| 1 | #!/bin/csh -f
|
|---|
| 2 | ###########################################################################
|
|---|
| 3 | #
|
|---|
| 4 | # mgbuild_4.sh -- Script used to build mg text collection in 4 passes.
|
|---|
| 5 | # Copyright (C) 1994 Neil Sharman; Modified by tes@kbs 1995
|
|---|
| 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 | #
|
|---|
| 22 | ###########################################################################
|
|---|
| 23 | #
|
|---|
| 24 | # Does an mgbuild in 4 passes instead of 2 passes.
|
|---|
| 25 | # This is useful if one does not have as much RAM and a large collection.
|
|---|
| 26 | #
|
|---|
| 27 |
|
|---|
| 28 | set complex = ""
|
|---|
| 29 |
|
|---|
| 30 | # ************ parse the command line arguments ************
|
|---|
| 31 |
|
|---|
| 32 | while ($#argv >= 1)
|
|---|
| 33 | if ("$1" == "-s" || "$1" == "--source") then
|
|---|
| 34 | shift
|
|---|
| 35 | if ($#argv >= 1) then
|
|---|
| 36 | set source = $1
|
|---|
| 37 | shift
|
|---|
| 38 | endif
|
|---|
| 39 | else if ("$1" == "-g" || "$1" == "--get") then
|
|---|
| 40 | shift
|
|---|
| 41 | if ($#argv >= 1) then
|
|---|
| 42 | set get = $1
|
|---|
| 43 | shift
|
|---|
| 44 | endif
|
|---|
| 45 | else if ("$1" == "-c" || "$1" == "--complex") then
|
|---|
| 46 | set complex = "-text"
|
|---|
| 47 | shift
|
|---|
| 48 | else
|
|---|
| 49 | if (! $?text ) then
|
|---|
| 50 | set text = $1
|
|---|
| 51 | endif
|
|---|
| 52 | shift
|
|---|
| 53 | endif
|
|---|
| 54 |
|
|---|
| 55 | end
|
|---|
| 56 |
|
|---|
| 57 | if (! $?text ) then
|
|---|
| 58 | set prog = $0
|
|---|
| 59 | echo "USAGE:"
|
|---|
| 60 | echo " "$prog:t" [-s config-script] [-g get-program] [-c] source"
|
|---|
| 61 | echo ""
|
|---|
| 62 | echo " The config-script is only needed if a non-standard build is required."
|
|---|
| 63 | echo " The get-program defaults to mg_get if not specified."
|
|---|
| 64 | exit 1
|
|---|
| 65 | endif
|
|---|
| 66 |
|
|---|
| 67 | set bindir = `which $0` # find which one we are executing
|
|---|
| 68 | set bindir = $bindir:h # use it as the bindir path
|
|---|
| 69 |
|
|---|
| 70 | # ************ set up shell variables ************
|
|---|
| 71 | # Override these by providing a script to source using the -s option
|
|---|
| 72 |
|
|---|
| 73 | # if $pipe == 1 then pipe in the source text using $get and $text otherwise
|
|---|
| 74 | # read the source text directly from the file names specified in $input_files
|
|---|
| 75 | set pipe = 1
|
|---|
| 76 |
|
|---|
| 77 | set get_options = ""
|
|---|
| 78 |
|
|---|
| 79 |
|
|---|
| 80 | if (! $?get ) then
|
|---|
| 81 | set complex = "-text"
|
|---|
| 82 | set get = $bindir/mg_get $get_options
|
|---|
| 83 | endif
|
|---|
| 84 |
|
|---|
| 85 |
|
|---|
| 86 | if (-e $MGDATA/${text}.chunks) then
|
|---|
| 87 | set input_files = `cat $MGDATA/${text}.chunks`
|
|---|
| 88 | endif
|
|---|
| 89 |
|
|---|
| 90 | # Set the stemming method
|
|---|
| 91 | # Bit 0 = case folding
|
|---|
| 92 | # Bit 1 = S stemmer
|
|---|
| 93 | set stem_method = 3
|
|---|
| 94 |
|
|---|
| 95 |
|
|---|
| 96 | # [RPAP - Jan 97: Stem Index Change]
|
|---|
| 97 | # If do_indexes == 1 then build collection with full indexes to blocked file
|
|---|
| 98 | # overriding stem_method (stem_method will be set to 0)
|
|---|
| 99 | # Otherwise don't build with indexes
|
|---|
| 100 | set do_indexes = 0
|
|---|
| 101 |
|
|---|
| 102 |
|
|---|
| 103 | # whether to build weights or not
|
|---|
| 104 | set do_weights = 1
|
|---|
| 105 |
|
|---|
| 106 | # If do_pass1 == 1 then do pass1 of building a mg database.
