| 1 | n/a | """Internal classes used by the gzip, lzma and bz2 modules""" |
|---|
| 2 | n/a | |
|---|
| 3 | n/a | import io |
|---|
| 4 | n/a | |
|---|
| 5 | n/a | |
|---|
| 6 | n/a | BUFFER_SIZE = io.DEFAULT_BUFFER_SIZE # Compressed data read chunk size |
|---|
| 7 | n/a | |
|---|
| 8 | n/a | |
|---|
| 9 | n/a | class BaseStream(io.BufferedIOBase): |
|---|
| 10 | n/a | """Mode-checking helper functions.""" |
|---|
| 11 | n/a | |
|---|
| 12 | n/a | def _check_not_closed(self): |
|---|
| 13 | n/a | if self.closed: |
|---|
| 14 | n/a | raise ValueError("I/O operation on closed file") |
|---|
| 15 | n/a | |
|---|
| 16 | n/a | def _check_can_read(self): |
|---|
| 17 | n/a | if not self.readable(): |
|---|
| 18 | n/a | raise io.UnsupportedOperation("File not open for reading") |
|---|
| 19 | n/a | |
|---|
| 20 | n/a | def _check_can_write(self): |
|---|
| 21 | n/a | if not self.writable(): |
|---|
| 22 | n/a | raise io.UnsupportedOperation("File not open for writing") |
|---|
| 23 | n/a | |
|---|
| 24 | n/a | def _check_can_seek(self): |
|---|
| 25 | n/a | if not self.readable(): |
|---|
| 26 | n/a | raise io.UnsupportedOperation("Seeking is only supported " |
|---|
| 27 | n/a | "on files open for reading") |
|---|
| 28 | n/a | if not self.seekable(): |
|---|
| 29 | n/a | raise io.UnsupportedOperation("The underlying file object " |
|---|
| 30 | n/a | "does not support seeking") |
|---|
| 31 | n/a | |
|---|
| 32 | n/a | |
|---|
| 33 | n/a | class DecompressReader(io.RawIOBase): |
|---|
| 34 | n/a | """Adapts the decompressor API to a RawIOBase reader API""" |
|---|
| 35 | n/a | |
|---|
| 36 | n/a | def readable(self): |
|---|
| 37 | n/a | return True |
|---|
| 38 | n/a | |
|---|
| 39 | n/a | def __init__(self, fp, decomp_factory, trailing_error=(), **decomp_args): |
|---|
| 40 | n/a | self._fp = fp |
|---|
| 41 | n/a | self._eof = False |
|---|
| 42 | n/a | self._pos = 0 # Current offset in decompressed stream |
|---|
| 43 | n/a | |
|---|
| 44 | n/a | # Set to size of decompressed stream once it is known, for SEEK_END |
|---|
| 45 | n/a | self._size = -1 |
|---|
| 46 | n/a | |
|---|
| 47 | n/a | # Save the decompressor factory and arguments. |
|---|
| 48 | n/a | # If the file contains multiple compressed streams, each |
|---|
| 49 | n/a | # stream will need a separate decompressor object. A new decompressor |
|---|
| 50 | n/a | # object is also needed when implementing a backwards seek(). |
|---|
| 51 | n/a | self._decomp_factory = decomp_factory |
|---|
| 52 | n/a | self._decomp_args = decomp_args |
|---|
| 53 | n/a | self._decompressor = self._decomp_factory(**self._decomp_args) |
|---|
| 54 | n/a | |
|---|
| 55 | n/a | # Exception class to catch from decompressor signifying invalid |
|---|
| 56 | n/a | # trailing data to ignore |
|---|
| 57 | n/a | self._trailing_error = trailing_error |
|---|
| 58 | n/a | |
|---|
| 59 | n/a | def close(self): |
|---|
| 60 | n/a | self._decompressor = None |
|---|
| 61 | n/a | return super().close() |
|---|
| 62 | n/a | |
|---|
| 63 | n/a | def seekable(self): |
|---|
| 64 | n/a | return self._fp.seekable() |
|---|
| 65 | n/a | |
|---|
| 66 | n/a | def readinto(self, b): |
|---|
| 67 | n/a | with memoryview(b) as view, view.cast("B") as byte_view: |
|---|
| 68 | n/a | data = self.read(len(byte_view)) |
|---|
| 69 | n/a | byte_view[:len(data)] = data |
|---|
| 70 | n/a | return len(data) |
|---|
| 71 | n/a | |
|---|
| 72 | n/a | def read(self, size=-1): |
|---|
| 73 | n/a | if size < 0: |
|---|
| 74 | n/a | return self.readall() |
|---|
| 75 | n/a | |
|---|
| 76 | n/a | if not size or self._eof: |
|---|
