| 1 | #!/bin/csh -f
|
|---|
| 2 | ###########################################################################
|
|---|
| 3 | #
|
|---|
| 4 | # mgbuild.sh -- Script used to build mg text collection.
|
|---|
| 5 | # Copyright (C) 1994 Neil Sharman
|
|---|
| 6 | #
|
|---|
| 7 | # This program is free software; you can redistribute it and/or modify
|
|---|
| 8 | # it under the terms of the GNU General Public License as published by
|
|---|
| 9 | # the Free Software Foundation; either version 2 of the License, or
|
|---|
| 10 | # (at your option) any later version.
|
|---|
| 11 | #
|
|---|
| 12 | # This program is distributed in the hope that it will be useful,
|
|---|
| 13 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|---|
| 14 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|---|
| 15 | # GNU General Public License for more details.
|
|---|
| 16 | #
|
|---|
| 17 | # You should have received a copy of the GNU General Public License
|
|---|
| 18 | # along with this program; if not, write to the Free Software
|
|---|
| 19 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
|---|
| 20 | #
|
|---|
| 21 | # @(#)mgbuild.sh 1.9 16 Mar 1994
|
|---|
| 22 | #
|
|---|
| 23 | ###########################################################################
|
|---|
| 24 | # mg.build
|
|---|
| 25 | # Creates a Full Text Retrieval database
|
|---|
| 26 | #
|
|---|
| 27 | # @(#)mgbuild.sh 1.9 16 Mar 1994
|
|---|
| 28 | #
|
|---|
| 29 |
|
|---|
| 30 | set complex = ""
|
|---|
| 31 |
|
|---|
| 32 | # Parse the command line arguments
|
|---|
| 33 | while ($#argv >= 1)
|
|---|
| 34 | if ("$1" == "-s") then
|
|---|
| 35 | shift
|
|---|
| 36 | if ($#argv >= 1) then
|
|---|
| 37 | set source = $1
|
|---|
| 38 | shift
|
|---|
| 39 | endif
|
|---|
| 40 | else if ("$1" == "-d") then
|
|---|
| 41 | shift
|
|---|
| 42 | if ($#argv >= 1) then
|
|---|
| 43 | set mgdata = $1
|
|---|
| 44 | shift
|
|---|
| 45 | endif
|
|---|
| 46 | else if ("$1" == "-g") then
|
|---|
| 47 | shift
|
|---|
| 48 | if ($#argv >= 1) then
|
|---|
| 49 | set get = $1
|
|---|
| 50 | shift
|
|---|
| 51 | endif
|
|---|
| 52 | else if ("$1" == "-c") then
|
|---|
| 53 | set complex = "-text"
|
|---|
| 54 | shift
|
|---|
| 55 | else
|
|---|
| 56 | if ($?text == "0") then
|
|---|
| 57 | set text = $1
|
|---|
| 58 | endif
|
|---|
| 59 | shift
|
|---|
| 60 | endif
|
|---|
| 61 |
|
|---|
| 62 | end
|
|---|
| 63 |
|
|---|
| 64 | if ($?text == "0") then
|
|---|
| 65 | set prog = $0
|
|---|
| 66 | echo "USAGE:"
|
|---|
| 67 | echo " "$prog:t" [-s config-script] [-g get-program] [-d mgdata dir] [-c] source"
|
|---|
| 68 | echo ""
|
|---|
| 69 | echo " The config-script is only needed if a non-standard build is required."
|
|---|
| 70 | echo " The get-program defaults to mg_get if not specified."
|
|---|
| 71 | exit 1
|
|---|
| 72 | endif
|
|---|
| 73 |
|
|---|
| 74 | set bindir = $0
|
|---|
| 75 | set bindir = $bindir:h
|
|---|
| 76 |
|
|---|
| 77 | # if $pipe == 1 then pipe in the source text using $get and $text otherwise
|
|---|
| 78 | # read the source text directly from the file names specified in $input_files
|
|---|
| 79 | set pipe = 1
|
|---|
| 80 |
|
|---|
| 81 | if ($?get == "0") then
|
|---|
| 82 | set complex = "-text"
|
|---|
| 83 | set get = $bindir/mg_get
|
|---|
| 84 | endif
|
|---|
| 85 |
|
|---|
| 86 | if ($?mgdata) then
|
|---|
| 87 | setenv MGDATA $mgdata
|
|---|
| 88 | endif
|
|---|
| 89 |
|
|---|
| 90 | if (-e $MGDATA/${text}.chunks) then
|
|---|
| 91 | set input_files = `cat $MGDATA/${text}.chunks`
|
|---|
| 92 | endif
|
|---|
| 93 |
|
|---|
| 94 | # Set the stemming method
|
|---|
| 95 | # Bit 0 = case folding
|
|---|
| 96 | # Bit 1 = S stemmer
|
|---|
| 97 | set stem_method = 3
|
|---|
| 98 |
|
|---|
| 99 |
|
|---|
| 100 | # [RPAP - Jan 97: Stem Index Change]
|
|---|
| 101 | # If do_indexes == 1 then build collection with full indexes to blocked file
|
|---|
| 102 | # overriding stem_method (stem_method will be set to 0)
|
|---|
| 103 | # Otherwise don't build with indexes
|
|---|
| 104 | set do_indexes = 0
|
|---|
| 105 |
|
|---|
| 106 |
|
|---|
| 107 | # If do_text == 1 then do the text component of building a mg database.
