| 1 |
import io |
| 2 |
from os import PathLike |
| 3 |
from _zstd import ZstdCompressor, ZstdDecompressor, ZSTD_DStreamOutSize |
| 4 |
from compression._common import _streams |
| 5 |
|
| 6 |
__all__ = ('ZstdFile', 'open') |
| 7 |
|
| 8 |
_MODE_CLOSED = 0 |
| 9 |
_MODE_READ = 1 |
| 10 |
_MODE_WRITE = 2 |
| 11 |
|
| 12 |
|
| 13 |
def _nbytes(dat, /): |
| 14 |
if isinstance(dat, (bytes, bytearray)): |
| 15 |
return len(dat) |
| 16 |
with memoryview(dat) as mv: |
| 17 |
return mv.nbytes |
| 18 |
|
| 19 |
|
| 20 |
class ZstdFile(_streams.BaseStream): |
| 21 |
"""A file-like object providing transparent Zstandard (de)compression. |
| 22 |
|
| 23 |
A ZstdFile can act as a wrapper for an existing file object, or refer |
| 24 |
directly to a named file on disk. |
| 25 |
|
| 26 |
ZstdFile provides a *binary* file interface. Data is read and returned as |
| 27 |
bytes, and may only be written to objects that support the Buffer Protocol. |
| 28 |
""" |
| 29 |
|
| 30 |
FLUSH_BLOCK = ZstdCompressor.FLUSH_BLOCK |
| 31 |
FLUSH_FRAME = ZstdCompressor.FLUSH_FRAME |
| 32 |
|
| 33 |
def __init__(self, file, /, mode='r', *, |
| 34 |
level=None, options=None, zstd_dict=None): |
| 35 |
"""Open a Zstandard compressed file in binary mode. |
| 36 |
|
| 37 |
*file* can be either an file-like object, or a file name to open. |
| 38 |
|
| 39 |
*mode* can be 'r' for reading (default), 'w' for (over)writing, 'x' for |
| 40 |
creating exclusively, or 'a' for appending. These can equivalently be |
| 41 |
given as 'rb', 'wb', 'xb' and 'ab' respectively. |
| 42 |
|
| 43 |
*level* is an optional int specifying the compression level to use, |
| 44 |
or COMPRESSION_LEVEL_DEFAULT if not given. |
| 45 |
|
| 46 |
*options* is an optional dict for advanced compression parameters. |
| 47 |
See CompressionParameter and DecompressionParameter for the possible |
| 48 |
options. |
| 49 |
|
| 50 |
*zstd_dict* is an optional ZstdDict object, a pre-trained Zstandard |
| 51 |
dictionary. See train_dict() to train ZstdDict on sample data. |
| 52 |
""" |
| 53 |
self._fp = None |
| 54 |
self._close_fp = False |
| 55 |
self._mode = _MODE_CLOSED |
| 56 |
self._buffer = None |
| 57 |
|
| 58 |
if not isinstance(mode, str): |
| 59 |
raise ValueError('mode must be a str') |
| 60 |
if options is not None and not isinstance(options, dict): |
| 61 |
raise TypeError('options must be a dict or None') |
| 62 |
mode = mode.removesuffix('b') # handle rb, wb, xb, ab |
| 63 |
if mode == 'r': |
| 64 |
if level is not None: |
| 65 |
raise TypeError('level is illegal in read mode') |
| 66 |
self._mode = _MODE_READ |
| 67 |
elif mode in {'w', 'a', 'x'}: |
| 68 |
if level is not None and not isinstance(level, int): |
| 69 |
raise TypeError('level must be int or None') |
| 70 |
self._mode = _MODE_WRITE |
| 71 |
self._compressor = ZstdCompressor(level=level, options=options, |
| 72 |
zstd_dict=zstd_dict) |
| 73 |
self._pos = 0 |
| 74 |
else: |
| 75 |
raise ValueError(f'Invalid mode: {mode!r}') |
| 76 |
|
| 77 |
if isinstance(file, (str, bytes, PathLike)): |
| 78 |
self._fp = io.open(file, f'{mode}b') |
| 79 |
self._close_fp = True |
| 80 |
elif ((mode == 'r' and hasattr(file, 'read')) |
| 81 |
or (mode != 'r' and hasattr(file, 'write'))): |
| 82 |
self._fp = file |
| 83 |
else: |
| 84 |
raise TypeError('file must be a file-like object ' |
