308 lines
13 KiB
Python
308 lines
13 KiB
Python
|
""" Python Character Mapping Codec iso8859_3 generated from 'MAPPINGS/ISO8859/8859-3.TXT' with gencodec.py.
|
||
|
|
|
||
|
|
"""#"
|
||
|
|
|
||
|
|
import codecs
|
||
|
|
|
||
|
|
### Codec APIs
|
||
|
|
|
||
|
|
class Codec(codecs.Codec):
|
||
|
|
|
||
|
|
def encode(self,input,errors='strict'):
|
||
|
|
return codecs.charmap_encode(input,errors,encoding_table)
|
||
|
|
|
||
|
|
def decode(self,input,errors='strict'):
|
||
|
|
return codecs.charmap_decode(input,errors,decoding_table)
|
||
|
|
|
||
|
|
class IncrementalEncoder(codecs.IncrementalEncoder):
|
||
|
|
def encode(self, input, final=False):
|
||
|
|
return codecs.charmap_encode(input,self.errors,encoding_table)[0]
|
||
|
|
|
||
|
|
class IncrementalDecoder(codecs.IncrementalDecoder):
|
||
|
|
def decode(self, input, final=False):
|
||
|
|
return codecs.charmap_decode(input,self.errors,decoding_table)[0]
|
||
|
|
|
||
|
|
class StreamWriter(Codec,codecs.StreamWriter):
|
||
|
|
pass
|
||
|
|
|
||
|
|
class StreamReader(Codec,codecs.StreamReader):
|
||
|
|
pass
|
||
|
|
|
||
|
|
### encodings module API
|
||
|
|
|
||
|
|
def getregentry():
|
||
|
|
return codecs.CodecInfo(
|
||
|
|
name='iso8859-3',
|
||
|
|
encode=Codec().encode,
|
||
|
|
decode=Codec().decode,
|
||
|
|
incrementalencoder=IncrementalEncoder,
|
||
|
|
incrementaldecoder=IncrementalDecoder,
|
||
|
|
streamreader=StreamReader,
|
||
|
|
streamwriter=StreamWriter,
|
||
|
|
)
|
||
|
|
|
||
|
|
|
||
|
|
### Decoding Table
|
||
|
|
|
||
|
|
decoding_table = (
|
||
|
|
u'\x00' # 0x00 -> NULL
|
||
|
|
u'\x01' # 0x01 -> START OF HEADING
|
||
|
|
u'\x02' # 0x02 -> START OF TEXT
|
||
|
|
u'\x03' # 0x03 -> END OF TEXT
|
||
|
|
u'\x04' # 0x04 -> END OF TRANSMISSION
|
||
|
|
u'\x05' # 0x05 -> ENQUIRY
|
||
|
|
u'\x06' # 0x06 -> ACKNOWLEDGE
|
||
|
|
u'\x07' # 0x07 -> BELL
|
||
|
|
u'\x08' # 0x08 -> BACKSPACE
|
||
|
|
u'\t' # 0x09 -> HORIZONTAL TABULATION
|
||
|
|
u'\n' # 0x0A -> LINE FEED
|
||
|
|
u'\x0b' # 0x0B -> VERTICAL TABULATION
|
||
|
|
u'\x0c' # 0x0C -> FORM FEED
|
||
|
|
u'\r' # 0x0D -> CARRIAGE RETURN
|
||
|
|
u'\x0e' # 0x0E -> SHIFT OUT
|
||
|
|
u'\x0f' # 0x0F -> SHIFT IN
|
||
|
|
u'\x10' # 0x10 -> DATA LINK ESCAPE
|
||
|
|
u'\x11' # 0x11 -> DEVICE CONTROL ONE