|
|---|
| 108 | # Otherwise don't do the text component.
|
|---|
| 109 | set do_text = 1
|
|---|
| 110 |
|
|---|
| 111 |
|
|---|
| 112 | # If do_invf == 1 then do the inversion component of building a mg database.
|
|---|
| 113 | # Otherwise don't do the inversion component.
|
|---|
| 114 | set do_invf = 1
|
|---|
| 115 |
|
|---|
| 116 |
|
|---|
| 117 | # If do_pass1 == 1 then do pass1 of building a mg database.
|
|---|
| 118 | # Otherwise don't do pass1.
|
|---|
| 119 | set do_pass1 = 1
|
|---|
| 120 |
|
|---|
| 121 | # If do_pass2 == 1 then do pass2 of building a mg database.
|
|---|
| 122 | # Otherwise don't do pass2.
|
|---|
| 123 | set do_pass2 = 1
|
|---|
| 124 |
|
|---|
| 125 |
|
|---|
| 126 | # $invf_mem specifies the amount of memory to use for the pass2 inversion
|
|---|
| 127 | set invf_mem = 32
|
|---|
| 128 |
|
|---|
| 129 |
|
|---|
| 130 | # $num_chunks specifies the number of interium chunks of inverted file that
|
|---|
| 131 | # may be written to disc before a merge into the invf file is done
|
|---|
| 132 | set num_chunks = 3
|
|---|
| 133 |
|
|---|
| 134 | # $invf_level specifies the level of the inverted file that will be generated.
|
|---|
| 135 | set invf_level = 2
|
|---|
| 136 |
|
|---|
| 137 |
|
|---|
| 138 | # If $strip_sgml == 1 then sgml tags are stripped from the inversion phase.
|
|---|
| 139 | # Otherwise sgml tags are kepted.
|
|---|
| 140 | set strip_sgml = 1
|
|---|
| 141 |
|
|---|
| 142 |
|
|---|
| 143 | # $trace specifies the interval between trace entries in Mb.
|
|---|
| 144 | # If this is not set no trace entries will be generated.
|
|---|
| 145 | set trace = 10
|
|---|
| 146 |
|
|---|
| 147 |
|
|---|
| 148 | # $weight_bits specifies the number of bits of precision bo be given to the
|
|---|
| 149 | # approximate weights.
|
|---|
| 150 | set weight_bits = 6
|
|---|
| 151 |
|
|---|
| 152 |
|
|---|
| 153 | # $mcd specifies the commandline arguments for the mg_compression_dict program
|
|---|
| 154 | #### Note: Set -S so novel words can be encoded, as this will happen if
|
|---|
| 155 | #### mgmerge is used.
|
|---|
| 156 | set mcd = -S
|
|---|
| 157 |
|
|---|
| 158 |
|
|---|
| 159 | # Source the parameter file to modify parameters of the build.
|
|---|
| 160 | if ($?source) then
|
|---|
| 161 | source ${source}
|
|---|
| 162 | endif
|
|---|
| 163 |
|
|---|
| 164 |
|
|---|
| 165 | # [RPAP - Jan 97: Stem Index Change]
|
|---|
| 166 | # If do_indexes == 1 then set stem_method = 0
|
|---|
| 167 | if ("$do_indexes" == "1") then
|
|---|
| 168 | set stem_method = 0
|
|---|
| 169 | endif
|
|---|
| 170 |
|
|---|
| 171 |
|
|---|
| 172 | # Generate the collection name.
|
|---|
| 173 | set coll_name = ${text}
|
|---|
| 174 | if ("$invf_level" == "3") then
|
|---|
| 175 | set coll_name = ${text}-p
|
|---|
| 176 | endif
|
|---|
| 177 |
|
|---|
| 178 | # Generate the directory for the collection.