| 85 |
'or a str, bytes, or PathLike object') |
| 86 |
|
| 87 |
if self._mode == _MODE_READ: |
| 88 |
raw = _streams.DecompressReader( |
| 89 |
self._fp, |
| 90 |
ZstdDecompressor, |
| 91 |
zstd_dict=zstd_dict, |
| 92 |
options=options, |
| 93 |
) |
| 94 |
self._buffer = io.BufferedReader(raw) |
| 95 |
|
| 96 |
def close(self): |
| 97 |
"""Flush and close the file. |
| 98 |
|
| 99 |
May be called multiple times. Once the file has been closed, |
| 100 |
any other operation on it will raise ValueError. |
| 101 |
""" |
| 102 |
if self._fp is None: |
| 103 |
return |
| 104 |
try: |
| 105 |
if self._mode == _MODE_READ: |
| 106 |
if getattr(self, '_buffer', None): |
| 107 |
self._buffer.close() |
| 108 |
self._buffer = None |
| 109 |
elif self._mode == _MODE_WRITE: |
| 110 |
self.flush(self.FLUSH_FRAME) |
| 111 |
self._compressor = None |
| 112 |
finally: |
| 113 |
self._mode = _MODE_CLOSED |
| 114 |
try: |
| 115 |
if self._close_fp: |
| 116 |
self._fp.close() |
| 117 |
finally: |
| 118 |
self._fp = None |
| 119 |
self._close_fp = False |
| 120 |
|
| 121 |
def write(self, data, /): |
| 122 |
"""Write a bytes-like object *data* to the file. |
| 123 |
|
| 124 |
Returns the number of uncompressed bytes written, which is |
| 125 |
always the length of data in bytes. Note that due to buffering, |
| 126 |
the file on disk may not reflect the data written until .flush() |
| 127 |
or .close() is called. |
| 128 |
""" |
| 129 |
self._check_can_write() |
| 130 |
|
| 131 |
length = _nbytes(data) |
| 132 |
|
| 133 |
compressed = self._compressor.compress(data) |
| 134 |
self._fp.write(compressed) |
| 135 |
self._pos += length |
| 136 |
return length |
| 137 |
|
| 138 |
def flush(self, mode=FLUSH_BLOCK): |
| 139 |
"""Flush remaining data to the underlying stream. |
| 140 |
|
| 141 |
The mode argument can be FLUSH_BLOCK or FLUSH_FRAME. Abuse of this |
| 142 |
method will reduce compression ratio, use it only when necessary. |
| 143 |
|
| 144 |
If the program is interrupted afterwards, all data can be recovered. |
| 145 |
To ensure saving to disk, also need to use os.fsync(fd). |
| 146 |
|
| 147 |
This method does nothing in reading mode. |
| 148 |
""" |
| 149 |
if self._mode == _MODE_READ: |
| 150 |
return |
| 151 |
self._check_not_closed() |
| 152 |
if mode not in {self.FLUSH_BLOCK, self.FLUSH_FRAME}: |
| 153 |
raise ValueError('Invalid mode argument, expected either ' |
| 154 |
'ZstdFile.FLUSH_FRAME or ' |
| 155 |
'ZstdFile.FLUSH_BLOCK') |
| 156 |
if self._compressor.last_mode == mode: |
| 157 |
return |
| 158 |
# Flush zstd block/frame, and write. |
| 159 |
data = self._compressor.flush(mode) |
| 160 |
self._fp.write(data) |
| 161 |
if hasattr(self._fp, 'flush'): |
| 162 |
self._fp.flush() |
| 163 |
|
| 164 |
def read(self, size=-1): |
| 165 |
"""Read up to size uncompressed bytes from the file. |
| 166 |
|
| 167 |
If size is negative or omitted, read until EOF is reached. |
| 168 |
Returns b'' if the file is already at EOF. |
| 169 |
""" |
| 170 |
if size is None: |
| 171 |
size = -1 |
| 172 |
self._check_can_read() |
| 173 |
return self._buffer.read(size) |
| 174 |
|
| 175 |
def read1(self, size=-1): |
| 176 |
"""Read up to size uncompressed bytes, while trying to avoid |
| 177 |
making multiple reads from the underlying stream. Reads up to a |
