Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 1 | """ codecs -- Python Codec Registry, API and helpers. |
| 2 | |
| 3 | |
| 4 | Written by Marc-Andre Lemburg (mal@lemburg.com). |
| 5 | |
| 6 | (c) Copyright CNRI, All Rights Reserved. NO WARRANTY. |
| 7 | |
| 8 | """#" |
| 9 | |
Martin v. Löwis | 02d893c | 2001-08-02 07:15:29 +0000 | [diff] [blame] | 10 | import struct, __builtin__ |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 11 | |
| 12 | ### Registry and builtin stateless codec functions |
| 13 | |
Guido van Rossum | b95de4f | 2000-03-31 17:25:23 +0000 | [diff] [blame] | 14 | try: |
| 15 | from _codecs import * |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 16 | except ImportError, why: |
Guido van Rossum | b95de4f | 2000-03-31 17:25:23 +0000 | [diff] [blame] | 17 | raise SystemError,\ |
| 18 | 'Failed to load the builtin codecs: %s' % why |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 19 | |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 20 | __all__ = ["register", "lookup", "open", "EncodedFile", "BOM", "BOM_BE", |
| 21 | "BOM_LE", "BOM32_BE", "BOM32_LE", "BOM64_BE", "BOM64_LE"] |
Skip Montanaro | e99d5ea | 2001-01-20 19:54:20 +0000 | [diff] [blame] | 22 | |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 23 | ### Constants |
| 24 | |
| 25 | # |
| 26 | # Byte Order Mark (BOM) and its possible values (BOM_BE, BOM_LE) |
| 27 | # |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 28 | BOM = struct.pack('=H', 0xFEFF) |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 29 | # |
| 30 | BOM_BE = BOM32_BE = '\376\377' |
Tim Peters | 88869f9 | 2001-01-14 23:36:06 +0000 | [diff] [blame] | 31 | # corresponds to Unicode U+FEFF in UTF-16 on big endian |
| 32 | # platforms == ZERO WIDTH NO-BREAK SPACE |
Guido van Rossum | 1c89b0e | 2000-04-11 15:41:38 +0000 | [diff] [blame] | 33 | BOM_LE = BOM32_LE = '\377\376' |
Tim Peters | 88869f9 | 2001-01-14 23:36:06 +0000 | [diff] [blame] | 34 | # corresponds to Unicode U+FFFE in UTF-16 on little endian |
| 35 | # platforms == defined as being an illegal Unicode character |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 36 | |
| 37 | # |
| 38 | # 64-bit Byte Order Marks |
| 39 | # |
| 40 | BOM64_BE = '\000\000\376\377' |
Tim Peters | 88869f9 | 2001-01-14 23:36:06 +0000 | [diff] [blame] | 41 | # corresponds to Unicode U+0000FEFF in UCS-4 |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 42 | BOM64_LE = '\377\376\000\000' |
Tim Peters | 88869f9 | 2001-01-14 23:36:06 +0000 | [diff] [blame] | 43 | # corresponds to Unicode U+0000FFFE in UCS-4 |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 44 | |
| 45 | |
| 46 | ### Codec base classes (defining the API) |
| 47 | |
| 48 | class Codec: |
| 49 | |
| 50 | """ Defines the interface for stateless encoders/decoders. |
| 51 | |
| 52 | The .encode()/.decode() methods may implement different error |
| 53 | handling schemes by providing the errors argument. These |
| 54 | string values are defined: |
| 55 | |
Guido van Rossum | d8855fd | 2000-03-24 22:14:19 +0000 | [diff] [blame] | 56 | 'strict' - raise a ValueError error (or a subclass) |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 57 | 'ignore' - ignore the character and continue with the next |
| 58 | 'replace' - replace with a suitable replacement character; |
| 59 | Python will use the official U+FFFD REPLACEMENT |
| 60 | CHARACTER for the builtin Unicode codecs. |
| 61 | |
| 62 | """ |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 63 | def encode(self, input, errors='strict'): |
Guido van Rossum | 1c89b0e | 2000-04-11 15:41:38 +0000 | [diff] [blame] | 64 | |
Fred Drake | 3e74c0d | 2000-03-17 15:40:35 +0000 | [diff] [blame] | 65 | """ Encodes the object input and returns a tuple (output |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 66 | object, length consumed). |
| 67 | |
| 68 | errors defines the error handling to apply. It defaults to |
| 69 | 'strict' handling. |
| 70 | |
| 71 | The method may not store state in the Codec instance. Use |
| 72 | StreamCodec for codecs which have to keep state in order to |
| 73 | make encoding/decoding efficient. |
| 74 | |
| 75 | The encoder must be able to handle zero length input and |
| 76 | return an empty object of the output object type in this |
| 77 | situation. |
| 78 | |
| 79 | """ |
| 80 | raise NotImplementedError |
| 81 | |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 82 | def decode(self, input, errors='strict'): |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 83 | |
