[3365] | 1 | /**************************************************************************
|
---|
| 2 | *
|
---|
| 3 | * mgpp_passes.cpp -- Driver for the various passes
|
---|
| 4 | * Copyright (C) 1994 Neil Sharman
|
---|
| 5 | * Copyright (C) 1999 Rodger McNab
|
---|
| 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 | #define _XOPEN_SOURCE 1
|
---|
| 24 | #define _XOPEN_SOURCE_EXTENDED 1
|
---|
| 25 |
|
---|
[8692] | 26 | #if defined(WIN32) || defined(_WIN32) || defined(__WIN32__)
|
---|
| 27 | #pragma warning(disable:4786)
|
---|
| 28 | #endif
|
---|
| 29 |
|
---|
[3365] | 30 | // need this to avoid bizarre compiler problems under VC++ 6.0
|
---|
| 31 | #if defined (__WIN32__) && !defined (GSDL_USE_IOS_H)
|
---|
| 32 | # include <iostream>
|
---|
| 33 | #endif
|
---|
| 34 |
|
---|
| 35 | #include "sysfuncs.h"
|
---|
| 36 |
|
---|
| 37 | #ifdef HAVE_MALLINFO
|
---|
| 38 | # include <malloc.h>
|
---|
| 39 | #endif
|
---|
| 40 |
|
---|
| 41 | #if defined __WIN32__
|
---|
| 42 | # include <io.h>
|
---|
| 43 | # include "getopt_old.h"
|
---|
| 44 | # define close _close
|
---|
| 45 | # define open _open
|
---|
| 46 | #elif defined __CYGWIN__
|
---|
| 47 | #include "getopt_old.h"
|
---|
| 48 | #else
|
---|
| 49 | # include <unistd.h>
|
---|
| 50 | #endif
|
---|
| 51 |
|
---|
| 52 | #include "memlib.h"
|
---|
| 53 | #include "messages.h"
|
---|
| 54 | #include "longlong.h"
|
---|
| 55 | #include "mg_files.h"
|
---|
| 56 | #include "mg.h"
|
---|
| 57 | #include "build.h"
|
---|
| 58 | #include "text.h"
|
---|
| 59 | #include "stemmer.h"
|
---|
| 60 | #include "FileBuf.h"
|
---|
| 61 | #include "TextEl.h"
|
---|
| 62 | #include "TagInfo.h"
|
---|
[12321] | 63 | #include "words.h"
|
---|
[3365] | 64 | #define MAX_PASSES 5
|
---|
| 65 |
|
---|
| 66 | #define SPECIAL 1
|
---|
| 67 | #define TEXT_PASS_1 2
|
---|
| 68 | #define TEXT_PASS_2 4
|
---|
| 69 | #define IVF_PASS_1 8
|
---|
| 70 | #define IVF_PASS_2 16
|
---|
| 71 |
|
---|
| 72 | #define MIN_BUF 8192
|
---|
| 73 |
|
---|
| 74 |
|
---|
| 75 | unsigned long invf_buffer_size = 5 * 1024 * 1024; /* 5Mb */
|
---|
| 76 |
|
---|
| 77 | static char Passes = 0;
|
---|
| 78 | static char **files = NULL;
|
---|
| 79 | static int num_files = 0;
|
---|
| 80 |
|
---|
| 81 |
|
---|
| 82 | struct pass_data {
|
---|
| 83 | char *name;
|
---|
| 84 | int (*init) (const TagInfo &tagInfo, char *);
|
---|
| 85 | int (*process) (const TagInfo &tagInfo, const TextElArray &doc);
|
---|
| 86 | int (*done) (const TagInfo &tagInfo, char *);
|
---|
| 87 | };
|
---|
| 88 |
|
---|
| 89 |
|
---|
| 90 | static pass_data PassData[MAX_PASSES] = {
|
---|
| 91 | {"special", init_special, process_special, done_special},
|
---|
| 92 | {"text.pass1", init_text_1, process_text_1, done_text_1},
|
---|
| 93 | {"text.pass2", init_text_2, process_text_2, done_text_2},
|
---|
| 94 | {"ivf.pass1", init_ivf_1, process_ivf_1, done_ivf_1},
|
---|