|
|---|
| 179 | if (-e $MGDATA/${coll_name}) then
|
|---|
| 180 | else
|
|---|
| 181 | mkdir $MGDATA/${coll_name}
|
|---|
| 182 | endif
|
|---|
| 183 |
|
|---|
| 184 | # Generate the base name for the collection.
|
|---|
| 185 | set bname = ${coll_name}/${coll_name}
|
|---|
| 186 |
|
|---|
| 187 |
|
|---|
| 188 | #build up the command lines for pass 1 and 2
|
|---|
| 189 | set pass1 = (-f ${bname} -${invf_level} -m ${invf_mem} -s ${stem_method})
|
|---|
| 190 | set pass2 = (-f ${bname} -${invf_level} -c ${num_chunks})
|
|---|
| 191 |
|
|---|
| 192 | if ($strip_sgml) then
|
|---|
| 193 | set pass1 = (${pass1} -G)
|
|---|
| 194 | set pass2 = (${pass2} -G)
|
|---|
| 195 | endif
|
|---|
| 196 |
|
|---|
| 197 | if ($?trace) then
|
|---|
| 198 | set pass1 = (${pass1} -t ${trace})
|
|---|
| 199 | set pass2 = (${pass2} -t ${trace})
|
|---|
| 200 | endif
|
|---|
| 201 |
|
|---|
| 202 | if ($do_text) then
|
|---|
| 203 | set pass1 = (${pass1} -T1)
|
|---|
| 204 | set pass2 = (${pass2} -T2)
|
|---|
| 205 | endif
|
|---|
| 206 |
|
|---|
| 207 | if ($do_invf) then
|
|---|
| 208 | set pass1 = (${pass1} -I1)
|
|---|
| 209 | set pass2 = (${pass2} -I2)
|
|---|
| 210 | endif
|
|---|
| 211 |
|
|---|
| 212 | if ($?trace_name) then
|
|---|
| 213 | set pass1 = (${pass1} -n ${trace_name})
|
|---|
| 214 | set pass2 = (${pass2} -n ${trace_name})
|
|---|
| 215 | endif
|
|---|
| 216 |
|
|---|
| 217 | if ($?comp_stats) then
|
|---|
| 218 | set pass2 = (${pass2} -C ${comp_stats})
|
|---|
| 219 | endif
|
|---|
| 220 |
|
|---|
| 221 | echo "-----------------------------------"
|
|---|
| 222 | echo "`uname -n`, `date`"
|
|---|
| 223 | echo "${text} --> ${bname}"
|
|---|
| 224 | echo "-----------------------------------"
|
|---|
| 225 | if ($pipe) then
|
|---|
| 226 | if ("$complex" != "") then
|
|---|
| 227 | echo "$get $text -init"
|
|---|
| 228 | $get $text -init
|
|---|
| 229 | if ("$status" != "0") exit 1
|
|---|
| 230 | echo "-----------------------------------"
|
|---|
| 231 | endif
|
|---|
| 232 | endif
|
|---|
| 233 |
|
|---|
| 234 | if (${do_pass1}) then
|
|---|
| 235 | if ($pipe) then
|
|---|
| 236 | if ($?pass1filter) then
|
|---|
| 237 | echo "$get $text $complex | $pass1filter | mg_passes ${pass1}"
|
|---|
| 238 | $get $text $complex| $pass1filter | $bindir/mg_passes ${pass1}
|
|---|
| 239 | if ("$status" != "0") exit 1
|
|---|
| 240 | else
|
|---|
| 241 | echo "$get $text $complex | mg_passes ${pass1}"
|
|---|
| 242 | $get $text $complex| $bindir/mg_passes ${pass1}
|
|---|
| 243 | if ("$status" != "0") exit 1
|
|---|
| 244 | endif
|
|---|
| 245 | else
|
|---|
| 246 | echo mg_passes ${pass1} ${input_files}
|
|---|
| 247 | $bindir/mg_passes ${pass1} ${input_files}
|
|---|
| 248 | if ("$status" != "0") exit 1
|
|---|
| 249 | endif
|
|---|
| 250 | echo "-----------------------------------"
|
|---|
| 251 |
|
|---|
| 252 | echo "mg_perf_hash_build -f ${bname}"
|
|---|
| 253 | $bindir/mg_perf_hash_build -f ${bname}
|
|---|
| 254 | if ("$status" != "0") exit 1
|
|---|
| 255 |
|
|---|
| 256 | echo "-----------------------------------"
|
|---|
| 257 | echo "mg_compression_dict -f ${bname} ${mcd}"
|
|---|
| 258 | $bindir/mg_compression_dict -f ${bname} ${mcd}
|
|---|
| 259 | if ("$status" != "0") exit 1
|
|---|