|
||
|
|
u'\x12' # 0x12 -> DEVICE CONTROL TWO
|
||
|
|
u'\x13' # 0x13 -> DEVICE CONTROL THREE
|
||
|
|
u'\x14' # 0x14 -> DEVICE CONTROL FOUR
|
||
|
|
u'\x15' # 0x15 -> NEGATIVE ACKNOWLEDGE
|
||
|
|
u'\x16' # 0x16 -> SYNCHRONOUS IDLE
|
||
|
|
u'\x17' # 0x17 -> END OF TRANSMISSION BLOCK
|
||
|
|
u'\x18' # 0x18 -> CANCEL
|
||
|
|
u'\x19' # 0x19 -> END OF MEDIUM
|
||
|
|
u'\x1a' # 0x1A -> SUBSTITUTE
|
||
|
|
u'\x1b' # 0x1B -> ESCAPE
|
||
|
|
u'\x1c' # 0x1C -> FILE SEPARATOR
|
||
|
|
u'\x1d' # 0x1D -> GROUP SEPARATOR
|
||
|
|
u'\x1e' # 0x1E -> RECORD SEPARATOR
|
||
|
|
u'\x1f' # 0x1F -> UNIT SEPARATOR
|
||
|
|
u' ' # 0x20 -> SPACE
|
||
|
|
u'!' # 0x21 -> EXCLAMATION MARK
|
||
|
|
u'"' # 0x22 -> QUOTATION MARK
|
||
|
|
u'#' # 0x23 -> NUMBER SIGN
|
||
|
|
u'$' # 0x24 -> DOLLAR SIGN
|
||
|
|
u'%' # 0x25 -> PERCENT SIGN
|
||
|
|
u'&' # 0x26 -> AMPERSAND
|
||
|
|
u"'" # 0x27 -> APOSTROPHE
|
||
|
|
u'(' # 0x28 -> LEFT PARENTHESIS
|
||
|
|
u')' # 0x29 -> RIGHT PARENTHESIS
|
||
|
|
u'*' # 0x2A -> ASTERISK
|
||
|
|
u'+' # 0x2B -> PLUS SIGN
|
||
|
|
u',' # 0x2C -> COMMA
|
||
|
|
u'-' # 0x2D -> HYPHEN-MINUS
|
||
|
|
u'.' # 0x2E -> FULL STOP
|
||
|
|
u'/' # 0x2F -> SOLIDUS
|
||
|
|
u'0' # 0x30 -> DIGIT ZERO
|
||
|
|
u'1' # 0x31 -> DIGIT ONE
|
||
|
|
u'2' # 0x32 -> DIGIT TWO
|
||
|
|
u'3' # 0x33 -> DIGIT THREE
|
||
|
|
u'4' # 0x34 -> DIGIT FOUR
|
||
|
|
u'5' # 0x35 -> DIGIT FIVE
|
||
|
|
u'6' # 0x36 -> DIGIT SIX
|
||
|
|
u'7' # 0x37 -> DIGIT SEVEN
|
||
|
|
u'8' # 0x38 -> DIGIT EIGHT
|
||
|
|
u'9' # 0x39 -> DIGIT NINE
|
||
|
|
u':' # 0x3A -> COLON
|
||
|
|
u';' # 0x3B -> SEMICOLON
|
||
|
|
u'<' # 0x3C -> LESS-THAN SIGN
|
||
|
|
u'=' # 0x3D -> EQUALS SIGN
|
||
|
|
u'>' # 0x3E -> GREATER-THAN SIGN
|
||
|
|
u'?' # 0x3F -> QUESTION MARK
|
||
|
|
u'@' # 0x40 -> COMMERCIAL AT
|
||
|
|
u'A' # 0x41 -> LATIN CAPITAL LETTER A
|
||
|
|
u'B' # 0x42 -> LATIN CAPITAL LETTER B
|
||
|
|
u'C' # 0x43 -> LATIN CAPITAL LETTER C
|
||
|
|
u'D' # 0x44 -> LATIN CAPITAL LETTER D
|
||
|
|
u'E' # 0x45 -> LATIN CAPITAL LETTER E
|
||
|
|
u'F' # 0x46 -> LATIN CAPITAL LETTER F
|
||
|
|
u'G' # 0x47 -> LATIN CAPITAL LETTER G
|
||
|
|
u'H' # 0x48 -> LATIN CAPITAL LETTER H
|
||
|
|
u'I' # 0x49 -> LATIN CAPITAL LETTER I
|
||
|
|
u'J' # 0x4A -> LATIN CAPITAL LETTER J