| 95 | {"ivf.pass2", init_ivf_2, process_ivf_2, done_ivf_2},
|
---|
| 96 | };
|
---|
| 97 |
|
---|
| 98 | static char *usage_str = "\nUSAGE:\n"
|
---|
| 99 | " %s [-J doc-tag] [-K level-tag] [-L index-level]\n"
|
---|
| 100 | " %*s [-m invf-memory] [-T1] [-T2] [-I1] [-I2] [-S]\n"
|
---|
[12321] | 101 | " %*s [-C] [-h] [-d directory] [-M maxnumeric] -f name\n\n";
|
---|
[3365] | 102 |
|
---|
| 103 |
|
---|
| 104 |
|
---|
| 105 | static void Usage (char *err) {
|
---|
| 106 | if (err) Message (err);
|
---|
| 107 | fprintf (stderr, usage_str, msg_prefix, strlen (msg_prefix), "",
|
---|
| 108 | strlen (msg_prefix), "");
|
---|
| 109 | exit (1);
|
---|
| 110 | }
|
---|
| 111 |
|
---|
| 112 |
|
---|
| 113 |
|
---|
| 114 |
|
---|
| 115 | int OpenNextFile (int in_fd) {
|
---|
| 116 | if (in_fd > 0) close (in_fd);
|
---|
| 117 | if (num_files == 0) return (-1);
|
---|
| 118 | if ((in_fd = open (files[0], O_RDONLY)) == -1)
|
---|
| 119 | FatalError (1, "Cannot open %s", files[0]);
|
---|
[9613] | 120 | ++files;
|
---|
| 121 | --num_files;
|
---|
[3365] | 122 | return (in_fd);
|
---|
| 123 | }
|
---|
| 124 |
|
---|
| 125 |
|
---|
| 126 | static void Driver (int in_fd, char *file_name,
|
---|
| 127 | const TagInfo &tagInfo, bool compatMode) {
|
---|
| 128 | // cout << tagInfo;
|
---|
| 129 |
|
---|
| 130 | int pass;
|
---|
| 131 |
|
---|
| 132 | unsigned long numBytes = 0;
|
---|
| 133 | unsigned long numDocs = 0;
|
---|
| 134 |
|
---|
| 135 | // initialise all the passes
|
---|
[9613] | 136 | for (pass = 0; pass < MAX_PASSES; ++pass) {
|
---|
[3365] | 137 | if (Passes & (1 << pass)) {
|
---|
| 138 | if (PassData[pass].init (tagInfo, file_name) == COMPERROR)
|
---|
| 139 | FatalError (1, "Error during init of \"%s\"", PassData[pass].name);
|
---|
| 140 | }
|
---|
| 141 | }
|
---|
| 142 |
|
---|
| 143 |
|
---|
| 144 | // set up various variables
|
---|
| 145 | FileBuf buf;
|
---|
| 146 | TextElArray doc;
|
---|
| 147 | unsigned long docLen = 0;
|
---|
| 148 |
|
---|
| 149 | // read and process each file (start with an open file)
|
---|
| 150 | do {
|
---|
| 151 |
|
---|
| 152 | // read and process each document in this file
|
---|
| 153 | buf.SetFD (in_fd);
|
---|
| 154 | while (ReadDoc (buf, tagInfo.docTag, doc, docLen, compatMode)) {
|
---|
| 155 |
|
---|
| 156 | // give this document to each pass
|
---|
[9613] | 157 | for (pass = 0; pass < MAX_PASSES; ++pass) {
|
---|
[3365] | 158 | if (Passes & (1 << pass)) {
|
---|
| 159 | if (PassData[pass].process (tagInfo, doc) == COMPERROR)
|
---|
| 160 | FatalError(1,"Error during processing of \"%s\"",PassData[pass].name);
|
---|
| 161 | }
|
---|
| 162 | }
|
---|
| 163 |
|
---|
| 164 | // another document has been processed
|
---|
| 165 | numBytes += docLen;
|
---|
[9613] | 166 | ++numDocs;
|
---|
[3365] | 167 | }
|
---|
| 168 |
|
---|
| 169 | } while ((in_fd = OpenNextFile (in_fd)) > 0);
|
---|
| 170 |
|
---|
| 171 |
|
---|