|
|---|
| 107 | # Otherwise don't do pass1.
|
|---|
| 108 | # Text 1st pass
|
|---|
| 109 | set do_pass1 = 1
|
|---|
| 110 |
|
|---|
| 111 | # If do_pass2 == 1 then do pass2 of building a mg database.
|
|---|
| 112 | # Otherwise don't do pass2.
|
|---|
| 113 | # Text 2nd pass
|
|---|
| 114 | set do_pass2 = 1
|
|---|
| 115 |
|
|---|
| 116 | # If do_pass3 == 1 then do pass3 of building a mg database.
|
|---|
| 117 | # Otherwise don't do pass3.
|
|---|
| 118 | # Invf 1st pass
|
|---|
| 119 | set do_pass3 = 1
|
|---|
| 120 |
|
|---|
| 121 | # If do_pass4 == 1 then do pass4 of building a mg database.
|
|---|
| 122 | # Otherwise don't do pass4.
|
|---|
| 123 | # Invf 2nd pass
|
|---|
| 124 | set do_pass4 = 1
|
|---|
| 125 |
|
|---|
| 126 | # Buffer size used by mg_passes
|
|---|
| 127 | # Should be big enough to hold largest document
|
|---|
| 128 | # In Kilobytes
|
|---|
| 129 | set buf_size = 3072
|
|---|
| 130 |
|
|---|
| 131 | # $invf_mem specifies the amount of memory to use for the pass2 inversion
|
|---|
| 132 | set invf_mem = 5
|
|---|
| 133 |
|
|---|
| 134 | # $num_chunks specifies the number of interium chunks of inverted file that
|
|---|
| 135 | # may be written to disc before a merge into the invf file is done
|
|---|
| 136 | set num_chunks = 3
|
|---|
| 137 |
|
|---|
| 138 | # $invf_level specifies the level of the inverted file that will be generated.
|
|---|
| 139 | set invf_level = 2
|
|---|
| 140 |
|
|---|
| 141 |
|
|---|
| 142 | # If $strip_sgml == 1 then sgml tags are stripped from the inversion phase.
|
|---|
| 143 | # Otherwise sgml tags are kepted.
|
|---|
| 144 | set strip_sgml = 1
|
|---|
| 145 |
|
|---|
| 146 |
|
|---|
| 147 | # $trace specifies the interval between trace entries in Mb.
|
|---|
| 148 | # If this is not set no trace entries will be generated.
|
|---|
| 149 | set trace = 10
|
|---|
| 150 |
|
|---|
| 151 |
|
|---|
| 152 | # $weight_bits specifies the number of bits of precision bo be given to the
|
|---|
| 153 | # approximate weights.
|
|---|
| 154 | set weight_bits = 6
|
|---|
| 155 |
|
|---|
| 156 |
|
|---|
| 157 | # $mcd specifies the commandline arguments for the mg_compression_dict program
|
|---|
| 158 | set mcd = -C
|
|---|
| 159 |
|
|---|
| 160 |
|
|---|
| 161 | # Source the parameter file to modify parameters of the build.