|
||
|
|
u'K' # 0x4B -> LATIN CAPITAL LETTER K
|
||
|
|
u'L' # 0x4C -> LATIN CAPITAL LETTER L
|
||
|
|
u'M' # 0x4D -> LATIN CAPITAL LETTER M
|
||
|
|
u'N' # 0x4E -> LATIN CAPITAL LETTER N
|
||
|
|
u'O' # 0x4F -> LATIN CAPITAL LETTER O
|
||
|
|
u'P' # 0x50 -> LATIN CAPITAL LETTER P
|
||
|
|
u'Q' # 0x51 -> LATIN CAPITAL LETTER Q
|
||
|
|
u'R' # 0x52 -> LATIN CAPITAL LETTER R
|
||
|
|
u'S' # 0x53 -> LATIN CAPITAL LETTER S
|
||
|
|
u'T' # 0x54 -> LATIN CAPITAL LETTER T
|
||
|
|
u'U' # 0x55 -> LATIN CAPITAL LETTER U
|
||
|
|
u'V' # 0x56 -> LATIN CAPITAL LETTER V
|
||
|
|
u'W' # 0x57 -> LATIN CAPITAL LETTER W
|
||
|
|
u'X' # 0x58 -> LATIN CAPITAL LETTER X
|
||
|
|
u'Y' # 0x59 -> LATIN CAPITAL LETTER Y
|
||
|
|
u'Z' # 0x5A -> LATIN CAPITAL LETTER Z
|
||
|
|
u'[' # 0x5B -> LEFT SQUARE BRACKET
|
||
|
|
u'\\' # 0x5C -> REVERSE SOLIDUS
|
||
|
|
u']' # 0x5D -> RIGHT SQUARE BRACKET
|
||
|
|
u'^' # 0x5E -> CIRCUMFLEX ACCENT
|
||
|
|
u'_' # 0x5F -> LOW LINE
|
||
|
|
u'`' # 0x60 -> GRAVE ACCENT
|
||
|
|
u'a' # 0x61 -> LATIN SMALL LETTER A
|
||
|
|
u'b' # 0x62 -> LATIN SMALL LETTER B
|
||
|
|
u'c' # 0x63 -> LATIN SMALL LETTER C
|
||
|
|
u'd' # 0x64 -> LATIN SMALL LETTER D
|
||
|
|
u'e' # 0x65 -> LATIN SMALL LETTER E
|
||
|
|
u'f' # 0x66 -> LATIN SMALL LETTER F
|
||
|
|
u'g' # 0x67 -> LATIN SMALL LETTER G
|
||
|
|
u'h' # 0x68 -> LATIN SMALL LETTER H
|
||
|
|
u'i' # 0x69 -> LATIN SMALL LETTER I
|
||
|
|
u'j' # 0x6A -> LATIN SMALL LETTER J
|
||
|
|
u'k' # 0x6B -> LATIN SMALL LETTER K
|
||
|
|
u'l' # 0x6C -> LATIN SMALL LETTER L
|
||
|
|
u'm' # 0x6D -> LATIN SMALL LETTER M
|
||
|
|
u'n' # 0x6E -> LATIN SMALL LETTER N
|
||
|
|
u'o' # 0x6F -> LATIN SMALL LETTER O
|
||
|
|
u'p' # 0x70 -> LATIN SMALL LETTER P
|
||
|
|
u'q' # 0x71 -> LATIN SMALL LETTER Q
|
||
|
|
u'r' # 0x72 -> LATIN SMALL LETTER R
|
||
|
|
u's' # 0x73 -> LATIN SMALL LETTER S
|
||
|
|
u't' # 0x74 -> LATIN SMALL LETTER T
|
||
|
|
u'u' # 0x75 -> LATIN SMALL LETTER U
|
||
|
|
u'v' # 0x76 -> LATIN SMALL LETTER V
|
||
|
|
u'w' # 0x77 -> LATIN SMALL LETTER W
|
||
|
|
u'x' # 0x78 -> LATIN SMALL LETTER X
|
||
|
|
u'y' # 0x79 -> LATIN SMALL LETTER Y
|
||
|
|
u'z' # 0x7A -> LATIN SMALL LETTER Z
|
||
|
|
u'{' # 0x7B -> LEFT CURLY BRACKET
|
||
|
|
u'|' # 0x7C -> VERTICAL LINE