| 84 | """ Decodes the object input and returns a tuple (output |
| 85 | object, length consumed). |
| 86 | |
| 87 | input must be an object which provides the bf_getreadbuf |
| 88 | buffer slot. Python strings, buffer objects and memory |
| 89 | mapped files are examples of objects providing this slot. |
Guido van Rossum | 1c89b0e | 2000-04-11 15:41:38 +0000 | [diff] [blame] | 90 | |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 91 | errors defines the error handling to apply. It defaults to |
| 92 | 'strict' handling. |
| 93 | |
| 94 | The method may not store state in the Codec instance. Use |
| 95 | StreamCodec for codecs which have to keep state in order to |
| 96 | make encoding/decoding efficient. |
| 97 | |
| 98 | The decoder must be able to handle zero length input and |
| 99 | return an empty object of the output object type in this |
| 100 | situation. |
| 101 | |
Guido van Rossum | 1c89b0e | 2000-04-11 15:41:38 +0000 | [diff] [blame] | 102 | """ |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 103 | raise NotImplementedError |
| 104 | |
| 105 | # |
| 106 | # The StreamWriter and StreamReader class provide generic working |
Andrew M. Kuchling | 97c5635 | 2001-09-18 20:29:48 +0000 | [diff] [blame] | 107 | # interfaces which can be used to implement new encoding submodules |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 108 | # very easily. See encodings/utf_8.py for an example on how this is |
| 109 | # done. |
Guido van Rossum | 1c89b0e | 2000-04-11 15:41:38 +0000 | [diff] [blame] | 110 | # |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 111 | |
| 112 | class StreamWriter(Codec): |
| 113 | |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 114 | def __init__(self, stream, errors='strict'): |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 115 | |
| 116 | """ Creates a StreamWriter instance. |
| 117 | |
| 118 | stream must be a file-like object open for writing |
| 119 | (binary) data. |
| 120 | |
| 121 | The StreamWriter may implement different error handling |
| 122 | schemes by providing the errors keyword argument. These |
| 123 | parameters are defined: |
| 124 | |
| 125 | 'strict' - raise a ValueError (or a subclass) |
| 126 | 'ignore' - ignore the character and continue with the next |
| 127 | 'replace'- replace with a suitable replacement character |
| 128 | |
| 129 | """ |
| 130 | self.stream = stream |
| 131 | self.errors = errors |
| 132 | |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 133 | def write(self, object): |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 134 | |
| 135 | """ Writes the object's contents encoded to self.stream. |
| 136 | """ |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 137 | data, consumed = self.encode(object, self.errors) |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 138 | self.stream.write(data) |
| 139 | |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 140 | def writelines(self, list): |
| 141 | |
| 142 | """ Writes the concatenated list of strings to the stream |
| 143 | using .write(). |
| 144 | """ |
| 145 | self.write(''.join(list)) |
Guido van Rossum | 1c89b0e | 2000-04-11 15:41:38 +0000 | [diff] [blame] | 146 | |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 147 | def reset(self): |
| 148 | |
| 149 | """ Flushes and resets the codec buffers used for keeping state. |
| 150 | |
| 151 | Calling this method should ensure that the data on the |
| 152 | output is put into a clean state, that allows appending |
| 153 | of new fresh data without having to rescan the whole |
| 154 | stream to recover state. |
| 155 | |
| 156 | """ |
| 157 | pass |
| 158 | |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 159 | def __getattr__(self, name, |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 160 | getattr=getattr): |
| 161 | |
| 162 | """ Inherit all other methods from the underlying stream. |
| 163 | """ |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 164 | return getattr(self.stream, name) |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 165 | |
| 166 | ### |
| 167 | |
| 168 | class StreamReader(Codec): |
| 169 | |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 170 | def __init__(self, stream, errors='strict'): |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 171 | |
| 172 | """ Creates a StreamReader instance. |
| 173 | |
