Commit | Line | Data |
---|---|---|
86530b38 AT |
1 | """ Python Character Mapping Codec generated from '8859-10.TXT' with gencodec.py. |
2 | ||
3 | Written by Marc-Andre Lemburg (mal@lemburg.com). | |
4 | ||
5 | (c) Copyright CNRI, All Rights Reserved. NO WARRANTY. | |
6 | (c) Copyright 2000 Guido van Rossum. | |
7 | ||
8 | """#" | |
9 | ||
10 | import codecs | |
11 | ||
12 | ### Codec APIs | |
13 | ||
14 | class Codec(codecs.Codec): | |
15 | ||
16 | def encode(self,input,errors='strict'): | |
17 | ||
18 | return codecs.charmap_encode(input,errors,encoding_map) | |
19 | ||
20 | def decode(self,input,errors='strict'): | |
21 | ||
22 | return codecs.charmap_decode(input,errors,decoding_map) | |
23 | ||
24 | class StreamWriter(Codec,codecs.StreamWriter): | |
25 | pass | |
26 | ||
27 | class StreamReader(Codec,codecs.StreamReader): | |
28 | pass | |
29 | ||
30 | ### encodings module API | |
31 | ||
32 | def getregentry(): | |
33 | ||
34 | return (Codec().encode,Codec().decode,StreamReader,StreamWriter) | |
35 | ||
36 | ### Decoding Map | |
37 | ||
38 | decoding_map = codecs.make_identity_dict(range(256)) | |
39 | decoding_map.update({ | |
40 | 0x00a1: 0x0104, # LATIN CAPITAL LETTER A WITH OGONEK | |
41 | 0x00a2: 0x0112, # LATIN CAPITAL LETTER E WITH MACRON | |
42 | 0x00a3: 0x0122, # LATIN CAPITAL LETTER G WITH CEDILLA | |
43 | 0x00a4: 0x012a, # LATIN CAPITAL LETTER I WITH MACRON | |
44 | 0x00a5: 0x0128, # LATIN CAPITAL LETTER I WITH TILDE | |
45 | 0x00a6: 0x0136, # LATIN CAPITAL LETTER K WITH CEDILLA | |
46 | 0x00a8: 0x013b, # LATIN CAPITAL LETTER L WITH CEDILLA | |
47 | 0x00a9: 0x0110, # LATIN CAPITAL LETTER D WITH STROKE | |
48 | 0x00aa: 0x0160, # LATIN CAPITAL LETTER S WITH CARON | |
49 | 0x00ab: 0x0166, # LATIN CAPITAL LETTER T WITH STROKE | |
50 | 0x00ac: 0x017d, # LATIN CAPITAL LETTER Z WITH CARON | |
51 | 0x00ae: 0x016a, # LATIN CAPITAL LETTER U WITH MACRON | |
52 | 0x00af: 0x014a, # LATIN CAPITAL LETTER ENG | |
53 | 0x00b1: 0x0105, # LATIN SMALL LETTER A WITH OGONEK | |
54 | 0x00b2: 0x0113, # LATIN SMALL LETTER E WITH MACRON | |
55 | 0x00b3: 0x0123, # LATIN SMALL LETTER G WITH CEDILLA | |
56 | 0x00b4: 0x012b, # LATIN SMALL LETTER I WITH MACRON | |
57 | 0x00b5: 0x0129, # LATIN SMALL LETTER I WITH TILDE | |
58 | 0x00b6: 0x0137, # LATIN SMALL LETTER K WITH CEDILLA | |
59 | 0x00b8: 0x013c, # LATIN SMALL LETTER L WITH CEDILLA | |
60 | 0x00b9: 0x0111, # LATIN SMALL LETTER D WITH STROKE | |
61 | 0x00ba: 0x0161, # LATIN SMALL LETTER S WITH CARON | |
62 | 0x00bb: 0x0167, # LATIN SMALL LETTER T WITH STROKE | |
63 | 0x00bc: 0x017e, # LATIN SMALL LETTER Z WITH CARON | |
64 | 0x00bd: 0x2015, # HORIZONTAL BAR | |
65 | 0x00be: 0x016b, # LATIN SMALL LETTER U WITH MACRON | |
66 | 0x00bf: 0x014b, # LATIN SMALL LETTER ENG | |
67 | 0x00c0: 0x0100, # LATIN CAPITAL LETTER A WITH MACRON | |
68 | 0x00c7: 0x012e, # LATIN CAPITAL LETTER I WITH OGONEK | |
69 | 0x00c8: 0x010c, # LATIN CAPITAL LETTER C WITH CARON | |
70 | 0x00ca: 0x0118, # LATIN CAPITAL LETTER E WITH OGONEK | |
71 | 0x00cc: 0x0116, # LATIN CAPITAL LETTER E WITH DOT ABOVE | |
72 | 0x00d1: 0x0145, # LATIN CAPITAL LETTER N WITH CEDILLA | |
73 | 0x00d2: 0x014c, # LATIN CAPITAL LETTER O WITH MACRON | |
74 | 0x00d7: 0x0168, # LATIN CAPITAL LETTER U WITH TILDE | |
75 | 0x00d9: 0x0172, # LATIN CAPITAL LETTER U WITH OGONEK | |
76 | 0x00e0: 0x0101, # LATIN SMALL LETTER A WITH MACRON | |
77 | 0x00e7: 0x012f, # LATIN SMALL LETTER I WITH OGONEK | |
78 | 0x00e8: 0x010d, # LATIN SMALL LETTER C WITH CARON | |
79 | 0x00ea: 0x0119, # LATIN SMALL LETTER E WITH OGONEK | |
80 | 0x00ec: 0x0117, # LATIN SMALL LETTER E WITH DOT ABOVE | |
81 | 0x00f1: 0x0146, # LATIN SMALL LETTER N WITH CEDILLA | |
82 | 0x00f2: 0x014d, # LATIN SMALL LETTER O WITH MACRON | |
83 | 0x00f7: 0x0169, # LATIN SMALL LETTER U WITH TILDE | |
84 | 0x00f9: 0x0173, # LATIN SMALL LETTER U WITH OGONEK | |
85 | 0x00ff: 0x0138, # LATIN SMALL LETTER KRA | |
86 | }) | |
87 | ||
88 | ### Encoding Map | |
89 | ||
90 | encoding_map = codecs.make_encoding_map(decoding_map) |