1
# Copyright (C) 2008 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
18
"""ChunkWriter: write compressed data out with a fixed upper bound."""
21
from zlib import Z_FINISH, Z_SYNC_FLUSH
24
class ChunkWriter(object):
25
"""ChunkWriter allows writing of compressed data with a fixed size.
27
If less data is supplied than fills a chunk, the chunk is padded with
28
NULL bytes. If more data is supplied, then the writer packs as much
29
in as it can, but never splits any item it was given.
31
The algorithm for packing is open to improvement! Current it is:
32
- write the bytes given
33
- if the total seen bytes so far exceeds the chunk size, flush.
35
:cvar _max_repack: To fit the maximum number of entries into a node, we
36
will sometimes start over and compress the whole list to get tighter
37
packing. We get diminishing returns after a while, so this limits the
38
number of times we will try.
39
The default is to try to avoid recompressing entirely, but setting this
40
to something like 20 will give maximum compression.
42
:cvar _max_zsync: Another tunable nob. If _max_repack is set to 0, then you
43
can limit the number of times we will try to pack more data into a
44
node. This allows us to do a single compression pass, rather than
45
trying until we overflow, and then recompressing again.
47
# In testing, some values for bzr.dev::
48
# repack time MB max full
55
# zsync time MB repack stop_for_z
69
# In testing, some values for mysql-unpacked::
71
# repack time MB full stop_for_repack
77
# 0 29.5 116.5 0 29782
84
# 10 29.4 18.6 195 4526
85
# 11 29.2 18.0 421 4143
86
# 12 28.0 17.5 702 3738
87
# 15 28.9 16.5 1223 2969
88
# 20 29.6 15.7 2182 1810
89
# 30 31.4 15.4 3891 23
91
# Tuple of (num_repack_attempts, num_zsync_attempts)
92
# num_zsync_attempts only has meaning if num_repack_attempts is 0.
93
_repack_opts_for_speed = (0, 8)
94
_repack_opts_for_size = (20, 0)
96
def __init__(self, chunk_size, reserved=0, optimize_for_size=False):
97
"""Create a ChunkWriter to write chunk_size chunks.
99
:param chunk_size: The total byte count to emit at the end of the
101
:param reserved: How many bytes to allow for reserved data. reserved
102
data space can only be written to via the write(...,
105
self.chunk_size = chunk_size
106
self.compressor = zlib.compressobj()
109
self.bytes_out_len = 0
110
# bytes that have been seen, but not included in a flush to out yet
111
self.unflushed_in_bytes = 0
114
self.unused_bytes = None
115
self.reserved_size = reserved
116
# Default is to make building fast rather than compact
117
self.set_optimize(for_size=optimize_for_size)
122
This returns the final compressed chunk, and either None, or the
123
bytes that did not fit in the chunk.
125
:return: (compressed_bytes, unused_bytes, num_nulls_needed)
127
* compressed_bytes: a list of bytes that were output from the
128
compressor. If the compressed length was not exactly chunk_size,
129
the final string will be a string of all null bytes to pad this
131
* unused_bytes: None, or the last bytes that were added, which we
133
* num_nulls_needed: How many nulls are padded at the end
135
self.bytes_in = None # Free the data cached so far, we don't need it
136
out = self.compressor.flush(Z_FINISH)
137
self.bytes_list.append(out)
138
self.bytes_out_len += len(out)
140
if self.bytes_out_len > self.chunk_size:
141
raise AssertionError('Somehow we ended up with too much'
142
' compressed data, %d > %d'
143
% (self.bytes_out_len, self.chunk_size))
144
nulls_needed = self.chunk_size - self.bytes_out_len
146
self.bytes_list.append(b"\x00" * nulls_needed)
147
return self.bytes_list, self.unused_bytes, nulls_needed
149
def set_optimize(self, for_size=True):
150
"""Change how we optimize our writes.
152
:param for_size: If True, optimize for minimum space usage, otherwise
153
optimize for fastest writing speed.