|
||
|
|
u'}' # 0x7D -> RIGHT CURLY BRACKET
|
||
|
|
u'~' # 0x7E -> TILDE
|
||
|
|
u'\x7f' # 0x7F -> DELETE
|
||
|
|
u'\x80' # 0x80 -> <control>
|
||
|
|
u'\x81' # 0x81 -> <control>
|
||
|
|
u'\x82' # 0x82 -> <control>
|
||
|
|
u'\x83' # 0x83 -> <control>
|
||
|
|
u'\x84' # 0x84 -> <control>
|
||
|
|
u'\x85' # 0x85 -> <control>
|
||
|
|
u'\x86' # 0x86 -> <control>
|
||
|
|
u'\x87' # 0x87 -> <control>
|
||
|
|
u'\x88' # 0x88 -> <control>
|
||
|
|
u'\x89' # 0x89 -> <control>
|
||
|
|
u'\x8a' # 0x8A -> <control>
|
||
|
|
u'\x8b' # 0x8B -> <control>
|
||
|
|
u'\x8c' # 0x8C -> <control>
|
||
|
|
u'\x8d' # 0x8D -> <control>
|
||
|
|
u'\x8e' # 0x8E -> <control>
|
||
|
|
u'\x8f' # 0x8F -> <control>
|
||
|
|
u'\x90' # 0x90 -> <control>
|
||
|
|
u'\x91' # 0x91 -> <control>
|
||
|
|
u'\x92' # 0x92 -> <control>
|
||
|
|
u'\x93' # 0x93 -> <control>
|
||
|
|
u'\x94' # 0x94 -> <control>
|
||
|
|
u'\x95' # 0x95 -> <control>
|
||
|
|
u'\x96' # 0x96 -> <control>
|
||
|
|
u'\x97' # 0x97 -> <control>
|
||
|
|
u'\x98' # 0x98 -> <control>
|
||
|
|
u'\x99' # 0x99 -> <control>
|
||
|
|
u'\x9a' # 0x9A -> <control>
|
||
|
|
u'\x9b' # 0x9B -> <control>
|
||
|
|
u'\x9c' # 0x9C -> <control>
|
||
|
|
u'\x9d' # 0x9D -> <control>
|
||
|
|
u'\x9e' # 0x9E -> <control>
|
||
|
|
u'\x9f' # 0x9F -> <control>
|
||
|
|
u'\xa0' # 0xA0 -> NO-BREAK SPACE
|
||
|
|
u'\u0126' # 0xA1 -> LATIN CAPITAL LETTER H WITH STROKE
|
||
|
|
u'\u02d8' # 0xA2 -> BREVE
|
||
|
|
u'\xa3' # 0xA3 -> POUND SIGN
|
||
|
|
u'\xa4' # 0xA4 -> CURRENCY SIGN
|
||
|
|
u'\ufffe'
|
||
|
|
u'\u0124' # 0xA6 -> LATIN CAPITAL LETTER H WITH CIRCUMFLEX
|
||
|
|
u'\xa7' # 0xA7 -> SECTION SIGN
|
||
|
|
u'\xa8' # 0xA8 -> DIAERESIS
|
||
|
|
u'\u0130' # 0xA9 -> LATIN CAPITAL LETTER I WITH DOT ABOVE
|
||
|
|
u'\u015e' # 0xAA -> LATIN CAPITAL LETTER S WITH CEDILLA
|
||
|
|
u'\u011e' # 0xAB -> LATIN CAPITAL LETTER G WITH BREVE
|
||
|
|
u'\u0134' # 0xAC -> LATIN CAPITAL LETTER J WITH CIRCUMFLEX
|
||
|
|
u'\xad' # 0xAD -> SOFT HYPHEN
|
||
|
|
u'\ufffe'
|
||
|
|
u'\u017b' # 0xAF -> LATIN CAPITAL LETTER Z WITH DOT ABOVE
|
||
|
|
u'\xb0' # 0xB0 -> DEGREE SIGN
|
||
|
|
u'\u0127' # 0xB1 -> LATIN SMALL LETTER H WITH STROKE
|
||
|
|
u'\xb2' # 0xB2 -> SUPERSCRIPT TWO
|
||
|
|
u'\xb3' # 0xB3 -> SUPERSCRIPT THREE
