[3365] | 1 | .\"------------------------------------------------------------
|
---|
| 2 | .\" Id - set Rv,revision, and Dt, Date using rcs-Id tag.
|
---|
| 3 | .de Id
|
---|
| 4 | .ds Rv \\$3
|
---|
| 5 | .ds Dt \\$4
|
---|
| 6 | ..
|
---|
| 7 | .\"------------------------------------------------------------
|
---|
| 8 | .TH mgpp_passes 1 \*(Dt CITRI
|
---|
| 9 | .SH NAME
|
---|
| 10 | mgpp_passes \- builds mgpp databases
|
---|
| 11 | .SH SYNOPSIS
|
---|
| 12 | .B mgpp_passes
|
---|
| 13 | [
|
---|
| 14 | .BI \-J " doc-tag"
|
---|
| 15 | ]
|
---|
| 16 | [
|
---|
| 17 | .BI \-K " level-tag"
|
---|
| 18 | ]
|
---|
| 19 | .if n .ti +10n
|
---|
| 20 | [
|
---|
| 21 | .BI \-L " index-level"
|
---|
| 22 | ]
|
---|
| 23 | [
|
---|
| 24 | .BI \-m " invf-mem-buffer"
|
---|
| 25 | ]
|
---|
| 26 | .if n .ti +10n
|
---|
| 27 | [
|
---|
| 28 | .B \-T1
|
---|
| 29 | ]
|
---|
| 30 | [
|
---|
| 31 | .B \-T2
|
---|
| 32 | ]
|
---|
| 33 | [
|
---|
| 34 | .B \-I1
|
---|
| 35 | ]
|
---|
| 36 | [
|
---|
| 37 | .B \-I2
|
---|
| 38 | ]
|
---|
| 39 | [
|
---|
| 40 | .B \-S
|
---|
| 41 | ]
|
---|
| 42 | [
|
---|
| 43 | .B \-C
|
---|
| 44 | ]
|
---|
| 45 | .if n .ti +10n
|
---|
| 46 | [
|
---|
| 47 | .B \-h
|
---|
| 48 | ]
|
---|
| 49 | [
|
---|
| 50 | .BI \-d " directory"
|
---|
| 51 | ]
|
---|
| 52 | .BI \-f " name"
|
---|
| 53 | [
|
---|
| 54 | .I filename(s)
|
---|
| 55 | ]
|
---|
| 56 | .SH DESCRIPTION
|
---|
| 57 | .B mgpp_passes
|
---|
| 58 | is the program that does most of the work when building mgpp
|
---|
| 59 | database systems. The input documents can come from either
|
---|
| 60 | .I stdin
|
---|
| 61 | or from a list of files on the command line. In general,
|
---|
| 62 | .B mgpp_passes
|
---|
| 63 | must be run twice to build a database, first with the
|
---|
| 64 | .B \-T1
|
---|
| 65 | and
|
---|
| 66 | .B \-I1
|
---|
| 67 | options, and second with the
|
---|
| 68 | .B \-T2
|
---|
| 69 | and
|
---|
| 70 | .B \-I2
|
---|
| 71 | options. Several other programs must be run in order to get an
|
---|
| 72 | mgpp database. The
|
---|
| 73 | .SB EXAMPLE
|
---|
| 74 | section below gives an example of how to build a complete
|
---|
| 75 | mgpp database.
|
---|
| 76 | .SH OPTIONS
|
---|
| 77 | Options may appear in any order, but the
|
---|
| 78 | .IR filename(s) ,
|
---|
| 79 | if specified, must be last.
|
---|
| 80 | .TP "\w'\fB\-C\fP \fIcompstatpointt\fP'u+2n"
|
---|
| 81 | .BI \-J " doc-tag"
|
---|
| 82 | Specifies the SGML tag that encloses each document. Text appearing
|
---|
| 83 | outside this tag is ignored. The document tag defines the highest
|
---|
| 84 | level document that can be queried and printed. The default document
|
---|
| 85 | tag is 'Document'.
