| 1 | #!/bin/csh -f
|
|---|
| 2 | ###########################################################################
|
|---|
| 3 | #
|
|---|
| 4 | # mgmerge.sh -- Script used to build mg text collection.
|
|---|
| 5 | # Copyright (C) 1994 Neil Sharman, Shane Hudson (mods. for merging)
|
|---|
| 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 | set complex = ""
|
|---|
| 25 | set slow_merge = ""
|
|---|
| 26 | set guess_weights = "-w "
|
|---|
| 27 | # [RPAP - Feb 97: Level 3 Merge]
|
|---|
| 28 | set text_pass = 1
|
|---|
| 29 |
|
|---|
| 30 | # Parse the command line arguments
|
|---|
| 31 | while ($#argv >= 1)
|
|---|
| 32 | if ("$1" == "-s") then
|
|---|
| 33 | shift
|
|---|
| 34 | if ($#argv >= 1) then
|
|---|
| 35 | set source = $1
|
|---|
| 36 | shift
|
|---|
| 37 | endif
|
|---|
| 38 | else if ("$1" == "-g") then
|
|---|
| 39 | shift
|
|---|
| 40 | if ($#argv >= 1) then
|
|---|
| 41 | set get = $1
|
|---|
| 42 | shift
|
|---|
| 43 | endif
|
|---|
| 44 | else if ("$1" == "-c") then
|
|---|
| 45 | set complex = "-text"
|
|---|
| 46 | shift
|
|---|
| 47 | else if ("$1" == "-S") then
|
|---|
| 48 | # set slow merge option in mg_invf_merge
|
|---|
| 49 | set slow_merge = "-s "
|
|---|
| 50 | shift
|
|---|
| 51 | else if ("$1" == "-w") then
|
|---|
| 52 | # turn weights guessing off (rebuild weights file from scratch)
|
|---|
| 53 | set guess_weights = ""
|
|---|
| 54 | shift
|
|---|
| 55 |
|
|---|
| 56 | # [RPAP - Feb 97: Level 3 Merge]
|
|---|
| 57 | else if ("$1" == "-T") then
|
|---|
| 58 | # turn text pass off
|
|---|
| 59 | set text_pass = 0
|
|---|
| 60 | shift
|
|---|
| 61 |
|
|---|
| 62 | else
|
|---|
| 63 | if ($?text == "0") then
|
|---|
| 64 | set text = $1
|
|---|
| 65 | endif
|
|---|
| 66 | shift
|
|---|
| 67 | endif
|
|---|
| 68 |
|
|---|
| 69 | end
|
|---|
| 70 |
|
|---|
| 71 | if ($?text == "0") then
|
|---|
| 72 | set prog = $0
|
|---|
| 73 | echo "USAGE:"
|
|---|
| 74 | echo " "$prog:t" [-s config-script] [-g get-program] [-c] source"
|
|---|
| 75 | echo ""
|
|---|
| 76 | echo " The config-script is only needed if a non-standard build is required."
|
|---|
| 77 | echo " The get-program defaults to mg_get if not specified."