|
|---|
| 162 | if ($?source) then
|
|---|
| 163 | source ${source}
|
|---|
| 164 | endif
|
|---|
| 165 |
|
|---|
| 166 |
|
|---|
| 167 | # [RPAP - Jan 97: Stem Index Change]
|
|---|
| 168 | # If do_indexes == 1 then set stem_method = 0
|
|---|
| 169 | if ("$do_indexes" == "1") then
|
|---|
| 170 | set stem_method = 0
|
|---|
| 171 | endif
|
|---|
| 172 |
|
|---|
| 173 |
|
|---|
| 174 | # Generate the collection name.
|
|---|
| 175 | set coll_name = ${text}
|
|---|
| 176 | if ("$invf_level" == "3") then
|
|---|
| 177 | set coll_name = ${text}-p
|
|---|
| 178 | endif
|
|---|
| 179 |
|
|---|
| 180 | # Generate the directory for the collection.
|
|---|
| 181 | if (-e $MGDATA/${coll_name}) then
|
|---|
| 182 | else
|
|---|
| 183 | mkdir $MGDATA/${coll_name}
|
|---|
| 184 | endif
|
|---|
| 185 |
|
|---|
| 186 | # Generate the base name for the collection.
|
|---|
| 187 | set bname = ${coll_name}/${coll_name}
|
|---|
| 188 |
|
|---|
| 189 | # ************ set up pass parameters ************
|
|---|
| 190 |
|
|---|
| 191 | if ($do_pass1) then
|
|---|
| 192 | set pass1_text = (-f ${bname} -T1)
|
|---|
| 193 | endif
|
|---|
| 194 |
|
|---|
| 195 | if ($do_pass2) then
|
|---|
| 196 | set pass2_text = (-f ${bname} -T2)
|
|---|
| 197 | endif
|
|---|
| 198 |
|
|---|
| 199 | if ($do_pass3) then
|
|---|
| 200 | set pass3_invf = (-f ${bname} -${invf_level} -m ${invf_mem} -s ${stem_method})
|
|---|
| 201 | set pass3_invf = (${pass3_invf} -N1)
|
|---|
| 202 | endif
|
|---|
| 203 |
|
|---|
| 204 | if ($do_pass4) then
|
|---|
| 205 | set pass4_invf = (-f ${bname} -${invf_level} -N2)
|
|---|
| 206 | endif
|
|---|
| 207 |
|
|---|
| 208 | if ($strip_sgml) then
|
|---|
| 209 | set pass1_text = (${pass1_text} -G)
|
|---|
| 210 | set pass2_text = (${pass2_text} -G)
|
|---|
| 211 | set pass3_invf = (${pass3_invf} -G)
|
|---|
| 212 | set pass4_invf = (${pass4_invf} -G)
|
|---|
| 213 | endif
|
|---|
| 214 |
|
|---|
| 215 | if ($?trace) then
|
|---|
| 216 | set pass1_text = (${pass1_text} -t ${trace})
|
|---|
| 217 | set pass2_text = (${pass2_text} -t ${trace})
|
|---|
| 218 | set pass3_invf = (${pass3_invf} -t ${trace})
|
|---|
| 219 | set pass4_invf = (${pass4_invf} -t ${trace})
|
|---|
| 220 | endif
|
|---|
| 221 |
|
|---|
| 222 | if ($?trace_name) then
|
|---|
| 223 | set pass1_text = (${pass1_text} -n ${trace_name})
|
|---|
| 224 | set pass2_text = (${pass2_text} -n ${trace_name})
|
|---|
| 225 | set pass3_invf = (${pass3_invf} -n ${trace_name})
|
|---|
| 226 | set pass4_invf = (${pass4_invf} -n ${trace_name})
|
|---|
| 227 | endif
|
|---|
| 228 |
|
|---|
| 229 | if ($buf_size) then
|
|---|
| 230 | set pass1_text = (${pass1_text} -b ${buf_size})
|
|---|
| 231 | set pass2_text = (${pass2_text} -b ${buf_size})
|
|---|
| 232 | set pass3_invf = (${pass3_invf} -b ${buf_size})
|
|---|
| 233 | set pass4_invf = (${pass4_invf} -b ${buf_size})