|
||
|
|
u'\xb4' # 0xB4 -> ACUTE ACCENT
|
||
|
|
u'\xb5' # 0xB5 -> MICRO SIGN
|
||
|
|
u'\u0125' # 0xB6 -> LATIN SMALL LETTER H WITH CIRCUMFLEX
|
||
|
|
u'\xb7' # 0xB7 -> MIDDLE DOT
|
||
|
|
u'\xb8' # 0xB8 -> CEDILLA
|
||
|
|
u'\u0131' # 0xB9 -> LATIN SMALL LETTER DOTLESS I
|
||
|
|
u'\u015f' # 0xBA -> LATIN SMALL LETTER S WITH CEDILLA
|
||
|
|
u'\u011f' # 0xBB -> LATIN SMALL LETTER G WITH BREVE
|
||
|
|
u'\u0135' # 0xBC -> LATIN SMALL LETTER J WITH CIRCUMFLEX
|
||
|
|
u'\xbd' # 0xBD -> VULGAR FRACTION ONE HALF
|
||
|
|
u'\ufffe'
|
||
|
|
u'\u017c' # 0xBF -> LATIN SMALL LETTER Z WITH DOT ABOVE
|
||
|
|
u'\xc0' # 0xC0 -> LATIN CAPITAL LETTER A WITH GRAVE
|
||
|
|
u'\xc1' # 0xC1 -> LATIN CAPITAL LETTER A WITH ACUTE
|
||
|
|
u'\xc2' # 0xC2 -> LATIN CAPITAL LETTER A WITH CIRCUMFLEX
|
||
|
|
u'\ufffe'
|
||
|
|
u'\xc4' # 0xC4 -> LATIN CAPITAL LETTER A WITH DIAERESIS
|
||
|
|
u'\u010a' # 0xC5 -> LATIN CAPITAL LETTER C WITH DOT ABOVE
|
||
|
|
u'\u0108' # 0xC6 -> LATIN CAPITAL LETTER C WITH CIRCUMFLEX
|
||
|
|
u'\xc7' # 0xC7 -> LATIN CAPITAL LETTER C WITH CEDILLA
|
||
|
|
u'\xc8' # 0xC8 -> LATIN CAPITAL LETTER E WITH GRAVE
|
||
|
|
u'\xc9' # 0xC9 -> LATIN CAPITAL LETTER E WITH ACUTE
|
||
|
|
u'\xca' # 0xCA -> LATIN CAPITAL LETTER E WITH CIRCUMFLEX
|
||
|
|
u'\xcb' # 0xCB -> LATIN CAPITAL LETTER E WITH DIAERESIS
|
||
|
|
u'\xcc' # 0xCC -> LATIN CAPITAL LETTER I WITH GRAVE
|
||
|
|
u'\xcd' # 0xCD -> LATIN CAPITAL LETTER I WITH ACUTE
|
||
|
|
u'\xce' # 0xCE -> LATIN CAPITAL LETTER I WITH CIRCUMFLEX
|
||
|
|
u'\xcf' # 0xCF -> LATIN CAPITAL LETTER I WITH DIAERESIS
|
||
|
|
u'\ufffe'
|
||
|
|
u'\xd1' # 0xD1 -> LATIN CAPITAL LETTER N WITH TILDE
|
||
|
|
u'\xd2' # 0xD2 -> LATIN CAPITAL LETTER O WITH GRAVE
|
||
|
|
u'\xd3' # 0xD3 -> LATIN CAPITAL LETTER O WITH ACUTE
|
||
|
|
u'\xd4' # 0xD4 -> LATIN CAPITAL LETTER O WITH CIRCUMFLEX
|
||
|
|
u'\u0120' # 0xD5 -> LATIN CAPITAL LETTER G WITH DOT ABOVE
|
||
|
|
u'\xd6' # 0xD6 -> LATIN CAPITAL LETTER O WITH DIAERESIS
|
||
|
|
u'\xd7' # 0xD7 -> MULTIPLICATION SIGN
|
||
|
|
u'\u011c' # 0xD8 -> LATIN CAPITAL LETTER G WITH CIRCUMFLEX
|
||
|
|
u'\xd9' # 0xD9 -> LATIN CAPITAL LETTER U WITH GRAVE
|
||
|
|
u'\xda' # 0xDA -> LATIN CAPITAL LETTER U WITH ACUTE
|
||
|
|