|
|---|
| 78 | exit 1
|
|---|
| 79 | endif
|
|---|
| 80 |
|
|---|
| 81 | set bindir = $0
|
|---|
| 82 | set bindir = $bindir:h
|
|---|
| 83 |
|
|---|
| 84 | # if $pipe == 1 then pipe in the source text using $get and $text otherwise
|
|---|
| 85 | # read the source text directly from the file names specified in $input_files
|
|---|
| 86 | set pipe = 1
|
|---|
| 87 |
|
|---|
| 88 | if ($?get == "0") then
|
|---|
| 89 | set complex = "-text"
|
|---|
| 90 | set get = $bindir/mg_get_merge
|
|---|
| 91 | endif
|
|---|
| 92 |
|
|---|
| 93 | if (-e $MGDATA/${text}.chunks) then
|
|---|
| 94 | set input_files = `cat $MGDATA/${text}.chunks`
|
|---|
| 95 | endif
|
|---|
| 96 |
|
|---|
| 97 | # Set the inversion method this may be either "I" or "N"
|
|---|
| 98 | # N is memory-efficient, I is faster for small collections
|
|---|
| 99 | set invf_method = N
|
|---|
| 100 |
|
|---|
| 101 |
|
|---|
| 102 | # Set the stemming method
|
|---|
| 103 | # Bit 0 = case folding
|
|---|
| 104 | # Bit 1 = S stemmer
|
|---|
| 105 | set stem_method = 3
|
|---|
| 106 |
|
|---|
| 107 |
|
|---|
| 108 | # [RPAP - Jan 97: Stem Index Change]
|
|---|
| 109 | # If do_indexes == 1 then build collection with full indexes to blocked file
|
|---|
| 110 | # overriding stem_method (stem_method will be set to 0)
|
|---|
| 111 | # Otherwise don't build with indexes
|
|---|
| 112 | set do_indexes = 0
|
|---|
| 113 |
|
|---|
| 114 |
|
|---|
| 115 | # $invf_mem specifies the amount of memory to use for the pass2 inversion
|
|---|
| 116 | # This only has an effect if $invf_method is "N".
|
|---|
| 117 | set invf_mem = 32
|
|---|
| 118 |
|
|---|
| 119 |
|
|---|
| 120 | # $num_chunks specifies the number of interium chunks of inverted file that
|
|---|
| 121 | # may be written to disc before a merge into the invf file is done
|
|---|
| 122 | # This only has an effect if $invf_method is "N".
|
|---|
| 123 | set num_chunks = 3
|
|---|
| 124 |
|
|---|
| 125 | # $invf_level specifies the level of the inverted file that will be generated.
|
|---|
| 126 | # Note: The value *MUST* be the same as was used to build the collection
|
|---|
| 127 | # that is being merged.
|
|---|
| 128 | # In the current mg_invf_merge implementation the maximum level is 2.
|
|---|
| 129 | # Paragraph-level inversion (invf_level = 3) is NOT supported.
|
|---|
| 130 | set invf_level = 2
|
|---|
| 131 |
|
|---|
| 132 |
|
|---|
| 133 | # If $strip_sgml == 1 then sgml tags are stripped from the inversion phase.
|
|---|
| 134 | # Otherwise sgml tags are kepted.
|
|---|
| 135 | set strip_sgml = 1
|
|---|
| 136 |
|
|---|
| 137 |
|
|---|
| 138 | # $trace specifies the interval between trace entries in Mb.
|
|---|
| 139 | # If this is not set no trace entries will be generated.
|
|---|
| 140 | set trace = 10
|
|---|
| 141 |
|
|---|
| 142 |
|
|---|
| 143 | # $weight_bits specifies the number of bits of precision bo be given to the
|
|---|
| 144 | # approximate weights.
|
|---|
| 145 | set weight_bits = 6
|
|---|
| 146 |
|
|---|
| 147 |
|
|---|
| 148 | # $mcd specifies the commandline arguments for the mg_compression_dict program
|
|---|
| 149 | set mcd = -S
|
|---|
| 150 |
|
|---|
| 151 | # $merge_name is the name of the subdirectory under ${MGDATA} where
|
|---|
| 152 | # the mergeing steps will be performed. There should NOT be an mg
|
|---|
| 153 | # collection with this name! The default is "MERGE"
|
|---|
| 154 | set merge = "MERGE"
|
|---|
| 155 |
|
|---|
| 156 | # Source the parameter file to modify parameters of the build.
|
|---|
| 157 | if ($?source) then
|
|---|
| 158 | source ${source}
|
|---|
| 159 | endif
|
|---|
| 160 |
|
|---|
| 161 |
|
|---|
| 162 | # [RPAP - Jan 97: Stem Index Change]
|
|---|
| 163 | # If do_indexes == 1 then set stem_method = 0
|
|---|
| 164 | if ($do_indexes) then
|
|---|
| 165 | set stem_method = 0
|
|---|
| 166 | endif
|
|---|
| 167 |
|
|---|
| 168 |
|
|---|
| 169 | # Generate the collection name.