| 174 | stream must be a file-like object open for reading |
| 175 | (binary) data. |
| 176 | |
| 177 | The StreamReader may implement different error handling |
| 178 | schemes by providing the errors keyword argument. These |
| 179 | parameters are defined: |
| 180 | |
| 181 | 'strict' - raise a ValueError (or a subclass) |
| 182 | 'ignore' - ignore the character and continue with the next |
| 183 | 'replace'- replace with a suitable replacement character; |
| 184 | |
| 185 | """ |
| 186 | self.stream = stream |
| 187 | self.errors = errors |
| 188 | |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 189 | def read(self, size=-1): |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 190 | |
| 191 | """ Decodes data from the stream self.stream and returns the |
| 192 | resulting object. |
| 193 | |
| 194 | size indicates the approximate maximum number of bytes to |
| 195 | read from the stream for decoding purposes. The decoder |
| 196 | can modify this setting as appropriate. The default value |
| 197 | -1 indicates to read and decode as much as possible. size |
| 198 | is intended to prevent having to decode huge files in one |
| 199 | step. |
| 200 | |
| 201 | The method should use a greedy read strategy meaning that |
| 202 | it should read as much data as is allowed within the |
| 203 | definition of the encoding and the given size, e.g. if |
| 204 | optional encoding endings or state markers are available |
| 205 | on the stream, these should be read too. |
| 206 | |
| 207 | """ |
| 208 | # Unsliced reading: |
| 209 | if size < 0: |
Andrew M. Kuchling | c6c2838 | 2000-12-10 15:12:14 +0000 | [diff] [blame] | 210 | return self.decode(self.stream.read(), self.errors)[0] |
Guido van Rossum | 1c89b0e | 2000-04-11 15:41:38 +0000 | [diff] [blame] | 211 | |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 212 | # Sliced reading: |
| 213 | read = self.stream.read |
| 214 | decode = self.decode |
| 215 | data = read(size) |
| 216 | i = 0 |
| 217 | while 1: |
| 218 | try: |
Andrew M. Kuchling | c6c2838 | 2000-12-10 15:12:14 +0000 | [diff] [blame] | 219 | object, decodedbytes = decode(data, self.errors) |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 220 | except ValueError, why: |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 221 | # This method is slow but should work under pretty much |
| 222 | # all conditions; at most 10 tries are made |
| 223 | i = i + 1 |
| 224 | newdata = read(1) |
| 225 | if not newdata or i > 10: |
| 226 | raise |
| 227 | data = data + newdata |
| 228 | else: |
| 229 | return object |
| 230 | |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 231 | def readline(self, size=None): |
| 232 | |
| 233 | """ Read one line from the input stream and return the |
| 234 | decoded data. |
| 235 | |
Fred Drake | 49fd107 | 2000-04-13 14:11:21 +0000 | [diff] [blame] | 236 | Note: Unlike the .readlines() method, this method inherits |
| 237 | the line breaking knowledge from the underlying stream's |
| 238 | .readline() method -- there is currently no support for |
| 239 | line breaking using the codec decoder due to lack of line |
| 240 | buffering. Sublcasses should however, if possible, try to |
| 241 | implement this method using their own knowledge of line |
| 242 | breaking. |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 243 | |
| 244 | size, if given, is passed as size argument to the stream's |
| 245 | .readline() method. |
Guido van Rossum | 1c89b0e | 2000-04-11 15:41:38 +0000 | [diff] [blame] | 246 | |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 247 | """ |
| 248 | if size is None: |
| 249 | line = self.stream.readline() |
| 250 | else: |
| 251 | line = self.stream.readline(size) |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 252 | return self.decode(line, self.errors)[0] |
Guido van Rossum | 1c89b0e | 2000-04-11 15:41:38 +0000 | [diff] [blame] | 253 | |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 254 | |
Martin v. Löwis | b786e61 | 2002-03-05 15:46:38 +0000 | [diff] [blame^] | 255 | def readlines(self, sizehint=None): |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 256 | |
| 257 | """ Read all lines available on the input stream |
| 258 | and return them as list of lines. |
| 259 | |
| 260 | Line breaks are implemented using the codec's decoder |
| 261 | method and are included in the list entries. |
Guido van Rossum | 1c89b0e | 2000-04-11 15:41:38 +0000 | [diff] [blame] | 262 | |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 263 | sizehint, if given, is passed as size argument to the |