| 77 | n/a | return b"" |
|---|
| 78 | n/a | data = None # Default if EOF is encountered |
|---|
| 79 | n/a | # Depending on the input data, our call to the decompressor may not |
|---|
| 80 | n/a | # return any data. In this case, try again after reading another block. |
|---|
| 81 | n/a | while True: |
|---|
| 82 | n/a | if self._decompressor.eof: |
|---|
| 83 | n/a | rawblock = (self._decompressor.unused_data or |
|---|
| 84 | n/a | self._fp.read(BUFFER_SIZE)) |
|---|
| 85 | n/a | if not rawblock: |
|---|
| 86 | n/a | break |
|---|
| 87 | n/a | # Continue to next stream. |
|---|
| 88 | n/a | self._decompressor = self._decomp_factory( |
|---|
| 89 | n/a | **self._decomp_args) |
|---|
| 90 | n/a | try: |
|---|
| 91 | n/a | data = self._decompressor.decompress(rawblock, size) |
|---|
| 92 | n/a | except self._trailing_error: |
|---|
| 93 | n/a | # Trailing data isn't a valid compressed stream; ignore it. |
|---|
| 94 | n/a | break |
|---|
| 95 | n/a | else: |
|---|
| 96 | n/a | if self._decompressor.needs_input: |
|---|
| 97 | n/a | rawblock = self._fp.read(BUFFER_SIZE) |
|---|
| 98 | n/a | if not rawblock: |
|---|
| 99 | n/a | raise EOFError("Compressed file ended before the " |
|---|
| 100 | n/a | "end-of-stream marker was reached") |
|---|
| 101 | n/a | else: |
|---|
| 102 | n/a | rawblock = b"" |
|---|
| 103 | n/a | data = self._decompressor.decompress(rawblock, size) |
|---|
| 104 | n/a | if data: |
|---|
| 105 | n/a | break |
|---|
| 106 | n/a | if not data: |
|---|
| 107 | n/a | self._eof = True |
|---|
| 108 | n/a | self._size = self._pos |
|---|
| 109 | n/a | return b"" |
|---|
| 110 | n/a | self._pos += len(data) |
|---|
| 111 | n/a | return data |
|---|
| 112 | n/a | |
|---|
| 113 | n/a | # Rewind the file to the beginning of the data stream. |
|---|
| 114 | n/a | def _rewind(self): |
|---|
| 115 | n/a | self._fp.seek(0) |
|---|
| 116 | n/a | self._eof = False |
|---|
| 117 | n/a | self._pos = 0 |
|---|
| 118 | n/a | self._decompressor = self._decomp_factory(**self._decomp_args) |
|---|
| 119 | n/a | |
|---|
| 120 | n/a | def seek(self, offset, whence=io.SEEK_SET): |
|---|
| 121 | n/a | # Recalculate offset as an absolute file position. |
|---|
| 122 | n/a | if whence == io.SEEK_SET: |
|---|
| 123 | n/a | pass |
|---|
| 124 | n/a | elif whence == io.SEEK_CUR: |
|---|
| 125 | n/a | offset = self._pos + offset |
|---|
| 126 | n/a | elif whence == io.SEEK_END: |
|---|
| 127 | n/a | # Seeking relative to EOF - we need to know the file's size. |
|---|
| 128 | n/a | if self._size < 0: |
|---|
| 129 | n/a | while self.read(io.DEFAULT_BUFFER_SIZE): |
|---|
| 130 | n/a | pass |
|---|
| 131 | n/a | offset = self._size + offset |
|---|
| 132 | n/a | else: |
|---|
| 133 | n/a | raise ValueError("Invalid value for whence: {}".format(whence)) |
|---|
| 134 | n/a | |
|---|
| 135 | n/a | # Make it so that offset is the number of bytes to skip forward. |
|---|
| 136 | n/a | if offset < self._pos: |
|---|
| 137 | n/a | self._rewind() |
|---|
| 138 | n/a | else: |
|---|
| 139 | n/a | offset -= self._pos |
|---|
| 140 | n/a | |
|---|
| 141 | n/a | # Read and discard data until we reach the desired position. |
|---|
| 142 | n/a | while offset > 0: |
|---|
| 143 | n/a | data = self.read(min(io.DEFAULT_BUFFER_SIZE, offset)) |
|---|
| 144 | n/a | if not data: |
|---|
| 145 | n/a | break |
|---|
| 146 | n/a | offset -= len(data) |
|---|
| 147 | n/a | |
|---|
| 148 | n/a | return self._pos |
|---|
| 149 | n/a | |
|---|
| 150 | n/a | def tell(self): |
|---|
| 151 | n/a | """Return the current file position.""" |
|---|
| 152 | n/a | return self._pos |
|---|