|
---|
| 86 | .TP
|
---|
| 87 | .BI \-K " level-tag"
|
---|
| 88 | Specifies the SGML tag of a sub document level. A level tag must
|
---|
| 89 | enclose all text enclosed by the document tag. Levels can be
|
---|
| 90 | queried and printed as if they were separate documents. Multiple
|
---|
| 91 | document levels can be specified (the document tag is always
|
---|
| 92 | added as a document level).
|
---|
| 93 | .TP
|
---|
| 94 | .BI \-L " index-level"
|
---|
| 95 | Specifies the SGML tag enclosing the smallest indexed element. The
|
---|
| 96 | index level should be no larger than the smallest document
|
---|
| 97 | level. An empty string can be used to specify a word level index
|
---|
| 98 | (which is the default).
|
---|
| 99 | .TP
|
---|
| 100 | .BI \-m " invf-mem-buffer"
|
---|
| 101 | Maximum amount of memory to use for the pass-2 file inversion in
|
---|
| 102 | megabytes. This option is only useful when used in conjunction with
|
---|
| 103 | the option
|
---|
| 104 | .BR \-I1 .
|
---|
| 105 | The larger this value, the faster the pass-2 inversion will proceed.
|
---|
| 106 | The default value is 5 MB.
|
---|
| 107 | .TP
|
---|
| 108 | .B \-T1
|
---|
| 109 | Generate the
|
---|
| 110 | .I *.text.stats
|
---|
| 111 | file.
|
---|
| 112 | .TP
|
---|
| 113 | .B \-T2
|
---|
| 114 | Generate the
|
---|
| 115 | .IR *.text ,
|
---|
| 116 | .IR *.text.idx ,
|
---|
| 117 | .IR *.text.level ,
|
---|
| 118 | and possibly the
|
---|
| 119 | .I *.text.dict.aux
|
---|
| 120 | files. Using this option requires that the
|
---|
| 121 | .I *.text.dict
|
---|
| 122 | file be present.
|
---|
| 123 | .TP
|
---|
| 124 | .B \-I1
|
---|
| 125 | Generate the
|
---|
| 126 | .IR *.invf.dict ,
|
---|
| 127 | .IR *.invf.level ,
|
---|
| 128 | .IR *.invf.chunk ,
|
---|
| 129 | and
|
---|
| 130 | .I *.invf.chunk.trans
|
---|
| 131 | files.
|
---|
| 132 | .TP
|
---|
| 133 | .B \-I2
|
---|
| 134 | Generate the
|
---|
| 135 | .I *.invf
|
---|
| 136 | and
|
---|
| 137 | .I *.invf.idx
|
---|
| 138 | files. Using this option requires
|
---|
| 139 | that the
|
---|
| 140 | .IR *.invf.dict.hash ,
|
---|
| 141 | .IR *.invf.level ,
|
---|
| 142 | .IR *.invf.chunk ,
|
---|
| 143 | and
|
---|
| 144 | .I *.invf.chunk.trans
|
---|
| 145 | files be present. The
|
---|
| 146 | .I *.invf.dict.hash
|
---|
| 147 | file is generated by
|
---|
| 148 | .BR mgpp_perf_hash_build (1)
|
---|
| 149 | from the
|
---|
| 150 | .I *.invf.dict
|
---|
| 151 | file.
|
---|
| 152 | .TP
|
---|
| 153 | .B \-S
|
---|
| 154 | This option causes a special pass to be executed. It is up to a user
|
---|
| 155 | to modify
|
---|
| 156 | .I mg.special.c
|
---|
| 157 | in the source code to do something with the documents it is given.
|
---|
| 158 | .TP
|
---|
| 159 | .B \-C
|
---|
| 160 | This activates the compatibility parsing mode. When using this
|
---|
| 161 | mode documents are separated by control-B and paragraphs are separated
|
---|
| 162 | by control-C. Internally these are converted to documents surrounded
|
---|
| 163 | by 'Document' tags and paragraphs surrounded by 'Paragraph' tags.
