Technetra

CodeMagic: serialize_with_cache - an approach to serializing and de-serializing arbitrary data in bash using an MRU cache

Code Guru,  November 9th, 2008 at 10:49 pm

The program serialize_with_cache demonstrates an approach to serializing and de-serializing arbitrary data in bash. This version explores the use of a cache to accelerate conversion of characters in the serialization string. The cache improves serialization performance on typical text by about 25%. However, performance is still quite slow so that string serialization using native bash mechanisms remains infeasible for anything but small datasets.

After making it executable, a simple way to run the script is by invoking:

./serialize_with_cache.sh

serialize_with_cache then runs a small self-test harness. The test exercises several small string serializations and de-serializations and then times a cycle of deflating and re-inflating the contents of a common system file (read only, of course) as well as measuring the performance speed-up due to cache effects.

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
#!/bin/bash
#
# serialize_with_cache demonstrates an approach to serializing
# and de-serializing arbitrary data in bash.  This version explores
# the use of a cache to accelerate conversion of characters in the
# serialization string. The cache improves serialization performance
# on typical text by about 25%. However, performance is still quite
# slow so that string serialization using native bash mechanisms
# remains infeasible for anything but small datasets. 
#
# A simple way to run the script is by invoking:
#
#     ./serialize_with_cache.sh
#
# serialize_with_cache then runs a small self-test harness. The test
# exercises several small string serializations and de-serializations
# and then times a cycle of deflating and re-inflating the contents of
# a common system file (read only, of course) as well as measuring the
# performance speed-up due to cache effects.
# 
# Copyright (c) 2008 Technetra Corp
# 
# Permission is hereby granted, free of charge, to any person
# obtaining a copy of this software and associated documentation
# files (the "Software"), to deal in the Software without
# restriction, including without limitation the rights to use,
# copy, modify, merge, publish, distribute, sublicense, and/or sell
# copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following
# conditions:
# 
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
# 
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
# EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES
# OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
# NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
# HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
# WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
# OTHER DEALINGS IN THE SOFTWARE.
#
PASSED=0
TOTAL=0
I=1
MRU[0]="za61"
CACHE_SIZE=200
find_char_no_cache() {
  #sleep 1
  local hex_result=$1
  local in_char=$2
  local hex_digits1 hex_digits2=(0 1 2 3 4 5 6 7 8 9 A B C D E F)
  if [[ "$in_char" > $'\x6f' ]]; then
    hex_digits1=(7 8 9 A B C D E F)
  elif [[ "$in_char" > $'\x5f' ]]; then
    hex_digits1=(6)
  elif [[ "$in_char" > $'\x4f' ]]; then
    hex_digits1=(5)
  elif [[ "$in_char" > $'\x3f' ]]; then
    hex_digits1=(4)
  elif [[ "$in_char" > $'\x2f' ]]; then
    hex_digits1=(3)
  elif [[ "$in_char" > $'\x1f' ]]; then
    hex_digits1=(2)
  else
    hex_digits1=(0 1)
  fi
  local finished=0
  local c t targ hxi hxj
  for hxi in ${hex_digits1[@]}; do
    for hxj in ${hex_digits2[@]}; do
      c=$hxi$hxj
      #t="`eval echo z\$\'\\\x$c\'z`"
      t="`echo -e "z\x${c}z"`"
      targ=${t:1:1}
      if [[ "$in_char" == "$targ" ]]; then
        #echo "matched character index $c"
        finished=1
        break
      fi
    done
    [[ $finished -eq 1 ]] && break
  done
  eval $hex_result=$c
} 
find_char_with_cache() {
  #sleep 1
  local hex_result=$1
  local in_char=$2
  local j=$(((I+(CACHE_SIZE-1))%CACHE_SIZE))
  while [[ j -ne I%CACHE_SIZE ]]; do
    m="${MRU[j]}"
    if [[ "$m" ]]; then