| 264 | stream's .read() method. |
| 265 | |
| 266 | """ |
| 267 | if sizehint is None: |
| 268 | data = self.stream.read() |
| 269 | else: |
| 270 | data = self.stream.read(sizehint) |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 271 | return self.decode(data, self.errors)[0].splitlines(1) |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 272 | |
| 273 | def reset(self): |
| 274 | |
| 275 | """ Resets the codec buffers used for keeping state. |
| 276 | |
| 277 | Note that no stream repositioning should take place. |
Thomas Wouters | 7e47402 | 2000-07-16 12:04:32 +0000 | [diff] [blame] | 278 | This method is primarily intended to be able to recover |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 279 | from decoding errors. |
| 280 | |
| 281 | """ |
| 282 | pass |
| 283 | |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 284 | def __getattr__(self, name, |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 285 | getattr=getattr): |
| 286 | |
| 287 | """ Inherit all other methods from the underlying stream. |
| 288 | """ |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 289 | return getattr(self.stream, name) |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 290 | |
| 291 | ### |
| 292 | |
| 293 | class StreamReaderWriter: |
| 294 | |
Fred Drake | 49fd107 | 2000-04-13 14:11:21 +0000 | [diff] [blame] | 295 | """ StreamReaderWriter instances allow wrapping streams which |
| 296 | work in both read and write modes. |
| 297 | |
| 298 | The design is such that one can use the factory functions |
Thomas Wouters | 7e47402 | 2000-07-16 12:04:32 +0000 | [diff] [blame] | 299 | returned by the codec.lookup() function to construct the |
Fred Drake | 49fd107 | 2000-04-13 14:11:21 +0000 | [diff] [blame] | 300 | instance. |
| 301 | |
| 302 | """ |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 303 | # Optional attributes set by the file wrappers below |
| 304 | encoding = 'unknown' |
| 305 | |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 306 | def __init__(self, stream, Reader, Writer, errors='strict'): |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 307 | |
| 308 | """ Creates a StreamReaderWriter instance. |
| 309 | |
| 310 | stream must be a Stream-like object. |
| 311 | |
| 312 | Reader, Writer must be factory functions or classes |
| 313 | providing the StreamReader, StreamWriter interface resp. |
| 314 | |
| 315 | Error handling is done in the same way as defined for the |
| 316 | StreamWriter/Readers. |
| 317 | |
| 318 | """ |
| 319 | self.stream = stream |
| 320 | self.reader = Reader(stream, errors) |
| 321 | self.writer = Writer(stream, errors) |
| 322 | self.errors = errors |
| 323 | |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 324 | def read(self, size=-1): |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 325 | |
| 326 | return self.reader.read(size) |
| 327 | |
Guido van Rossum | d58c26f | 2000-05-01 16:17:32 +0000 | [diff] [blame] | 328 | def readline(self, size=None): |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 329 | |
| 330 | return self.reader.readline(size) |
| 331 | |
Guido van Rossum | d58c26f | 2000-05-01 16:17:32 +0000 | [diff] [blame] | 332 | def readlines(self, sizehint=None): |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 333 | |
| 334 | return self.reader.readlines(sizehint) |
| 335 | |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 336 | def write(self, data): |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 337 | |
| 338 | return self.writer.write(data) |
| 339 | |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 340 | def writelines(self, list): |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 341 | |
| 342 | return self.writer.writelines(list) |
| 343 | |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 344 | def reset(self): |
| 345 | |
| 346 | self.reader.reset() |
| 347 | self.writer.reset() |
| 348 | |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 349 | def __getattr__(self, name, |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 350 | getattr=getattr): |
| 351 | |
| 352 | """ Inherit all other methods from the underlying stream. |
| 353 | """ |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 354 | return getattr(self.stream, name) |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 355 | |
| 356 | ### |
| 357 | |
| 358 | class StreamRecoder: |