|
---|
| 164 | .TP
|
---|
| 165 | .B \-h
|
---|
| 166 | This displays a usage line on
|
---|
| 167 | .IR stderr .
|
---|
| 168 | .TP
|
---|
| 169 | .BI \-d " directory"
|
---|
| 170 | This specifies the directory where the document collection is to be
|
---|
| 171 | written.
|
---|
| 172 | .TP
|
---|
| 173 | .BI \-f " name"
|
---|
| 174 | This specifies the base name of the document collection that will be
|
---|
| 175 | created.
|
---|
| 176 | .TP
|
---|
| 177 | .I filename(s)
|
---|
| 178 | This specifies the source text. If this is not specified, then the
|
---|
| 179 | program expects the source text from
|
---|
| 180 | .IR stdin .
|
---|
| 181 | .SH EXAMPLE
|
---|
| 182 | What follows is a UNIX
|
---|
| 183 | .BR csh (1)
|
---|
| 184 | script as an example of how to build an mgpp document collection.
|
---|
| 185 | .LP
|
---|
| 186 | .nf
|
---|
| 187 | .DT
|
---|
| 188 | .ft B
|
---|
| 189 | .I #! /bin/csh
|
---|
| 190 | .I
|
---|
| 191 | # The first argument on the command line specifies the
|
---|
| 192 | .I
|
---|
| 193 | # source of the text
|
---|
| 194 | set source = ($1)
|
---|
| 195 | .PP
|
---|
| 196 | .I
|
---|
| 197 | # The second argument is the name of the collection
|
---|
| 198 | set text = ($2)
|
---|
| 199 | .PP
|
---|
| 200 | .I
|
---|
| 201 | # Create *.text.stats, *.invf.dict, *.invf.level
|
---|
| 202 | .I
|
---|
| 203 | # *.invf.chunk and *.invf.chunks.trans
|
---|
| 204 | ${source} | mgpp_passes -T1 -I1 -f ${text}
|
---|
| 205 | .PP
|
---|
| 206 | .I
|
---|
| 207 | # Create *.text.dict
|
---|
| 208 | mgpp_compression_dict -f ${text}
|
---|
| 209 | .PP
|
---|
| 210 | .I
|
---|
| 211 | # Create *.invf.dict.hash
|
---|
| 212 | mgpp_perf_hash_build -f ${text}
|
---|
| 213 | .PP
|
---|
| 214 | .I
|
---|
| 215 | # Create *.text, *.text.idx, *.text.level
|
---|
| 216 | .I
|
---|
| 217 | # *.invf and *.invf.idx
|
---|
| 218 | ${source} | mgpp_passes -T2 -I2 -f ${text}
|
---|
| 219 | .PP
|
---|
| 220 | .I
|
---|
| 221 | # Create *.text.weight and *.weight.approx
|
---|
| 222 | mgpp_weights_build -f ${text}
|
---|
| 223 | .PP
|
---|
| 224 | .I
|
---|
| 225 | # Create *.invf.dict.blocked
|
---|
| 226 | mgpp_invf_dict -f ${text}
|
---|
| 227 | .PP
|
---|
| 228 | .I
|
---|
| 229 | # Create *.invf.dict.blocked.1
|
---|
| 230 | mgpp_stem_idx -s 1 -f ${text}
|
---|
| 231 | .PP
|
---|
| 232 | .I
|
---|
| 233 | # Create *.invf.dict.blocked.2
|
---|
| 234 | mgpp_stem_idx -s 2 -f ${text}
|
---|
| 235 | .PP
|
---|
| 236 | .I
|
---|
| 237 | # Create *.invf.dict.blocked.3
|
---|
| 238 | mgpp_stem_idx -s 3 -f ${text}
|
---|
| 239 | .PP
|
---|
| 240 | .I
|
---|
| 241 | # Create *.text.dict.fast
|
---|
| 242 | mgpp_fast_comp_dict -f ${text}
|
---|
| 243 | .ft R
|
---|
| 244 | .fi
|
---|
| 245 | .SH ENVIRONMENT
|
---|
| 246 | .TP "\w'\fBMGDATA\fP'u+2n"
|
---|
| 247 | .SB MGDATA
|
---|
| 248 | If this environment variable exists, then its value is used as the
|
---|
| 249 | default directory where the mgpp
|
---|
| 250 | collection files are. If this variable does not exist, then the
|
---|
| 251 | directory \*(lq\fB.\fP\*(rq is used by default. The command line
|
---|
| 252 | option
|
---|
| 253 | .BI \-d " directory"
|
---|
| 254 | overrides the directory in
|
---|
| 255 | .BR MGDATA .