      j=$(((j+(CACHE_SIZE-1))%CACHE_SIZE))
      if [[ "${m:1:1}" == "$in_char" ]]; then
        eval $hex_result=${m:2:2}
        #echo "got cache hit \"$in_char\""
        I=$((I%CACHE_SIZE))
        MRU[I++]="`echo -e "z$in_char${m:2:2}"`"
        return
      fi 
    else
      #echo "missed cache for \"$in_char\""
      break
    fi
    #sleep 1
  done 
  local hex_digits1 hex_digits2=(0 1 2 3 4 5 6 7 8 9 A B C D E F)
  if [[ "$in_char" > $'\x6f' ]]; then
    hex_digits1=(7 8 9 A B C D E F)
  elif [[ "$in_char" > $'\x5f' ]]; then
    hex_digits1=(6)
  elif [[ "$in_char" > $'\x4f' ]]; then
    hex_digits1=(5)
  elif [[ "$in_char" > $'\x3f' ]]; then
    hex_digits1=(4)
  elif [[ "$in_char" > $'\x2f' ]]; then
    hex_digits1=(3)
  elif [[ "$in_char" > $'\x1f' ]]; then
    hex_digits1=(2)
  else
    hex_digits1=(0 1)
  fi
  local finished=0
  local c t targ hxi hxj
  for hxi in ${hex_digits1[@]}; do
    for hxj in ${hex_digits2[@]}; do
      c=$hxi$hxj
      #t="`eval echo z\$\'\\\x$c\'z`"
      t="`echo -e "z\x${c}z"`"
      targ=${t:1:1}
      if [[ "$in_char" == "$targ" ]]; then
        #echo "matched character index $c"
        finished=1
        I=$((I%CACHE_SIZE))
        MRU[I++]="`echo -e "z$targ$c"`"
        break
      fi
    done
    [[ $finished -eq 1 ]] && break
  done
  eval $hex_result=$c
}
deflate() {
  local result=$1
  local s="$2"
  local which_find_char=$3
  local i=0 deflated out_hex
  local c="${s:i++:1}"
  while [[ "$c" ]]; do
    $which_find_char out_hex "$c"
    deflated=$deflated$out_hex
    c="${s:i++:1}"
  done 
  eval $result=$deflated
}
inflate() {
  local result=$1
  local hex_string=$2 
  local i=0 inflated t targ
  local h=${hex_string:i:2}
  i=$((i+2))
  while [[ "$h" ]]; do
    #t="`eval echo z\$\'\\\x$h\'z`"
    t="`echo -e "z\x${h}z"`"
    targ=${t:1:1}
    [[ "$targ" == '"' ]] && targ='\"'
    inflated="$inflated$targ"
    h=${hex_string:i:2}
    i=$((i+2))
  done
  eval "$result=\"$inflated\""
}
test_serialize () {
  local in_orig="$1"
  local which_find_char=${2:-find_char_with_cache}
  local stat hex copy
  deflate hex "$in_orig" $which_find_char
  inflate copy $hex
  if [[ "$in_orig" == "$copy" ]]; then
    stat=passed
    ((PASSED++))
  else
    stat=failed
  fi
  ((TOTAL++))
  echo "$stat test $TOTAL (cache=$which_find_char), hex=$hex, original=\"$in_orig\""
}
time_file () {
  echo "timing serialization for file $1..."
  test_serialize "$(cat $1)" $2
}
###
### MAIN
###
test_serialize "a\"c"
test_serialize "a\'c"
test_serialize "$(echo -e " a\n\tb")"
test_serialize "'%' in a prerequisite of a pattern rule stands for the same stem that was matched by the"
time_file /etc/passwd
echo "checking performance improvement due to cache..."
# built-in bash reserved word, time, requires execution in a subshell to capture its stderr
# output
export TIMEFORMAT=%3R
t1=`(time test_serialize "$(head -n 25 /etc/passwd)" >/dev/null) 2>&1`
echo "continuing..."
t2=`(time test_serialize "$(head -n 25 /etc/passwd)" find_char_no_cache >/dev/null) 2>&1`
# use integer arithmetic and add leading 1 to guard against misinterpretation
# as octal because of leading zeros
l1=${#t1}
l2=${#t2}
max=$((l1>l2?l1:l2))
pad=100000
pad=${pad:0:max}
z1=${pad:0:max-l1+1}${t1/./}
z2=${pad:0:max-l2+1}${t2/./}
echo "TEST SUMMARY: passed $PASSED of $TOTAL tests" \
     "(cache speed-up approx: $((((z2-z1)*100)/(z2-pad)))%)"
# end serialize_with_cache

Copyright © 2008 Technetra. This code is covered by the MIT license. You can follow any responses to this entry through the RSS 2.0 feed. You can skip to the end and leave a response. Pings are currently closed.

CodeMagic: rename.rb - a Ruby version of Perl’s rename command Article Index CodeMagic: serialize - an approach to serializing and de-serializing arbitrary data entirely within bash

Comments

Be the first to post a comment.

Add a comment

Add your comments
  1. (required)
  2. (valid email required)
  3. (required)
  4. Captcha
 

cforms contact form by delicious:days

© 2000-2009 Technetra. All rights reserved. Contact | Terms of Use

WordPress