|
|---|
| 234 | endif
|
|---|
| 235 |
|
|---|
| 236 | if ($?comp_stats) then
|
|---|
| 237 | set pass2_text = (${pass2_text} -C ${comp_stats})
|
|---|
| 238 | endif
|
|---|
| 239 |
|
|---|
| 240 | echo "-----------------------------------"
|
|---|
| 241 | echo "`uname -n`, `date`"
|
|---|
| 242 | echo "${text} --> ${bname}"
|
|---|
| 243 | echo "-----------------------------------"
|
|---|
| 244 |
|
|---|
| 245 | # ************ init get ************
|
|---|
| 246 |
|
|---|
| 247 | if ($pipe) then
|
|---|
| 248 | if ("$complex" != "") then
|
|---|
| 249 | echo "$get $text -init"
|
|---|
| 250 | $get $text -init
|
|---|
| 251 | if ("$status" != "0") exit 1
|
|---|
| 252 | echo "-----------------------------------"
|
|---|
| 253 | endif
|
|---|
| 254 | endif
|
|---|
| 255 |
|
|---|
| 256 | # ************************************* T E X T **********************************
|
|---|
| 257 |
|
|---|
| 258 | # ************ pass 1_text ************
|
|---|
| 259 |
|
|---|
| 260 | if (${do_pass1}) then
|
|---|
| 261 | if ($pipe) then
|
|---|
| 262 | if ($?pass1filter) then
|
|---|
| 263 | echo "$get $text $complex | $pass1filter | mg_passes ${pass1_text}"
|
|---|
| 264 | $get $text $complex| $pass1filter | $bindir/mg_passes ${pass1_text}
|
|---|
| 265 | if ("$status" != "0") exit 1
|
|---|
| 266 | else
|
|---|
| 267 | echo "$get $text $complex | mg_passes ${pass1_text}"
|
|---|
| 268 | $get $text $complex| $bindir/mg_passes ${pass1_text}
|
|---|
| 269 | if ("$status" != "0") exit 1
|
|---|
| 270 | endif
|
|---|
| 271 | else
|
|---|
| 272 | echo mg_passes ${pass1_text} ${input_files}
|
|---|
| 273 | $bindir/mg_passes ${pass1_text} ${input_files}
|
|---|
| 274 | if ("$status" != "0") exit 1
|
|---|
| 275 | endif
|
|---|
| 276 | echo "-----------------------------------"
|
|---|
| 277 | endif
|
|---|
| 278 |
|
|---|
| 279 | # ************ compression dict ************
|
|---|
| 280 | if ($do_pass2) then
|
|---|
| 281 | echo "mg_compression_dict -f ${bname} ${mcd}"
|
|---|
| 282 | $bindir/mg_compression_dict -f ${bname} ${mcd}
|
|---|
| 283 | if ("$status" != "0") exit 1
|
|---|
| 284 | echo "-----------------------------------"
|
|---|
| 285 | endif
|
|---|
| 286 |
|
|---|
| 287 | # ************ pass 2_text ************
|
|---|
| 288 | if (${do_pass2}) then
|
|---|
| 289 | if ($pipe) then
|
|---|
| 290 | if ($?pass2filter) then
|
|---|
| 291 | echo "$get $text $complex | $pass2filter | mg_passes ${pass2_text}"
|
|---|
| 292 | $get $text $complex| $pass2filter | $bindir/mg_passes ${pass2_text}
|
|---|
| 293 | if ("$status" != "0") exit 1
|
|---|
| 294 | else
|
|---|
| 295 | echo "$get $text $complex | mg_passes ${pass2_text}"
|
|---|
| 296 | $get $text $complex| $bindir/mg_passes ${pass2_text}
|
|---|
| 297 | if ("$status" != "0") exit 1
|
|---|
| 298 | endif
|
|---|
| 299 | else
|
|---|
| 300 | echo mg_passes ${pass2_text} ${input_files}