| 178 |
buffer's worth of data if size is negative. |
| 179 |
|
| 180 |
Returns b'' if the file is at EOF. |
| 181 |
""" |
| 182 |
self._check_can_read() |
| 183 |
if size < 0: |
| 184 |
# Note this should *not* be io.DEFAULT_BUFFER_SIZE. |
| 185 |
# ZSTD_DStreamOutSize is the minimum amount to read guaranteeing |
| 186 |
# a full block is read. |
| 187 |
size = ZSTD_DStreamOutSize |
| 188 |
return self._buffer.read1(size) |
| 189 |
|
| 190 |
def readinto(self, b): |
| 191 |
"""Read bytes into b. |
| 192 |
|
| 193 |
Returns the number of bytes read (0 for EOF). |
| 194 |
""" |
| 195 |
self._check_can_read() |
| 196 |
return self._buffer.readinto(b) |
| 197 |
|
| 198 |
def readinto1(self, b): |
| 199 |
"""Read bytes into b, while trying to avoid making multiple reads |
| 200 |
from the underlying stream. |
| 201 |
|
| 202 |
Returns the number of bytes read (0 for EOF). |
| 203 |
""" |
| 204 |
self._check_can_read() |
| 205 |
return self._buffer.readinto1(b) |
| 206 |
|
| 207 |
def readline(self, size=-1): |
| 208 |
"""Read a line of uncompressed bytes from the file. |
| 209 |
|
| 210 |
The terminating newline (if present) is retained. If size is |
| 211 |
non-negative, no more than size bytes will be read (in which |
| 212 |
case the line may be incomplete). Returns b'' if already at EOF. |
| 213 |
""" |
| 214 |
self._check_can_read() |
| 215 |
return self._buffer.readline(size) |
| 216 |
|
| 217 |
def seek(self, offset, whence=io.SEEK_SET): |
| 218 |
"""Change the file position. |
| 219 |
|
| 220 |
The new position is specified by offset, relative to the |
| 221 |
position indicated by whence. Possible values for whence are: |
| 222 |
|
| 223 |
0: start of stream (default): offset must not be negative |
| 224 |
1: current stream position |
| 225 |
2: end of stream; offset must not be positive |
| 226 |
|
| 227 |
Returns the new file position. |
| 228 |
|
| 229 |
Note that seeking is emulated, so depending on the arguments, |
| 230 |
this operation may be extremely slow. |
| 231 |
""" |
| 232 |
self._check_can_read() |
| 233 |
|
| 234 |
# BufferedReader.seek() checks seekable |
| 235 |
return self._buffer.seek(offset, whence) |
| 236 |
|
| 237 |
def peek(self, size=-1): |
| 238 |
"""Return buffered data without advancing the file position. |
| 239 |
|
| 240 |
Always returns at least one byte of data, unless at EOF. |
| 241 |
The exact number of bytes returned is unspecified. |
| 242 |
""" |
| 243 |
# Relies on the undocumented fact that BufferedReader.peek() always |
| 244 |
# returns at least one byte (except at EOF) |
| 245 |
self._check_can_read() |
| 246 |
return self._buffer.peek(size) |
| 247 |
|
| 248 |
def __next__(self): |
| 249 |
if ret := self._buffer.readline(): |
| 250 |
return ret |
| 251 |
raise StopIteration |
| 252 |
|
| 253 |
def tell(self): |
| 254 |
"""Return the current file position.""" |
| 255 |
self._check_not_closed() |
| 256 |
if self._mode == _MODE_READ: |
| 257 |
return self._buffer.tell() |
| 258 |
elif self._mode == _MODE_WRITE: |
| 259 |
return self._pos |
| 260 |
|
| 261 |
def fileno(self): |
| 262 |
"""Return the file descriptor for the underlying file.""" |
| 263 |
self._check_not_closed() |
| 264 |
return self._fp.fileno() |
| 265 |
|
| 266 |
@property |
| 267 |
def name(self): |
| 268 |
self._check_not_closed() |