|
|---|
| 170 | set coll_name = ${text}
|
|---|
| 171 |
|
|---|
| 172 | # Generate the directory where merging will take place
|
|---|
| 173 | if (-e $MGDATA/${merge}) then
|
|---|
| 174 | else
|
|---|
| 175 | mkdir $MGDATA/${merge}
|
|---|
| 176 | endif
|
|---|
| 177 |
|
|---|
| 178 | # Generate the base name for the collection.
|
|---|
| 179 | # [RPAP - Feb 97: Level 3 Merge]
|
|---|
| 180 | set bname = ${coll_name}
|
|---|
| 181 |
|
|---|
| 182 | # generate the old, new and merge names in the $merge directory
|
|---|
| 183 | set oldname = ${merge}/${coll_name}.old
|
|---|
| 184 | set newname = ${merge}/${coll_name}.new
|
|---|
| 185 | set mergename = ${merge}/${coll_name}
|
|---|
| 186 |
|
|---|
| 187 | # build up the command lines for pass 1 and 2 to build a
|
|---|
| 188 | # dummy collection for the new documents
|
|---|
| 189 | set pass1 = (-f ${newname} -${invf_level} -m ${invf_mem} -s ${stem_method})
|
|---|
| 190 | set pass2 = (-f ${newname} -${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 ${trace})
|
|---|
| 200 | endif
|
|---|
| 201 |
|
|---|
| 202 | # Note that a -T1 pass isnt done since the old .text.stats file is used
|
|---|
| 203 | # as the compression model for compressing the new documents.
|
|---|
| 204 |
|
|---|
| 205 | # [RPAP - Feb 97: Level 3 Merge]
|
|---|
| 206 | # Do not perform -T2 pass if text_pass == 0
|
|---|
| 207 | if ($text_pass) then
|
|---|
| 208 | set pass2 = (${pass2} -T2)
|
|---|
| 209 | endif
|
|---|
| 210 |
|
|---|
| 211 | set pass1 = (${pass1} -${invf_method}1)
|
|---|
| 212 | set pass2 = (${pass2} -${invf_method}2)
|
|---|
| 213 |
|
|---|
| 214 |
|
|---|
| 215 | if ($?trace_name) then
|
|---|
| 216 | set pass1 = (${pass1} -n ${trace_name})
|
|---|
| 217 | set pass2 = (${pass2} -n ${trace_name})
|
|---|
| 218 | endif
|
|---|
| 219 |
|
|---|
| 220 | if ($?comp_stats) then
|
|---|
| 221 | set pass2 = (${pass2} -C ${comp_stats})
|
|---|
| 222 | endif
|
|---|
| 223 |
|
|---|
| 224 | ########################################################################
|
|---|
| 225 | # Here is where mgmerge goes to work.
