1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
|
import bz2
import struct
import typing
import zlib
import lzma
try:
import zstandard
except ImportError:
zstandard = None
class Decompressor(typing.Protocol):
def copy(self) -> "Decompressor":
...
def decompress(self, data: bytes) -> bytes:
...
def flush(self) -> bytes:
...
@property
def unused_data(self) -> bytes:
...
class GzipDecompressor:
"""An interface to gzip which is similar to bz2.BZ2Decompressor and
lzma.LZMADecompressor."""
def __init__(self) -> None:
self.sawheader = False
self.inbuffer = b""
self.decompressor: typing.Optional[Decompressor] = None
self.crc = 0
self.size = 0
def decompress(self, data: bytes) -> bytes:
"""
@raises ValueError: if no gzip magic is found
@raises zlib.error: from zlib invocations
"""
while True:
if self.decompressor:
data = self.decompressor.decompress(data)
self.crc = zlib.crc32(data, self.crc)
self.size += len(data)
unused_data = self.decompressor.unused_data
if not unused_data:
return data
self.decompressor = None
return data + self.decompress(unused_data)
self.inbuffer += data
skip = 10
if len(self.inbuffer) < skip:
return b""
if not self.inbuffer.startswith(b"\037\213\010"):
raise ValueError("gzip magic not found")
flag = ord(self.inbuffer[3:4])
if flag & 4:
if len(self.inbuffer) < skip + 2:
return b""
length, = struct.unpack("<H", self.inbuffer[skip:skip+2])
skip += 2 + length
for field in (8, 16):
if flag & field:
length = self.inbuffer.find(b"\0", skip)
if length < 0:
return b""
skip = length + 1
if flag & 2:
skip += 2
if len(self.inbuffer) < skip:
return b""
data = self.inbuffer[skip:]
self.inbuffer = b""
self.sawheader = True
self.decompressor = zlib.decompressobj(-zlib.MAX_WBITS)
@property
def unused_data(self) -> bytes:
if self.decompressor:
return self.decompressor.unused_data
elif not self.sawheader:
return self.inbuffer
else:
expect = struct.pack("<LL", self.crc, self.size)
if self.inbuffer.startswith(expect) and \
self.inbuffer[len(expect):].replace(b"\0", b"") == b"":
return b""
return self.inbuffer
def flush(self) -> bytes:
"""
@raises zlib.error: from zlib invocations
"""
if not self.decompressor:
return b""
return self.decompressor.flush()
def copy(self) -> "GzipDecompressor":
new = GzipDecompressor()
new.inbuffer = self.inbuffer
if self.decompressor:
new.decompressor = self.decompressor.copy()
new.sawheader = self.sawheader
new.crc = self.crc
new.size = self.size
return new
class DecompressedStream:
"""Turn a readable file-like into a decompressed file-like. It supports
read(optional length), tell, seek(forward only) and close."""
blocksize = 65536
def __init__(
self, fileobj: typing.BinaryIO, decompressor: Decompressor
) -> None:
"""
@param fileobj: a file-like object providing read(size)
@param decompressor: a bz2.BZ2Decompressor or lzma.LZMADecompressor
like object providing methods decompress and flush and an
attribute unused_data
"""
self.fileobj: typing.Optional[typing.BinaryIO] = fileobj
self.decompressor: typing.Optional[Decompressor] = decompressor
self.buff = bytearray()
self.pos = 0
def _fill_buff_until(
self, predicate: typing.Callable[[bytes], bool]
) -> None:
assert self.fileobj is not None
assert self.decompressor is not None
while not predicate(self.buff):
data = self.fileobj.read(self.blocksize)
if data:
self.buff += self.decompressor.decompress(data)
else:
if hasattr(self.decompressor, "flush"):
self.buff += self.decompressor.flush()
break
def _read_from_buff(self, length: int) -> bytes:
ret = bytes(self.buff[:length])
self.buff[:length] = b""
self.pos += length
return ret
def read(self, length: typing.Optional[int] = None) -> bytes:
if length is None:
self._fill_buff_until(lambda _: False)
length = len(self.buff)
else:
self._fill_buff_until(lambda b, l=length: len(b) >= l)
return self._read_from_buff(length)
def readline(self) -> bytes:
self._fill_buff_until(lambda b: b'\n' in b)
try:
length = self.buff.index(b'\n') + 1
except ValueError:
length = len(self.buff)
return self._read_from_buff(length)
def __iter__(self) -> typing.Iterator[bytes]:
return iter(self.readline, b'')
def tell(self) -> int:
assert self.fileobj is not None
return self.pos
def seek(self, pos: int) -> None:
"""Forward seeks by absolute position only."""
assert self.fileobj is not None
if pos < self.pos:
raise ValueError("negative seek not allowed on decompressed stream")
while True:
left = pos - self.pos
# Reading self.buff entirely avoids string concatenation.
size = len(self.buff) or self.blocksize
if left > size:
self.read(size)
else:
self.read(left)
return
def close(self) -> None:
if self.fileobj is not None:
self.fileobj.close()
self.fileobj = None
self.decompressor = None
self.buff = bytearray()
decompressors = {
'.gz': GzipDecompressor,
'.bz2': bz2.BZ2Decompressor,
'.lzma': lzma.LZMADecompressor,
'.xz': lzma.LZMADecompressor,
}
if zstandard is not None:
decompressors[".zst"] = zstandard.ZstdDecompressor
def decompress(filelike: typing.BinaryIO, extension: str) -> typing.BinaryIO:
"""Decompress a stream according to its extension.
@param filelike: is a read-only byte-stream. It must support read(size) and
close().
@param extension: permitted values are "", ".gz", ".bz2", ".lzma", and
".xz"
@returns: a read-only byte-stream with the decompressed contents of the
original filelike. It supports read(size) and close(). If the
original supports seek(pos) and tell(), then it also supports
those.
@raises ValueError: on unkown extensions
"""
if not extension:
return filelike
try:
decompressor = decompressors[extension]
except KeyError:
raise ValueError("unknown compression format with extension %r" %
extension)
return DecompressedStream(filelike, decompressor())
|