157
opts = ChunkWriter._repack_opts_for_size
159
opts = ChunkWriter._repack_opts_for_speed
160
self._max_repack, self._max_zsync = opts
162
def _recompress_all_bytes_in(self, extra_bytes=None):
163
"""Recompress the current bytes_in, and optionally more.
165
:param extra_bytes: Optional, if supplied we will add it with
167
:return: (bytes_out, bytes_out_len, alt_compressed)
169
* bytes_out: is the compressed bytes returned from the compressor
170
* bytes_out_len: the length of the compressed output
171
* compressor: An object with everything packed in so far, and
174
compressor = zlib.compressobj()
176
append = bytes_out.append
177
compress = compressor.compress
178
for accepted_bytes in self.bytes_in:
179
out = compress(accepted_bytes)
183
out = compress(extra_bytes)
184
out += compressor.flush(Z_SYNC_FLUSH)
186
bytes_out_len = sum(map(len, bytes_out))
187
return bytes_out, bytes_out_len, compressor
189
def write(self, bytes, reserved=False):
190
"""Write some bytes to the chunk.
192
If the bytes fit, False is returned. Otherwise True is returned
193
and the bytes have not been added to the chunk.
195
:param bytes: The bytes to include
196
:param reserved: If True, we can use the space reserved in the
199
if self.num_repack > self._max_repack and not reserved:
200
self.unused_bytes = bytes
203
capacity = self.chunk_size
205
capacity = self.chunk_size - self.reserved_size
206
comp = self.compressor
208
# Check to see if the currently unflushed bytes would fit with a bit of
209
# room to spare, assuming no compression.
210
next_unflushed = self.unflushed_in_bytes + len(bytes)
211
remaining_capacity = capacity - self.bytes_out_len - 10
212
if (next_unflushed < remaining_capacity):
213
# looks like it will fit
214
out = comp.compress(bytes)
216
self.bytes_list.append(out)
217
self.bytes_out_len += len(out)
218
self.bytes_in.append(bytes)
219
self.unflushed_in_bytes += len(bytes)
221
# This may or may not fit, try to add it with Z_SYNC_FLUSH
222
# Note: It is tempting to do this as a look-ahead pass, and to
223
# 'copy()' the compressor before flushing. However, it seems
224
# that Which means that it is the same thing as increasing
225
# repack, similar cost, same benefit. And this way we still
226
# have the 'repack' knob that can be adjusted, and not depend
227
# on a platform-specific 'copy()' function.
229
if self._max_repack == 0 and self.num_zsync > self._max_zsync:
231
self.unused_bytes = bytes
233
out = comp.compress(bytes)
234
out += comp.flush(Z_SYNC_FLUSH)
235
self.unflushed_in_bytes = 0
237
self.bytes_list.append(out)
238
self.bytes_out_len += len(out)
240
# We are a bit extra conservative, because it seems that you *can*
241
# get better compression with Z_SYNC_FLUSH than a full compress. It
242
# is probably very rare, but we were able to trigger it.
243
if self.num_repack == 0:
247
if self.bytes_out_len + safety_margin <= capacity:
248
# It fit, so mark it added
249
self.bytes_in.append(bytes)
251
# We are over budget, try to squeeze this in without any
254
(bytes_out, this_len,
255
compressor) = self._recompress_all_bytes_in(bytes)
256
if self.num_repack >= self._max_repack:
257
# When we get *to* _max_repack, bump over so that the
258
# earlier > _max_repack will be triggered.
260
if this_len + 10 > capacity:
261
(bytes_out, this_len,
262
compressor) = self._recompress_all_bytes_in()
263
self.compressor = compressor
264
# Force us to not allow more data
265
self.num_repack = self._max_repack + 1
266
self.bytes_list = bytes_out
267
self.bytes_out_len = this_len
268
self.unused_bytes = bytes
271
# This fits when we pack it tighter, so use the new packing
272
self.compressor = compressor
273
self.bytes_in.append(bytes)
274
self.bytes_list = bytes_out
275
self.bytes_out_len = this_len