|
|---|
| 226 |
|
|---|
| 227 | echo "--------------------------------------------------------------"
|
|---|
| 228 | echo "`uname -n`, `date`"
|
|---|
| 229 | echo "MGMERGE collection: ${bname}"
|
|---|
| 230 | echo "--------------------------------------------------------------"
|
|---|
| 231 | echo "FIRST PHASE: build new collection from new documents"
|
|---|
| 232 | echo "--------------------------------------------------------------"
|
|---|
| 233 |
|
|---|
| 234 | # move and rename the text dictionary file
|
|---|
| 235 | # [RPAP - Feb 97: Level 3 Merge] - only if performing text pass
|
|---|
| 236 | if ($text_pass) then
|
|---|
| 237 | mv $MGDATA/${bname}.text.dict $MGDATA/${newname}.text.dict
|
|---|
| 238 | endif
|
|---|
| 239 |
|
|---|
| 240 | if ($pipe) then
|
|---|
| 241 | if ("$complex" != "") then
|
|---|
| 242 | echo ">> $get $text -init"
|
|---|
| 243 | $get $text -init
|
|---|
| 244 | if ("$status" != "0") exit 1
|
|---|
| 245 | echo "-----------------------------------"
|
|---|
| 246 | endif
|
|---|
| 247 | endif
|
|---|
| 248 |
|
|---|
| 249 | if ($pipe) then
|
|---|
| 250 | if ($?pass1filter) then
|
|---|
| 251 | echo ">> $get $text $complex | $pass1filter | mg_passes ${pass1}"
|
|---|
| 252 | $get $text $complex| $pass1filter | $bindir/mg_passes ${pass1}
|
|---|
| 253 | if ("$status" != "0") exit 1
|
|---|
| 254 | else
|
|---|
| 255 | echo ">> $get $text $complex | mg_passes ${pass1}"
|
|---|
| 256 | $get $text $complex| $bindir/mg_passes ${pass1}
|
|---|
| 257 | if ("$status" != "0") exit 1
|
|---|
| 258 | endif
|
|---|
| 259 | else
|
|---|
| 260 | echo ">> mg_passes ${pass1} ${input_files}"
|
|---|
| 261 | $bindir/mg_passes ${pass1} ${input_files}
|
|---|
| 262 | if ("$status" != "0") exit 1
|
|---|
| 263 | endif
|
|---|
| 264 | echo "-----------------------------------"
|
|---|
| 265 |
|
|---|
| 266 | echo "mg_perf_hash_build -f ${newname}"
|
|---|
| 267 | $bindir/mg_perf_hash_build -f ${newname}
|
|---|
| 268 | if ("$status" != "0") exit 1
|
|---|
| 269 | echo "-----------------------------------"
|
|---|
| 270 |
|
|---|
| 271 | if ($pipe) then
|
|---|
| 272 | if ($?pass2filter) then
|
|---|
| 273 | echo ">> $get $text $complex | $pass2filter | mg_passes ${pass2}"
|
|---|
| 274 | $get $text $complex | $pass2filter | $bindir/mg_passes ${pass2}
|
|---|
| 275 | if ("$status" != "0") exit 1
|
|---|
| 276 | else
|
|---|
| 277 | echo ">> $get $text $complex | mg_passes ${pass2}"
|
|---|
| 278 | $get $text $complex | $bindir/mg_passes ${pass2}
|
|---|
| 279 | if ("$status" != "0") exit 1
|
|---|
| 280 | endif
|
|---|
| 281 | else
|
|---|
| 282 | echo ">> mg_passes ${pass2} ${input_files}"
|
|---|
| 283 | $bindir/mg_passes ${pass2} ${input_files}
|
|---|
| 284 | if ("$status" != "0") exit 1
|
|---|
| 285 | endif
|
|---|
| 286 | echo "-----------------------------------"
|
|---|
| 287 |
|
|---|
| 288 | if ($pipe) then
|
|---|
| 289 | if ("$complex" != "") then
|
|---|
| 290 | echo "-----------------------------------"
|
|---|
| 291 | echo ">> $get $text -cleanup"
|
|---|
| 292 | $get $text -cleanup
|
|---|
| 293 | if ("$status" != "0") exit 1
|
|---|
| 294 | endif
|
|---|
| 295 | endif
|
|---|
| 296 |
|
|---|
| 297 | echo "---------------------------------------"
|
|---|
| 298 | date
|
|---|
| 299 | echo "--------------------------------------------------------------"
|
|---|
| 300 | echo "SECOND PHASE: merge the two collections"
|
|---|
| 301 | echo "--------------------------------------------------------------"
|
|---|
| 302 |
|
|---|
| 303 | # move the appropriate files from the base directory to the merge dir
|
|---|
| 304 | echo "(moving files....)"
