test/test_utils.py

   1 #!/usr/bin/env python
   2
   3 # Various small unit tests
   4
   5 import sys
   6 import unittest
   7
   8 # Allow direct execution
   9 import os
  10 sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
  11
  12 #from youtube_dl.utils import htmlentity_transform
  13 from youtube_dl.utils import timeconvert
  14 from youtube_dl.utils import sanitize_filename
  15 from youtube_dl.utils import unescapeHTML
  16 from youtube_dl.utils import orderedSet
  17
  18 if sys.version_info < (3, 0):
  19     _compat_str = lambda b: b.decode('unicode-escape')
  20 else:
  21     _compat_str = lambda s: s
  22
  23
  24 class TestUtil(unittest.TestCase):
  25     def test_timeconvert(self):
  26         self.assertTrue(timeconvert('') is None)
  27         self.assertTrue(timeconvert('bougrg') is None)
  28
  29     def test_sanitize_filename(self):
  30         self.assertEqual(sanitize_filename('abc'), 'abc')
  31         self.assertEqual(sanitize_filename('abc_d-e'), 'abc_d-e')
  32
  33         self.assertEqual(sanitize_filename('123'), '123')
  34
  35         self.assertEqual('abc_de', sanitize_filename('abc/de'))
  36         self.assertFalse('/' in sanitize_filename('abc/de///'))
  37
  38         self.assertEqual('abc_de', sanitize_filename('abc/<>\\*|de'))
  39         self.assertEqual('xxx', sanitize_filename('xxx/<>\\*|'))
  40         self.assertEqual('yes no', sanitize_filename('yes? no'))
  41         self.assertEqual('this - that', sanitize_filename('this: that'))
  42
  43         self.assertEqual(sanitize_filename('AT&T'), 'AT&T')
  44         aumlaut = _compat_str('\xe4')
  45         self.assertEqual(sanitize_filename(aumlaut), aumlaut)
  46         tests = _compat_str('\u043a\u0438\u0440\u0438\u043b\u043b\u0438\u0446\u0430')
  47         self.assertEqual(sanitize_filename(tests), tests)
  48
  49         forbidden = '"\0\\/'
  50         for fc in forbidden:
  51             for fbc in forbidden:
  52                 self.assertTrue(fbc not in sanitize_filename(fc))
  53
  54     def test_sanitize_filename_restricted(self):
  55         self.assertEqual(sanitize_filename('abc', restricted=True), 'abc')
  56         self.assertEqual(sanitize_filename('abc_d-e', restricted=True), 'abc_d-e')
  57
  58         self.assertEqual(sanitize_filename('123', restricted=True), '123')
  59
  60         self.assertEqual('abc_de', sanitize_filename('abc/de', restricted=True))
  61         self.assertFalse('/' in sanitize_filename('abc/de///', restricted=True))
  62
  63         self.assertEqual('abc_de', sanitize_filename('abc/<>\\*|de', restricted=True))
  64         self.assertEqual('xxx', sanitize_filename('xxx/<>\\*|', restricted=True))
  65         self.assertEqual('yes_no', sanitize_filename('yes? no', restricted=True))
  66         self.assertEqual('this_-_that', sanitize_filename('this: that', restricted=True))
  67
  68         tests = _compat_str('a\xe4b\u4e2d\u56fd\u7684c')
  69         self.assertEqual(sanitize_filename(tests, restricted=True), 'a_b_c')
  70         self.assertTrue(sanitize_filename(_compat_str('\xf6'), restricted=True) != '')  # No empty filename
  71
  72         forbidden = '"\0\\/&!: \'\t\n()[]{}$;`^,#'
  73         for fc in forbidden:
  74             for fbc in forbidden:
  75                 self.assertTrue(fbc not in sanitize_filename(fc, restricted=True))
  76
  77         # Handle a common case more neatly
  78         self.assertEqual(sanitize_filename(_compat_str('\u5927\u58f0\u5e26 - Song'), restricted=True), 'Song')
  79         self.assertEqual(sanitize_filename(_compat_str('\u603b\u7edf: Speech'), restricted=True), 'Speech')
  80         # .. but make sure the file name is never empty
  81         self.assertTrue(sanitize_filename('-', restricted=True) != '')
  82         self.assertTrue(sanitize_filename(':', restricted=True) != '')
  83
  84     def test_sanitize_ids(self):
  85         self.assertEqual(sanitize_filename('_n_cd26wFpw', is_id=True), '_n_cd26wFpw')
  86         self.assertEqual(sanitize_filename('_BD_eEpuzXw', is_id=True), '_BD_eEpuzXw')
  87         self.assertEqual(sanitize_filename('N0Y__7-UOdI', is_id=True), 'N0Y__7-UOdI')
  88
  89     def test_ordered_set(self):
  90         self.assertEqual(orderedSet([1, 1, 2, 3, 4, 4, 5, 6, 7, 3, 5]), [1, 2, 3, 4, 5, 6, 7])
  91         self.assertEqual(orderedSet([]), [])
  92         self.assertEqual(orderedSet([1]), [1])
  93         #keep the list ordered
  94         self.assertEqual(orderedSet([135, 1, 1, 1]), [135, 1])
  95
  96     def test_unescape_html(self):
  97         self.assertEqual(unescapeHTML(_compat_str('%20;')), _compat_str('%20;'))
  98
  99 if __name__ == '__main__':
 100     unittest.main()