|
|---|
| 301 | $bindir/mg_passes ${pass2_text} ${input_files}
|
|---|
| 302 | if ("$status" != "0") exit 1
|
|---|
| 303 | endif
|
|---|
| 304 | echo "-----------------------------------"
|
|---|
| 305 | endif
|
|---|
| 306 |
|
|---|
| 307 | # ************************************* I N V F **********************************
|
|---|
| 308 |
|
|---|
| 309 | # ************ pass 3_invf ************
|
|---|
| 310 | if (${do_pass3}) then
|
|---|
| 311 | if ($pipe) then
|
|---|
| 312 | if ($?pass3filter) then
|
|---|
| 313 | echo "$get $text $complex | $pass3filter | mg_passes ${pass3_invf}"
|
|---|
| 314 | $get $text $complex| $pass3filter | $bindir/mg_passes ${pass3_invf}
|
|---|
| 315 | if ("$status" != "0") exit 1
|
|---|
| 316 | else
|
|---|
| 317 | echo "$get $text $complex | mg_passes ${pass3_invf}"
|
|---|
| 318 | $get $text $complex| $bindir/mg_passes ${pass3_invf}
|
|---|
| 319 | if ("$status" != "0") exit 1
|
|---|
| 320 | endif
|
|---|
| 321 | else
|
|---|
| 322 | echo mg_passes ${pass3_invf} ${input_files}
|
|---|
| 323 | $bindir/mg_passes ${pass3_invf} ${input_files}
|
|---|
| 324 | if ("$status" != "0") exit 1
|
|---|
| 325 | endif
|
|---|
| 326 | echo "-----------------------------------"
|
|---|
| 327 | endif
|
|---|
| 328 |
|
|---|
| 329 | # ************ perfect hash ************
|
|---|
| 330 | if ($do_pass4) then
|
|---|
| 331 | echo "mg_perf_hash_build -f ${bname}"
|
|---|
| 332 | $bindir/mg_perf_hash_build -f ${bname}
|
|---|
| 333 | if ("$status" != "0") exit 1
|
|---|
| 334 |
|
|---|
| 335 | echo "-----------------------------------"
|
|---|
| 336 | endif
|
|---|
| 337 |
|
|---|
| 338 | # ************ pass 4_invf ************
|
|---|
| 339 | if (${do_pass4}) then
|
|---|
| 340 | if ($pipe) then
|
|---|
| 341 | if ($?pass4filter) then
|
|---|
| 342 | echo "$get $text $complex | $pass4filter | mg_passes ${pass4_invf}"
|
|---|
| 343 | $get $text $complex| $pass4filter | $bindir/mg_passes ${pass4_invf}
|
|---|
| 344 | if ("$status" != "0") exit 1
|
|---|
| 345 | else
|
|---|
| 346 | echo "$get $text $complex | mg_passes ${pass4_invf}"
|
|---|
| 347 | $get $text $complex| $bindir/mg_passes ${pass4_invf}
|
|---|
| 348 | if ("$status" != "0") exit 1
|
|---|
| 349 | endif
|
|---|
| 350 | else
|
|---|
| 351 | echo mg_passes ${pass4_invf} ${input_files}
|
|---|
| 352 | $bindir/mg_passes ${pass4_invf} ${input_files}
|
|---|
| 353 | if ("$status" != "0") exit 1
|
|---|
| 354 | endif
|
|---|
| 355 | echo "-----------------------------------"
|
|---|
| 356 | endif
|
|---|
| 357 |
|
|---|
| 358 |
|
|---|
| 359 | # ************ build invf dictionary ************
|
|---|
| 360 |
|
|---|
| 361 | if ($do_pass4) then
|
|---|
| 362 | echo "mg_invf_dict -f ${bname} -b 4096"
|
|---|
| 363 | $bindir/mg_invf_dict -f ${bname} -b 4096
|
|---|
| 364 | if ("$status" != "0") exit 1
|
|---|