|
---|
| 256 | .SH FILES
|
---|
| 257 | .TP 22
|
---|
| 258 | .B *.invf
|
---|
| 259 | Inverted file.
|
---|
| 260 | .TP
|
---|
| 261 | .B *.invf.chunk
|
---|
| 262 | Inverted file chunk descriptor file. When the inverted file is
|
---|
| 263 | created it is created in chunks that use no more than a set amount of
|
---|
| 264 | memory. This file describes those chunks.
|
---|
| 265 | .TP
|
---|
| 266 | .B *.invf.chunk.trans
|
---|
| 267 | Word-occurrence-order to lexical-order translation file. The
|
---|
| 268 | .B *.invf.chunk
|
---|
| 269 | file is written in word-occurrence order but is required by
|
---|
| 270 | .B \-I2
|
---|
| 271 | to be in lexical order.
|
---|
| 272 | .TP
|
---|
| 273 | .B *.invf.dict
|
---|
| 274 | Compressed stemmed dictionary.
|
---|
| 275 | .TP
|
---|
| 276 | .B *.invf.dict.blocked
|
---|
| 277 | Compressed stemmed dictionary with index into the dictionary.
|
---|
| 278 | .TP
|
---|
| 279 | .B *.invf.dict.blocked.n
|
---|
| 280 | Transformation dictionary from words stemmed with method
|
---|
| 281 | .B n
|
---|
| 282 | to unstemmed words.
|
---|
| 283 | .TP
|
---|
| 284 | .B *.invf.dict.hash
|
---|
| 285 | Data for an order-preserving perfect hash function.
|
---|
| 286 | .TP
|
---|
| 287 | .B *.invf.idx
|
---|
| 288 | The index into the inverted file.
|
---|
| 289 | .TP
|
---|
| 290 | .B *.invf.level
|
---|
| 291 | Information about the document levels needed for querying.
|
---|
| 292 | .TP
|
---|
| 293 | .B *.text
|
---|
| 294 | Compressed text.
|
---|
| 295 | .TP
|
---|
| 296 | .B *.text.dict
|
---|
| 297 | Compressed compression dictionary.
|
---|
| 298 | .TP
|
---|
| 299 | .B *.text.dict.fast
|
---|
| 300 | A fast loading version of the compressed compression dictionary.
|
---|
| 301 | .TP
|
---|
| 302 | .B *.text.idx
|
---|
| 303 | Index into the compressed documents.
|
---|
| 304 | .TP
|
---|
| 305 | .B *.text.level
|
---|
| 306 | Information about the document levels needed for text decompression.
|
---|
| 307 | .TP
|
---|
| 308 | .B *.text.stats
|
---|
| 309 | Statistics about the text.
|
---|
| 310 | .TP
|
---|
| 311 | .B *.weight
|
---|
| 312 | The exact weights file.
|
---|
| 313 | .TP
|
---|
| 314 | .B *.weight.approx
|
---|
| 315 | The approximate weights file.
|
---|
| 316 | .SH "SEE ALSO"
|
---|
| 317 | .na
|
---|
| 318 | .BR mgpp_compression_dict (1),
|
---|
| 319 | .BR mgpp_fast_comp_dict (1),
|
---|
| 320 | .BR mgpp_invf_dict (1),
|
---|
| 321 | .BR mgpp_perf_hash_build (1),
|
---|
| 322 | .BR mgpp_stem_idx (1),
|
---|
| 323 | .BR mgpp_weights_build (1)
|
---|