u'\xdb' # 0xDB -> LATIN CAPITAL LETTER U WITH CIRCUMFLEX
|
||
|
|
u'\xdc' # 0xDC -> LATIN CAPITAL LETTER U WITH DIAERESIS
|
||
|
|
u'\u016c' # 0xDD -> LATIN CAPITAL LETTER U WITH BREVE
|
||
|
|
u'\u015c' # 0xDE -> LATIN CAPITAL LETTER S WITH CIRCUMFLEX
|
||
|
|
u'\xdf' # 0xDF -> LATIN SMALL LETTER SHARP S
|
||
|
|
u'\xe0' # 0xE0 -> LATIN SMALL LETTER A WITH GRAVE
|
||
|
|
u'\xe1' # 0xE1 -> LATIN SMALL LETTER A WITH ACUTE
|
||
|
|
u'\xe2' # 0xE2 -> LATIN SMALL LETTER A WITH CIRCUMFLEX
|
||
|
|
u'\ufffe'
|
||
|
|
u'\xe4' # 0xE4 -> LATIN SMALL LETTER A WITH DIAERESIS
|
||
|
|
u'\u010b' # 0xE5 -> LATIN SMALL LETTER C WITH DOT ABOVE
|
||
|
|
u'\u0109' # 0xE6 -> LATIN SMALL LETTER C WITH CIRCUMFLEX
|
||
|
|
u'\xe7' # 0xE7 -> LATIN SMALL LETTER C WITH CEDILLA
|
||
|
|
u'\xe8' # 0xE8 -> LATIN SMALL LETTER E WITH GRAVE
|
||
|
|
u'\xe9' # 0xE9 -> LATIN SMALL LETTER E WITH ACUTE
|
||
|
|
u'\xea' # 0xEA -> LATIN SMALL LETTER E WITH CIRCUMFLEX
|
||
|
|
u'\xeb' # 0xEB -> LATIN SMALL LETTER E WITH DIAERESIS
|
||
|
|
u'\xec' # 0xEC -> LATIN SMALL LETTER I WITH GRAVE
|
||
|
|
u'\xed' # 0xED -> LATIN SMALL LETTER I WITH ACUTE
|
||
|
|
u'\xee' # 0xEE -> LATIN SMALL LETTER I WITH CIRCUMFLEX
|
||
|
|
u'\xef' # 0xEF -> LATIN SMALL LETTER I WITH DIAERESIS
|
||
|
|
u'\ufffe'
|
||
|
|
u'\xf1' # 0xF1 -> LATIN SMALL LETTER N WITH TILDE
|
||
|
|
u'\xf2' # 0xF2 -> LATIN SMALL LETTER O WITH GRAVE
|
||
|
|
u'\xf3' # 0xF3 -> LATIN SMALL LETTER O WITH ACUTE
|
||
|
|
u'\xf4' # 0xF4 -> LATIN SMALL LETTER O WITH CIRCUMFLEX
|
||
|
|
u'\u0121' # 0xF5 -> LATIN SMALL LETTER G WITH DOT ABOVE
|
||
|
|
u'\xf6' # 0xF6 -> LATIN SMALL LETTER O WITH DIAERESIS
|
||
|
|
u'\xf7' # 0xF7 -> DIVISION SIGN
|
||
|
|
u'\u011d' # 0xF8 -> LATIN SMALL LETTER G WITH CIRCUMFLEX
|
||
|
|
u'\xf9' # 0xF9 -> LATIN SMALL LETTER U WITH GRAVE
|
||
|
|
u'\xfa' # 0xFA -> LATIN SMALL LETTER U WITH ACUTE
|
||
|
|
u'\xfb' # 0xFB -> LATIN SMALL LETTER U WITH CIRCUMFLEX
|
||
|
|
u'\xfc' # 0xFC -> LATIN SMALL LETTER U WITH DIAERESIS
|
||
|
|
u'\u016d' # 0xFD -> LATIN SMALL LETTER U WITH BREVE
|
||
|
|
u'\u015d' # 0xFE -> LATIN SMALL LETTER S WITH CIRCUMFLEX
|
||
|
|
u'\u02d9' # 0xFF -> DOT ABOVE
|
||
|
|
)
|
||
|
|
|
||
|
|
### Encoding table
|
||
|
|
encoding_table=codecs.charmap_build(decoding_table)
|