|
|---|
| 305 | # [RPAP - Feb 97: Level 3 Merge]
|
|---|
| 306 | # Move files if doing text pass
|
|---|
| 307 | if ($text_pass) then
|
|---|
| 308 | mv $MGDATA/${bname}.text $MGDATA/${oldname}.text
|
|---|
| 309 | mv $MGDATA/${bname}.text.idx $MGDATA/${oldname}.text.idx
|
|---|
| 310 | else
|
|---|
| 311 | # Need this file if building weights
|
|---|
| 312 | cp -d $MGDATA/${bname}.text.idx $MGDATA/${mergename}.text.idx
|
|---|
| 313 | endif
|
|---|
| 314 | mv $MGDATA/${bname}.invf $MGDATA/${oldname}.invf
|
|---|
| 315 | mv $MGDATA/${bname}.invf.idx $MGDATA/${oldname}.invf.idx
|
|---|
| 316 | mv $MGDATA/${bname}.invf.dict $MGDATA/${oldname}.invf.dict
|
|---|
| 317 | # the old weight file may be used for the new file
|
|---|
| 318 | mv $MGDATA/${bname}.weight $MGDATA/${mergename}.weight
|
|---|
| 319 | # EOC: the old invf.paragraph file may be used for the new file for Level 3 Invf
|
|---|
| 320 | if (-e $MGDATA/${bname}.invf.paragraph) then
|
|---|
| 321 | mv $MGDATA/${bname}.invf.paragraph $MGDATA/${mergename}.invf.paragraph
|
|---|
| 322 | endif
|
|---|
| 323 | echo "---------------------------------------"
|
|---|
| 324 |
|
|---|
| 325 | # [RPAP - Feb 97: Level 3 Merge]
|
|---|
| 326 | # Only merge text if doing text pass
|
|---|
| 327 | if ($text_pass) then
|
|---|
| 328 | echo ">> mg_text_merge -f ${mergename}"
|
|---|
| 329 | $bindir/mg_text_merge -f $mergename
|
|---|
| 330 | echo "---------------------------------------"
|
|---|
| 331 | endif
|
|---|
| 332 |
|
|---|
| 333 | echo ">> mg_invf_merge ${slow_merge}${guess_weights} -f ${mergename}"
|
|---|
| 334 | $bindir/mg_invf_merge ${slow_merge}${guess_weights} -f ${mergename}
|
|---|
| 335 | echo "---------------------------------------"
|
|---|
| 336 |
|
|---|
| 337 | #remove .weight file, so a new one will be created by mg_weights_build
|
|---|
| 338 | if ("$guess_weights" == "") then
|
|---|
| 339 | echo "rm ${mergename}.weight"
|
|---|
| 340 | rm $MGDATA/${mergename}.weight
|
|---|
| 341 | endif
|
|---|
| 342 |
|
|---|
| 343 | # mv $MGDATA/${mergename}.weight $MGDATA/${mergename}.w2
|
|---|
| 344 |
|
|---|
| 345 | echo ">> mg_weights_build -f ${mergename} -b ${weight_bits}"
|
|---|
| 346 | $bindir/mg_weights_build -f ${mergename} -b ${weight_bits}
|
|---|
| 347 | if ("$status" != "0") exit 1
|
|---|
| 348 |
|
|---|
| 349 | echo "---------------------------------------"
|
|---|
| 350 |
|
|---|
| 351 | echo ">> mg_invf_dict -f ${mergename} -b 4096"
|
|---|
| 352 | $bindir/mg_invf_dict -f ${mergename} -b 4096
|
|---|
| 353 |
|
|---|
| 354 | # [RPAP - Jan 97: Stem Index Change]
|
|---|
| 355 | if ($do_indexes) then
|
|---|
| 356 | echo "-----------------------------------"
|
|---|
| 357 |
|
|---|
| 358 | echo "mg_stem_idx -f ${mergename} -b 4096 -s1"
|
|---|
| 359 | $bindir/mg_stem_idx -f ${mergename} -b 4096 -s1
|
|---|
