| OLD | NEW |
| (Empty) |
| 1 #!/usr/bin/env python | |
| 2 # Copyright (c) 2012 The Chromium Authors. All rights reserved. | |
| 3 # Use of this source code is governed by a BSD-style license that can be | |
| 4 # found in the LICENSE file. | |
| 5 | |
| 6 '''Pseudotranslation support. Our pseudotranslations are based on the | |
| 7 P-language, which is a simple vowel-extending language. Examples of P: | |
| 8 - "hello" becomes "hepellopo" | |
| 9 - "howdie" becomes "hopowdiepie" | |
| 10 - "because" becomes "bepecaupause" (but in our implementation we don't | |
| 11 handle the silent e at the end so it actually would return "bepecaupausepe" | |
| 12 | |
| 13 The P-language has the excellent quality of increasing the length of text | |
| 14 by around 30-50% which is great for pseudotranslations, to stress test any | |
| 15 GUI layouts etc. | |
| 16 | |
| 17 To make the pseudotranslations more obviously "not a translation" and to make | |
| 18 them exercise any code that deals with encodings, we also transform all English | |
| 19 vowels into equivalent vowels with diacriticals on them (rings, acutes, | |
| 20 diaresis, and circumflex), and we write the "p" in the P-language as a Hebrew | |
| 21 character Qof. It looks sort of like a latin character "p" but it is outside | |
| 22 the latin-1 character set which will stress character encoding bugs. | |
| 23 ''' | |
| 24 | |
| 25 from grit import lazy_re | |
| 26 from grit import tclib | |
| 27 | |
| 28 | |
| 29 # An RFC language code for the P pseudolanguage. | |
| 30 PSEUDO_LANG = 'x-P-pseudo' | |
| 31 | |
| 32 # Hebrew character Qof. It looks kind of like a 'p' but is outside | |
| 33 # the latin-1 character set which is good for our purposes. | |
| 34 # TODO(joi) For now using P instead of Qof, because of some bugs it used. Find | |
| 35 # a better solution, i.e. one that introduces a non-latin1 character into the | |
| 36 # pseudotranslation. | |
| 37 #_QOF = u'\u05e7' | |
| 38 _QOF = u'P' | |
| 39 | |
| 40 # How we map each vowel. | |
| 41 _VOWELS = { | |
| 42 u'a' : u'\u00e5', # a with ring | |
| 43 u'e' : u'\u00e9', # e acute | |
| 44 u'i' : u'\u00ef', # i diaresis | |
| 45 u'o' : u'\u00f4', # o circumflex | |
| 46 u'u' : u'\u00fc', # u diaresis | |
| 47 u'y' : u'\u00fd', # y acute | |
| 48 u'A' : u'\u00c5', # A with ring | |
| 49 u'E' : u'\u00c9', # E acute | |
| 50 u'I' : u'\u00cf', # I diaresis | |
| 51 u'O' : u'\u00d4', # O circumflex | |
| 52 u'U' : u'\u00dc', # U diaresis | |
| 53 u'Y' : u'\u00dd', # Y acute | |
| 54 } | |
| 55 | |
| 56 # Matches vowels and P | |
| 57 _PSUB_RE = lazy_re.compile("(%s)" % '|'.join(_VOWELS.keys() + ['P'])) | |
| 58 | |
| 59 | |
| 60 # Pseudotranslations previously created. This is important for performance | |
| 61 # reasons, especially since we routinely pseudotranslate the whole project | |
| 62 # several or many different times for each build. | |
| 63 _existing_translations = {} | |
| 64 | |
| 65 | |
| 66 def MapVowels(str, also_p = False): | |
| 67 '''Returns a copy of 'str' where characters that exist as keys in _VOWELS | |
| 68 have been replaced with the corresponding value. If also_p is true, this | |
| 69 function will also change capital P characters into a Hebrew character Qof. | |
| 70 ''' | |
| 71 def Repl(match): | |
| 72 if match.group() == 'p': | |
| 73 if also_p: | |
| 74 return _QOF | |
| 75 else: | |
| 76 return 'p' | |
| 77 else: | |
| 78 return _VOWELS[match.group()] | |
| 79 return _PSUB_RE.sub(Repl, str) | |
| 80 | |
| 81 | |
| 82 def PseudoString(str): | |
| 83 '''Returns a pseudotranslation of the provided string, in our enhanced | |
| 84 P-language.''' | |
| 85 if str in _existing_translations: | |
| 86 return _existing_translations[str] | |
| 87 | |
| 88 outstr = u'' | |
| 89 ix = 0 | |
| 90 while ix < len(str): | |
| 91 if str[ix] not in _VOWELS.keys(): | |
| 92 outstr += str[ix] | |
| 93 ix += 1 | |
| 94 else: | |
| 95 # We want to treat consecutive vowels as one composite vowel. This is not | |
| 96 # always accurate e.g. in composite words but good enough. | |
| 97 consecutive_vowels = u'' | |
| 98 while ix < len(str) and str[ix] in _VOWELS.keys(): | |
| 99 consecutive_vowels += str[ix] | |
| 100 ix += 1 | |
| 101 changed_vowels = MapVowels(consecutive_vowels) | |
| 102 outstr += changed_vowels | |
| 103 outstr += _QOF | |
| 104 outstr += changed_vowels | |
| 105 | |
| 106 _existing_translations[str] = outstr | |
| 107 return outstr | |
| 108 | |
| 109 | |
| 110 def PseudoMessage(message): | |
| 111 '''Returns a pseudotranslation of the provided message. | |
| 112 | |
| 113 Args: | |
| 114 message: tclib.Message() | |
| 115 | |
| 116 Return: | |
| 117 tclib.Translation() | |
| 118 ''' | |
| 119 transl = tclib.Translation() | |
| 120 | |
| 121 for part in message.GetContent(): | |
| 122 if isinstance(part, tclib.Placeholder): | |
| 123 transl.AppendPlaceholder(part) | |
| 124 else: | |
| 125 transl.AppendText(PseudoString(part)) | |
| 126 | |
| 127 return transl | |
| 128 | |
| OLD | NEW |