| 359 | |
Fred Drake | 49fd107 | 2000-04-13 14:11:21 +0000 | [diff] [blame] | 360 | """ StreamRecoder instances provide a frontend - backend |
| 361 | view of encoding data. |
| 362 | |
| 363 | They use the complete set of APIs returned by the |
| 364 | codecs.lookup() function to implement their task. |
| 365 | |
| 366 | Data written to the stream is first decoded into an |
| 367 | intermediate format (which is dependent on the given codec |
| 368 | combination) and then written to the stream using an instance |
| 369 | of the provided Writer class. |
| 370 | |
| 371 | In the other direction, data is read from the stream using a |
| 372 | Reader instance and then return encoded data to the caller. |
| 373 | |
| 374 | """ |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 375 | # Optional attributes set by the file wrappers below |
| 376 | data_encoding = 'unknown' |
| 377 | file_encoding = 'unknown' |
| 378 | |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 379 | def __init__(self, stream, encode, decode, Reader, Writer, |
| 380 | errors='strict'): |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 381 | |
| 382 | """ Creates a StreamRecoder instance which implements a two-way |
| 383 | conversion: encode and decode work on the frontend (the |
Guido van Rossum | 1c89b0e | 2000-04-11 15:41:38 +0000 | [diff] [blame] | 384 | input to .read() and output of .write()) while |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 385 | Reader and Writer work on the backend (reading and |
Fred Drake | 908670c | 2000-03-17 15:42:11 +0000 | [diff] [blame] | 386 | writing to the stream). |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 387 | |
| 388 | You can use these objects to do transparent direct |
| 389 | recodings from e.g. latin-1 to utf-8 and back. |
| 390 | |
| 391 | stream must be a file-like object. |
| 392 | |
| 393 | encode, decode must adhere to the Codec interface, Reader, |
| 394 | Writer must be factory functions or classes providing the |
| 395 | StreamReader, StreamWriter interface resp. |
| 396 | |
| 397 | encode and decode are needed for the frontend translation, |
| 398 | Reader and Writer for the backend translation. Unicode is |
| 399 | used as intermediate encoding. |
| 400 | |
| 401 | Error handling is done in the same way as defined for the |
| 402 | StreamWriter/Readers. |
| 403 | |
| 404 | """ |
| 405 | self.stream = stream |
| 406 | self.encode = encode |
| 407 | self.decode = decode |
| 408 | self.reader = Reader(stream, errors) |
| 409 | self.writer = Writer(stream, errors) |
| 410 | self.errors = errors |
| 411 | |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 412 | def read(self, size=-1): |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 413 | |
| 414 | data = self.reader.read(size) |
| 415 | data, bytesencoded = self.encode(data, self.errors) |
| 416 | return data |
| 417 | |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 418 | def readline(self, size=None): |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 419 | |
| 420 | if size is None: |
| 421 | data = self.reader.readline() |
| 422 | else: |
| 423 | data = self.reader.readline(size) |
| 424 | data, bytesencoded = self.encode(data, self.errors) |
| 425 | return data |
| 426 | |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 427 | def readlines(self, sizehint=None): |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 428 | |
| 429 | if sizehint is None: |
| 430 | data = self.reader.read() |
| 431 | else: |
| 432 | data = self.reader.read(sizehint) |
| 433 | data, bytesencoded = self.encode(data, self.errors) |
| 434 | return data.splitlines(1) |
| 435 | |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 436 | def write(self, data): |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 437 | |
| 438 | data, bytesdecoded = self.decode(data, self.errors) |
| 439 | return self.writer.write(data) |
| 440 | |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 441 | def writelines(self, list): |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 442 | |
| 443 | data = ''.join(list) |
| 444 | data, bytesdecoded = self.decode(data, self.errors) |
| 445 | return self.writer.write(data) |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 446 | |
| 447 | def reset(self): |
| 448 | |
| 449 | self.reader.reset() |
| 450 | self.writer.reset() |
| 451 | |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 452 | def __getattr__(self, name, |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 453 | getattr=getattr): |