| 360 | if ("$status" != "0") exit 1
|
|---|
| 361 |
|
|---|
| 362 | echo ""
|
|---|
| 363 |
|
|---|
| 364 | echo "mg_stem_idx -f ${mergename} -b 4096 -s2"
|
|---|
| 365 | $bindir/mg_stem_idx -f ${mergename} -b 4096 -s2
|
|---|
| 366 | if ("$status" != "0") exit 1
|
|---|
| 367 |
|
|---|
| 368 | echo ""
|
|---|
| 369 |
|
|---|
| 370 | echo "mg_stem_idx -f ${mergename} -b 4096 -s3"
|
|---|
| 371 | $bindir/mg_stem_idx -f ${mergename} -b 4096 -s3
|
|---|
| 372 | if ("$status" != "0") exit 1
|
|---|
| 373 | endif
|
|---|
| 374 |
|
|---|
| 375 | echo "--------------------------------------------------------------"
|
|---|
| 376 |
|
|---|
| 377 | # move files back to base dir
|
|---|
| 378 | # [RPAP - Feb 97: Level 3 Merge]
|
|---|
| 379 | # mv the text files if doing text pass
|
|---|
| 380 | if ($text_pass) then
|
|---|
| 381 | mv $MGDATA/${oldname}.text $MGDATA/${bname}.text
|
|---|
| 382 | mv $MGDATA/${mergename}.text.idx $MGDATA/${bname}.text.idx
|
|---|
| 383 | mv $MGDATA/${newname}.text.dict $MGDATA/${bname}.text.dict
|
|---|
| 384 | endif
|
|---|
| 385 | mv $MGDATA/${mergename}.invf $MGDATA/${bname}.invf
|
|---|
| 386 | mv $MGDATA/${mergename}.invf.idx $MGDATA/${bname}.invf.idx
|
|---|
| 387 | mv $MGDATA/${mergename}.invf.dict $MGDATA/${bname}.invf.dict
|
|---|
| 388 | mv $MGDATA/${mergename}.invf.dict.blocked $MGDATA/${bname}.invf.dict.blocked
|
|---|
| 389 |
|
|---|
| 390 | # [RPAP - Jan 97: Stem Index Change]
|
|---|
| 391 | if ($do_indexes) then
|
|---|
| 392 | mv $MGDATA/${mergename}.invf.dict.blocked.1 $MGDATA/${bname}.invf.dict.blocked.1
|
|---|
| 393 | mv $MGDATA/${mergename}.invf.dict.blocked.2 $MGDATA/${bname}.invf.dict.blocked.2
|
|---|
| 394 | mv $MGDATA/${mergename}.invf.dict.blocked.3 $MGDATA/${bname}.invf.dict.blocked.3
|
|---|
| 395 | endif
|
|---|
| 396 |
|
|---|
| 397 | mv $MGDATA/${mergename}.weight $MGDATA/${bname}.weight
|
|---|
| 398 | mv $MGDATA/${mergename}.weight.approx $MGDATA/${bname}.weight.approx
|
|---|
| 399 | mv $MGDATA/${mergename}.text.idx.wgt $MGDATA/${bname}.text.idx.wgt
|
|---|
| 400 |
|
|---|
| 401 | # EOC: mv .invf.paragraph file if exists
|
|---|
| 402 | if (-e $MGDATA/${mergename}.invf.paragraph) then
|
|---|
| 403 | mv $MGDATA/${mergename}.invf.paragraph $MGDATA/${bname}.invf.paragraph
|
|---|
| 404 | endif
|
|---|
| 405 |
|
|---|
| 406 | # remove old files in the MERGE directory
|
|---|
| 407 | rm -f $MGDATA/${merge}/*
|
|---|
| 408 |
|
|---|
| 409 | echo "mgstat -f ${bname} -E"
|
|---|
| 410 | $bindir/mgstat -f ${bname} -E
|
|---|
| 411 | if ("$status" != "0") exit 1
|
|---|
| 412 |
|
|---|
| 413 | echo "--------------------------------------------------------------"
|
|---|
| 414 | echo "`uname -n`, `date`"
|
|---|
| 415 | echo "--------------------------------------------------------------"
|
|---|
| 416 |
|
|---|
| 417 | echo ""
|
|---|
