg == '--help': raise InputError

	args, checksum = args[:-1], args[-1]
	if not args: raise InputError('both url and checksum must be specified')
	args, url = args[:-1], args[-1]

	file_path = urlparse.urlparse(url).path
	file_path = urllib.unquote_plus(file_path.rsplit('/', 1)[-1])
	if not file_path: raise InputError('failed to detect output filename from url')
	args = ['wget'] + args + ['-O', file_path, url]

	return file_path, checksum, args


def validate_digest( checksum,
		_hash_funcs = [ ('md5', 16), ('sha1', 20),
			('sha256', 32), ('sha512', 64), ('sha224', 28), ('sha384', 48) ] ):
	digest = checksum.lower().decode('hex')

	for hash_name, hash_len in _hash_funcs:
		if len(digest) == hash_len: break
	else:
		raise InputError( 'Error: failed to pick hash'
			' for value (len: {}): {}'.format(len(checksum), checksum) )
	hash_func = getattr(hashlib, hash_name)

	return digest, hash_name, hash_func


class DigestError(Exception): pass

def check_hash(file_path, hash_func, digest, bs=4 * 2**20):
	chk = hash_func()
	with open(file_path) as src:
		for chunk in iter(ft.partial(src.read, bs), ''): chk.update(chunk)
	chk = chk.digest()
	if digest != chk:
		raise DigestError('{} (expected) != {} (real)'.format(digest.encode('hex'), chk.encode('hex')))


def main(args=None):
	try:
		file_path, checksum, wget_args = process_cli(args)
		digest, hash_name, hash_func = validate_digest(checksum)
	except InputError as err: err.print_usage()

	print('Using hash: {}'.format(hash_name))
	print('Using output filename: {}'.format(file_path))

	subprocess.check_call(wget_args)
	check_hash(file_path, hash_func, digest)

	print('Checksum matched')


if __name__ == '__main__': sys.exit(main())