| 454 | |
| 455 | """ Inherit all other methods from the underlying stream. |
| 456 | """ |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 457 | return getattr(self.stream, name) |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 458 | |
| 459 | ### Shortcuts |
| 460 | |
Marc-André Lemburg | 349a3d3 | 2000-06-21 21:21:04 +0000 | [diff] [blame] | 461 | def open(filename, mode='rb', encoding=None, errors='strict', buffering=1): |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 462 | |
| 463 | """ Open an encoded file using the given mode and return |
| 464 | a wrapped version providing transparent encoding/decoding. |
| 465 | |
| 466 | Note: The wrapped version will only accept the object format |
| 467 | defined by the codecs, i.e. Unicode objects for most builtin |
| 468 | codecs. Output is also codec dependent and will usually by |
| 469 | Unicode as well. |
| 470 | |
Marc-André Lemburg | 349a3d3 | 2000-06-21 21:21:04 +0000 | [diff] [blame] | 471 | Files are always opened in binary mode, even if no binary mode |
| 472 | was specified. Thisis done to avoid data loss due to encodings |
| 473 | using 8-bit values. The default file mode is 'rb' meaning to |
| 474 | open the file in binary read mode. |
| 475 | |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 476 | encoding specifies the encoding which is to be used for the |
| 477 | the file. |
| 478 | |
| 479 | errors may be given to define the error handling. It defaults |
| 480 | to 'strict' which causes ValueErrors to be raised in case an |
| 481 | encoding error occurs. |
| 482 | |
| 483 | buffering has the same meaning as for the builtin open() API. |
| 484 | It defaults to line buffered. |
| 485 | |
Fred Drake | 49fd107 | 2000-04-13 14:11:21 +0000 | [diff] [blame] | 486 | The returned wrapped file object provides an extra attribute |
| 487 | .encoding which allows querying the used encoding. This |
| 488 | attribute is only available if an encoding was specified as |
| 489 | parameter. |
| 490 | |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 491 | """ |
| 492 | if encoding is not None and \ |
| 493 | 'b' not in mode: |
| 494 | # Force opening of the file in binary mode |
| 495 | mode = mode + 'b' |
| 496 | file = __builtin__.open(filename, mode, buffering) |
| 497 | if encoding is None: |
| 498 | return file |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 499 | (e, d, sr, sw) = lookup(encoding) |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 500 | srw = StreamReaderWriter(file, sr, sw, errors) |
| 501 | # Add attributes to simplify introspection |
| 502 | srw.encoding = encoding |
| 503 | return srw |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 504 | |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 505 | def EncodedFile(file, data_encoding, file_encoding=None, errors='strict'): |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 506 | |
| 507 | """ Return a wrapped version of file which provides transparent |
| 508 | encoding translation. |
| 509 | |
| 510 | Strings written to the wrapped file are interpreted according |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 511 | to the given data_encoding and then written to the original |
| 512 | file as string using file_encoding. The intermediate encoding |
| 513 | will usually be Unicode but depends on the specified codecs. |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 514 | |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 515 | Strings are read from the file using file_encoding and then |
| 516 | passed back to the caller as string using data_encoding. |
| 517 | |
| 518 | If file_encoding is not given, it defaults to data_encoding. |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 519 | |
| 520 | errors may be given to define the error handling. It defaults |
| 521 | to 'strict' which causes ValueErrors to be raised in case an |
| 522 | encoding error occurs. |
| 523 | |
Fred Drake | 49fd107 | 2000-04-13 14:11:21 +0000 | [diff] [blame] | 524 | The returned wrapped file object provides two extra attributes |
| 525 | .data_encoding and .file_encoding which reflect the given |
| 526 | parameters of the same name. The attributes can be used for |
| 527 | introspection by Python programs. |
| 528 | |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 529 | """ |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 530 | if file_encoding is None: |