| 269 |
return self._fp.name |
| 270 |
|
| 271 |
@property |
| 272 |
def mode(self): |
| 273 |
return 'wb' if self._mode == _MODE_WRITE else 'rb' |
| 274 |
|
| 275 |
@property |
| 276 |
def closed(self): |
| 277 |
"""True if this file is closed.""" |
| 278 |
return self._mode == _MODE_CLOSED |
| 279 |
|
| 280 |
def seekable(self): |
| 281 |
"""Return whether the file supports seeking.""" |
| 282 |
return self.readable() and self._buffer.seekable() |
| 283 |
|
| 284 |
def readable(self): |
| 285 |
"""Return whether the file was opened for reading.""" |
| 286 |
self._check_not_closed() |
| 287 |
return self._mode == _MODE_READ |
| 288 |
|
| 289 |
def writable(self): |
| 290 |
"""Return whether the file was opened for writing.""" |
| 291 |
self._check_not_closed() |
| 292 |
return self._mode == _MODE_WRITE |
| 293 |
|
| 294 |
|
| 295 |
def open(file, /, mode='rb', *, level=None, options=None, zstd_dict=None, |
| 296 |
encoding=None, errors=None, newline=None): |
| 297 |
"""Open a Zstandard compressed file in binary or text mode. |
| 298 |
|
| 299 |
file can be either a file name (given as a str, bytes, or PathLike object), |
| 300 |
in which case the named file is opened, or it can be an existing file object |
| 301 |
to read from or write to. |
| 302 |
|
| 303 |
The mode parameter can be 'r', 'rb' (default), 'w', 'wb', 'x', 'xb', 'a', |
| 304 |
'ab' for binary mode, or 'rt', 'wt', 'xt', 'at' for text mode. |
| 305 |
|
| 306 |
The level, options, and zstd_dict parameters specify the settings the same |
| 307 |
as ZstdFile. |
| 308 |
|
| 309 |
When using read mode (decompression), the options parameter is a dict |
| 310 |
representing advanced decompression options. The level parameter is not |
| 311 |
supported in this case. When using write mode (compression), only one of |
| 312 |
level, an int representing the compression level, or options, a dict |
| 313 |
representing advanced compression options, may be passed. In both modes, |
| 314 |
zstd_dict is a ZstdDict instance containing a trained Zstandard dictionary. |
| 315 |
|
| 316 |
For binary mode, this function is equivalent to the ZstdFile constructor: |
| 317 |
ZstdFile(filename, mode, ...). In this case, the encoding, errors and |
| 318 |
newline parameters must not be provided. |
| 319 |
|
| 320 |
For text mode, an ZstdFile object is created, and wrapped in an |
| 321 |
io.TextIOWrapper instance with the specified encoding, error handling |
| 322 |
behavior, and line ending(s). |
| 323 |
""" |
| 324 |
|
| 325 |
text_mode = 't' in mode |
| 326 |
mode = mode.replace('t', '') |
| 327 |
|
| 328 |
if text_mode: |
| 329 |
if 'b' in mode: |
| 330 |
raise ValueError(f'Invalid mode: {mode!r}') |
| 331 |
else: |
| 332 |
if encoding is not None: |
| 333 |
raise ValueError('Argument "encoding" not supported in binary mode') |
| 334 |
if errors is not None: |
| 335 |
raise ValueError('Argument "errors" not supported in binary mode') |
| 336 |
if newline is not None: |
| 337 |
raise ValueError('Argument "newline" not supported in binary mode') |
| 338 |
|
| 339 |
binary_file = ZstdFile(file, mode, level=level, options=options, |
| 340 |
zstd_dict=zstd_dict) |
| 341 |
|
| 342 |
if text_mode: |
| 343 |
return io.TextIOWrapper(binary_file, encoding, errors, newline) |
| 344 |
else: |
| 345 |
return binary_file |