| 260 | echo "-----------------------------------"
|
|---|
| 261 | endif
|
|---|
| 262 |
|
|---|
| 263 | if (${do_pass2}) then
|
|---|
| 264 | if ($pipe) then
|
|---|
| 265 | if ($?pass2filter) then
|
|---|
| 266 | echo "$get $text $complex | $pass2filter | mg_passes ${pass2}"
|
|---|
| 267 | $get $text $complex | $pass2filter | $bindir/mg_passes ${pass2}
|
|---|
| 268 | if ("$status" != "0") exit 1
|
|---|
| 269 | else
|
|---|
| 270 | echo "$get $text $complex | mg_passes ${pass2}"
|
|---|
| 271 | $get $text $complex | $bindir/mg_passes ${pass2}
|
|---|
| 272 | if ("$status" != "0") exit 1
|
|---|
| 273 | endif
|
|---|
| 274 | else
|
|---|
| 275 | echo mg_passes ${pass2} ${input_files}
|
|---|
| 276 | $bindir/mg_passes ${pass2} ${input_files}
|
|---|
| 277 | if ("$status" != "0") exit 1
|
|---|
| 278 | endif
|
|---|
| 279 | echo "-----------------------------------"
|
|---|
| 280 | endif
|
|---|
| 281 |
|
|---|
| 282 | echo "mg_weights_build -f ${bname} -b ${weight_bits}"
|
|---|
| 283 | $bindir/mg_weights_build -f ${bname} -b ${weight_bits}
|
|---|
| 284 | if ("$status" != "0") exit 1
|
|---|
| 285 |
|
|---|
| 286 | echo "-----------------------------------"
|
|---|
| 287 |
|
|---|
| 288 | echo "mg_invf_dict -f ${bname} -b 4096"
|
|---|
| 289 | $bindir/mg_invf_dict -f ${bname} -b 4096
|
|---|
| 290 | if ("$status" != "0") exit 1
|
|---|
| 291 |
|
|---|
| 292 | echo "-----------------------------------"
|
|---|
| 293 |
|
|---|
| 294 | # [RPAP - Jan 97: Stem Index Change]
|
|---|
| 295 | if ("$do_indexes" == "1") then
|
|---|
| 296 | echo "mg_stem_idx -f ${bname} -b 4096 -s1"
|
|---|
| 297 | $bindir/mg_stem_idx -f ${bname} -b 4096 -s1
|
|---|
| 298 | if ("$status" != "0") exit 1
|
|---|
| 299 |
|
|---|
| 300 | echo ""
|
|---|
| 301 |
|
|---|
| 302 | echo "mg_stem_idx -f ${bname} -b 4096 -s2"
|
|---|
| 303 | $bindir/mg_stem_idx -f ${bname} -b 4096 -s2
|
|---|
| 304 | if ("$status" != "0") exit 1
|
|---|
| 305 |
|
|---|
| 306 | echo ""
|
|---|
| 307 |
|
|---|
| 308 | echo "mg_stem_idx -f ${bname} -b 4096 -s3"
|
|---|
| 309 | $bindir/mg_stem_idx -f ${bname} -b 4096 -s3
|
|---|
| 310 | if ("$status" != "0") exit 1
|
|---|
| 311 |
|
|---|
| 312 | echo "-----------------------------------"
|
|---|
| 313 | endif
|
|---|
| 314 |
|
|---|
| 315 | echo "mgstat -f ${bname} -E"
|
|---|
| 316 | $bindir/mgstat -f ${bname} -E
|
|---|
| 317 | if ("$status" != "0") exit 1
|
|---|
| 318 |
|
|---|
| 319 | if ($pipe) then
|
|---|
| 320 | if ("$complex" != "") then
|
|---|
| 321 | echo "-----------------------------------"
|
|---|
| 322 | echo "$get $text -cleanup"
|
|---|
| 323 | $get $text -cleanup
|
|---|
| 324 | if ("$status" != "0") exit 1
|
|---|
| 325 | endif
|
|---|
| 326 | endif
|
|---|
| 327 |
|
|---|
| 328 | echo "-----------------------------------"
|
|---|
| 329 | echo "`uname -n`, `date`"
|
|---|
| 330 | echo "-----------------------------------"
|
|---|
| 331 |
|
|---|
| 332 | echo "-- The fast-loading compression dictionary has not been built."
|
|---|
| 333 | echo "-- If you wish to build it, execute the following command:"
|
|---|
| 334 | echo "-- "$bindir/mg_fast_comp_dict -f ${bname}
|
|---|
| 335 |
|
|---|
| 336 | echo "-----------------------------------"
|
|---|
| 337 |
|
|---|
| 338 | echo ""
|
|---|
| 339 |
|
|---|