| 365 | echo "-----------------------------------"
|
|---|
| 366 | endif
|
|---|
| 367 |
|
|---|
| 368 | # ******** build indexes to stem dict *********
|
|---|
| 369 |
|
|---|
| 370 | # [RPAP - Jan 97: Stem Index Change]
|
|---|
| 371 | if ($do_pass4 && "$do_indexes" == "1") then
|
|---|
| 372 | echo "mg_stem_idx -f ${bname} -b 4096 -s1"
|
|---|
| 373 | $bindir/mg_stem_idx -f ${bname} -b 4096 -s1
|
|---|
| 374 | if ("$status" != "0") exit 1
|
|---|
| 375 |
|
|---|
| 376 | echo ""
|
|---|
| 377 |
|
|---|
| 378 | echo "mg_stem_idx -f ${bname} -b 4096 -s2"
|
|---|
| 379 | $bindir/mg_stem_idx -f ${bname} -b 4096 -s2
|
|---|
| 380 | if ("$status" != "0") exit 1
|
|---|
| 381 |
|
|---|
| 382 | echo ""
|
|---|
| 383 |
|
|---|
| 384 | echo "mg_stem_idx -f ${bname} -b 4096 -s3"
|
|---|
| 385 | $bindir/mg_stem_idx -f ${bname} -b 4096 -s3
|
|---|
| 386 | if ("$status" != "0") exit 1
|
|---|
| 387 |
|
|---|
| 388 | echo "-----------------------------------"
|
|---|
| 389 | endif
|
|---|
| 390 |
|
|---|
| 391 | # ************ build weights ************
|
|---|
| 392 |
|
|---|
| 393 | if ($do_pass4 && $do_weights) then
|
|---|
| 394 | echo "mg_weights_build -f ${bname} -b ${weight_bits} "
|
|---|
| 395 | $bindir/mg_weights_build -f ${bname} -b ${weight_bits}
|
|---|
| 396 | if ("$status" != "0") exit 1
|
|---|
| 397 | echo "-----------------------------------"
|
|---|
| 398 | endif
|
|---|
| 399 |
|
|---|
| 400 | # *******************************************************************************
|
|---|
| 401 |
|
|---|
| 402 |
|
|---|
| 403 | # ************ print out statistics ************
|
|---|
| 404 | if ($do_pass1 && $do_pass2 && $do_pass3 && $do_pass4) then
|
|---|
| 405 | echo "mgstat -f ${bname} -E"
|
|---|
| 406 | $bindir/mgstat -f ${bname} -E
|
|---|
| 407 | if ("$status" != "0") exit 1
|
|---|
| 408 | endif
|
|---|
| 409 |
|
|---|
| 410 |
|
|---|
| 411 | # ************ cleanup get ************
|
|---|
| 412 |
|
|---|
| 413 | if ($pipe) then
|
|---|
| 414 | if ("$complex" != "") then
|
|---|
| 415 | echo "-----------------------------------"
|
|---|
| 416 | echo "$get $text -cleanup"
|
|---|
| 417 | $get $text -cleanup
|
|---|
| 418 | if ("$status" != "0") exit 1
|
|---|
| 419 | endif
|
|---|
| 420 | endif
|
|---|
| 421 |
|
|---|
| 422 | echo "-----------------------------------"
|
|---|
| 423 | echo "`uname -n`, `date`"
|
|---|
| 424 | echo "-----------------------------------"
|
|---|
| 425 |
|
|---|
| 426 | if ($do_pass2) then
|
|---|
| 427 | echo "-- The fast-loading compression dictionary has not been built."
|
|---|
| 428 | echo "-- If you wish to build it, execute the following command:"
|
|---|
| 429 | echo "-- "$bindir/mg_fast_comp_dict -f ${bname}
|
|---|
| 430 |
|
|---|
| 431 | echo "-----------------------------------"
|
|---|
| 432 | endif
|
|---|
| 433 |
|
|---|
| 434 | echo ""
|
|---|
| 435 |
|
|---|
