Lib/json/tests/test_unicode.py

   1 from unittest import TestCase
   2
   3 import json
   4 from collections import OrderedDict
   5
   6 class TestUnicode(TestCase):
   7     def test_encoding1(self):
   8         encoder = json.JSONEncoder(encoding='utf-8')
   9         u = u'\N{GREEK SMALL LETTER ALPHA}\N{GREEK CAPITAL LETTER OMEGA}'
  10         s = u.encode('utf-8')
  11         ju = encoder.encode(u)
  12         js = encoder.encode(s)
  13         self.assertEquals(ju, js)
  14
  15     def test_encoding2(self):
  16         u = u'\N{GREEK SMALL LETTER ALPHA}\N{GREEK CAPITAL LETTER OMEGA}'
  17         s = u.encode('utf-8')
  18         ju = json.dumps(u, encoding='utf-8')
  19         js = json.dumps(s, encoding='utf-8')
  20         self.assertEquals(ju, js)
  21
  22     def test_encoding3(self):
  23         u = u'\N{GREEK SMALL LETTER ALPHA}\N{GREEK CAPITAL LETTER OMEGA}'
  24         j = json.dumps(u)
  25         self.assertEquals(j, '"\\u03b1\\u03a9"')
  26
  27     def test_encoding4(self):
  28         u = u'\N{GREEK SMALL LETTER ALPHA}\N{GREEK CAPITAL LETTER OMEGA}'
  29         j = json.dumps([u])
  30         self.assertEquals(j, '["\\u03b1\\u03a9"]')
  31
  32     def test_encoding5(self):
  33         u = u'\N{GREEK SMALL LETTER ALPHA}\N{GREEK CAPITAL LETTER OMEGA}'
  34         j = json.dumps(u, ensure_ascii=False)
  35         self.assertEquals(j, u'"{0}"'.format(u))
  36
  37     def test_encoding6(self):
  38         u = u'\N{GREEK SMALL LETTER ALPHA}\N{GREEK CAPITAL LETTER OMEGA}'
  39         j = json.dumps([u], ensure_ascii=False)
  40         self.assertEquals(j, u'["{0}"]'.format(u))
  41
  42     def test_big_unicode_encode(self):
  43         u = u'\U0001d120'
  44         self.assertEquals(json.dumps(u), '"\\ud834\\udd20"')
  45         self.assertEquals(json.dumps(u, ensure_ascii=False), u'"\U0001d120"')
  46
  47     def test_big_unicode_decode(self):
  48         u = u'z\U0001d120x'
  49         self.assertEquals(json.loads('"' + u + '"'), u)
  50         self.assertEquals(json.loads('"z\\ud834\\udd20x"'), u)
  51
  52     def test_unicode_decode(self):
  53         for i in range(0, 0xd7ff):
  54             u = unichr(i)
  55             s = '"\\u{0:04x}"'.format(i)
  56             self.assertEquals(json.loads(s), u)
  57
  58     def test_object_pairs_hook_with_unicode(self):
  59         s = u'{"xkd":1, "kcw":2, "art":3, "hxm":4, "qrt":5, "pad":6, "hoy":7}'
  60         p = [(u"xkd", 1), (u"kcw", 2), (u"art", 3), (u"hxm", 4),
  61              (u"qrt", 5), (u"pad", 6), (u"hoy", 7)]
  62         self.assertEqual(json.loads(s), eval(s))
  63         self.assertEqual(json.loads(s, object_pairs_hook = lambda x: x), p)
  64         od = json.loads(s, object_pairs_hook = OrderedDict)
  65         self.assertEqual(od, OrderedDict(p))
  66         self.assertEqual(type(od), OrderedDict)
  67         # the object_pairs_hook takes priority over the object_hook
  68         self.assertEqual(json.loads(s,
  69                                     object_pairs_hook = OrderedDict,
  70                                     object_hook = lambda x: None),
  71                          OrderedDict(p))
  72
  73     def test_default_encoding(self):
  74         self.assertEquals(json.loads(u'{"a": "\xe9"}'.encode('utf-8')),
  75             {'a': u'\xe9'})
  76
  77     def test_unicode_preservation(self):
  78         self.assertEquals(type(json.loads(u'""')), unicode)
  79         self.assertEquals(type(json.loads(u'"a"')), unicode)
  80         self.assertEquals(type(json.loads(u'["a"]')[0]), unicode)