| 531 | file_encoding = data_encoding |
| 532 | encode, decode = lookup(data_encoding)[:2] |
| 533 | Reader, Writer = lookup(file_encoding)[2:] |
| 534 | sr = StreamRecoder(file, |
Tim Peters | 30324a7 | 2001-05-15 17:19:16 +0000 | [diff] [blame] | 535 | encode, decode, Reader, Writer, |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 536 | errors) |
| 537 | # Add attributes to simplify introspection |
| 538 | sr.data_encoding = data_encoding |
| 539 | sr.file_encoding = file_encoding |
| 540 | return sr |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 541 | |
Marc-André Lemburg | aa32c5a | 2001-09-19 11:24:48 +0000 | [diff] [blame] | 542 | ### Helpers for codec lookup |
| 543 | |
| 544 | def getencoder(encoding): |
| 545 | |
| 546 | """ Lookup up the codec for the given encoding and return |
| 547 | its encoder function. |
| 548 | |
| 549 | Raises a LookupError in case the encoding cannot be found. |
| 550 | |
| 551 | """ |
| 552 | return lookup(encoding)[0] |
| 553 | |
| 554 | def getdecoder(encoding): |
| 555 | |
| 556 | """ Lookup up the codec for the given encoding and return |
| 557 | its decoder function. |
| 558 | |
| 559 | Raises a LookupError in case the encoding cannot be found. |
| 560 | |
| 561 | """ |
| 562 | return lookup(encoding)[1] |
| 563 | |
| 564 | def getreader(encoding): |
| 565 | |
| 566 | """ Lookup up the codec for the given encoding and return |
| 567 | its StreamReader class or factory function. |
| 568 | |
| 569 | Raises a LookupError in case the encoding cannot be found. |
| 570 | |
| 571 | """ |
| 572 | return lookup(encoding)[2] |
| 573 | |
| 574 | def getwriter(encoding): |
| 575 | |
| 576 | """ Lookup up the codec for the given encoding and return |
| 577 | its StreamWriter class or factory function. |
| 578 | |
| 579 | Raises a LookupError in case the encoding cannot be found. |
| 580 | |
| 581 | """ |
| 582 | return lookup(encoding)[3] |
| 583 | |
Marc-André Lemburg | a866df8 | 2001-01-03 21:29:14 +0000 | [diff] [blame] | 584 | ### Helpers for charmap-based codecs |
| 585 | |
| 586 | def make_identity_dict(rng): |
| 587 | |
| 588 | """ make_identity_dict(rng) -> dict |
| 589 | |
| 590 | Return a dictionary where elements of the rng sequence are |
| 591 | mapped to themselves. |
Tim Peters | 88869f9 | 2001-01-14 23:36:06 +0000 | [diff] [blame] | 592 | |
Marc-André Lemburg | a866df8 | 2001-01-03 21:29:14 +0000 | [diff] [blame] | 593 | """ |
| 594 | res = {} |
| 595 | for i in rng: |
| 596 | res[i]=i |
| 597 | return res |
| 598 | |
Marc-André Lemburg | 716cf91 | 2001-05-16 09:41:45 +0000 | [diff] [blame] | 599 | def make_encoding_map(decoding_map): |
| 600 | |
| 601 | """ Creates an encoding map from a decoding map. |
| 602 | |
| 603 | If a target mapping in the decoding map occurrs multiple |
| 604 | times, then that target is mapped to None (undefined mapping), |
| 605 | causing an exception when encountered by the charmap codec |
| 606 | during translation. |
| 607 | |
| 608 | One example where this happens is cp875.py which decodes |
| 609 | multiple character to \u001a. |
| 610 | |
| 611 | """ |
| 612 | m = {} |
| 613 | for k,v in decoding_map.items(): |
| 614 | if not m.has_key(v): |
| 615 | m[v] = k |
| 616 | else: |
| 617 | m[v] = None |
| 618 | return m |
Tim Peters | 3a2ab1a | 2001-05-29 06:06:54 +0000 | [diff] [blame] | 619 | |
Martin v. Löwis | 6cd441d | 2001-07-31 08:54:55 +0000 | [diff] [blame] | 620 | # Tell modulefinder that using codecs probably needs the encodings |
| 621 | # package |
| 622 | _false = 0 |
| 623 | if _false: |
| 624 | import encodings |
| 625 | |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 626 | ### Tests |
Guido van Rossum | 1c89b0e | 2000-04-11 15:41:38 +0000 | [diff] [blame] | 627 | |
Guido van Rossum | 0612d84 | 2000-03-10 23:20:43 +0000 | [diff] [blame] | 628 | if __name__ == '__main__': |
| 629 | |
| 630 | import sys |
Guido van Rossum | 1c89b0e | 2000-04-11 15:41:38 +0000 | [diff] [blame] | 631 | |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 632 | # Make stdout translate Latin-1 output into UTF-8 output |
| 633 | sys.stdout = EncodedFile(sys.stdout, 'latin-1', 'utf-8') |
Guido van Rossum | 1c89b0e | 2000-04-11 15:41:38 +0000 | [diff] [blame] | 634 | |
Guido van Rossum | a327713 | 2000-04-11 15:37:43 +0000 | [diff] [blame] | 635 | # Have stdin translate Latin-1 input into UTF-8 input |
| 636 | sys.stdin = EncodedFile(sys.stdin, 'utf-8', 'latin-1') |