| 172 | // do done for each pass
|
---|
[9613] | 173 | for (pass = 0; pass < MAX_PASSES; ++pass) {
|
---|
[3365] | 174 | if (Passes & (1 << pass)) {
|
---|
| 175 | if (PassData[pass].done (tagInfo, file_name) == COMPERROR)
|
---|
| 176 | FatalError (1, "Error during done of \"%s\"", PassData[pass].name);
|
---|
| 177 | }
|
---|
| 178 | }
|
---|
| 179 | }
|
---|
| 180 |
|
---|
| 181 | int main (int argc, char **argv) {
|
---|
[12321] | 182 | int ch, in_fd, maxnum;
|
---|
[3365] | 183 | char *filename = NULL;
|
---|
| 184 | bool compatMode = false;
|
---|
| 185 | TagInfo tagInfo;
|
---|
| 186 | tagInfo.SetDocTag ("Document");
|
---|
| 187 |
|
---|
| 188 | msg_prefix = argv[0];
|
---|
| 189 |
|
---|
| 190 | opterr = 0;
|
---|
[12321] | 191 | while ((ch=getopt(argc, argv, "J:K:L:M:f:d:m:I:T:SCh"))!=-1){
|
---|
[3365] | 192 | switch (ch) {
|
---|
| 193 | case 'J':
|
---|
| 194 | tagInfo.SetDocTag (optarg);
|
---|
| 195 | break;
|
---|
| 196 | case 'K':
|
---|
| 197 | tagInfo.AddLevelTag (optarg);
|
---|
| 198 | break;
|
---|
| 199 | case 'L':
|
---|
| 200 | tagInfo.SetIndexLevel (optarg);
|
---|
| 201 | break;
|
---|
[12321] | 202 | case 'M':
|
---|
| 203 | maxnum = atoi(optarg);
|
---|
| 204 | if (4 < maxnum < 512) {
|
---|
| 205 | MAXNUMERIC = maxnum;
|
---|
| 206 | }
|
---|
| 207 | break;
|
---|
[3365] | 208 | case 'f':
|
---|
| 209 | filename = optarg;
|
---|
| 210 | break;
|
---|
| 211 | case 'd':
|
---|
| 212 | set_basepath (optarg);
|
---|
| 213 | break;
|
---|
| 214 | case 'm':
|
---|
| 215 | invf_buffer_size = (int) (atof (optarg) * 1024 * 1024);
|
---|
| 216 | break;
|
---|
| 217 | case 'I':
|
---|
| 218 | if (*optarg == '1')
|
---|
| 219 | Passes |= IVF_PASS_1;
|
---|
| 220 | else if (*optarg == '2')
|
---|
| 221 | Passes |= IVF_PASS_2;
|
---|
| 222 | else
|
---|
| 223 | Usage ("Invalid pass number");
|
---|
| 224 | break;
|
---|
| 225 | case 'T':
|
---|
| 226 | if (*optarg == '1')
|
---|
| 227 | Passes |= TEXT_PASS_1;
|
---|
| 228 | else if (*optarg == '2')
|
---|
| 229 | Passes |= TEXT_PASS_2;
|
---|
| 230 | else
|
---|
| 231 | Usage ("Invalid pass number");
|
---|
| 232 | break;
|
---|
| 233 | case 'S':
|
---|
| 234 | Passes |= SPECIAL;
|
---|
| 235 | break;
|
---|
| 236 | case 'C':
|
---|
| 237 | compatMode = true;
|
---|
| 238 | break;
|
---|
| 239 | case 'h':
|
---|
| 240 | case '?':
|
---|
| 241 | Usage (NULL);
|
---|
| 242 | }
|
---|
| 243 | }
|
---|
| 244 |
|
---|
| 245 | if (!filename || *filename == '\0')
|
---|
| 246 | FatalError (1, "A document collection name must be specified.");
|
---|
| 247 |
|
---|
| 248 | if ((Passes & (IVF_PASS_1 | IVF_PASS_2)) == (IVF_PASS_1 | IVF_PASS_2))
|
---|
| 249 | FatalError (1, "I1 and I2 cannot be done simultaneously.");
|
---|
| 250 |
|
---|
| 251 | if ((Passes & (TEXT_PASS_1 | TEXT_PASS_2)) == (TEXT_PASS_1 | TEXT_PASS_2))
|
---|
| 252 | FatalError (1, "T1 and T2 cannot be done simultaneously.");
|
---|
| 253 |
|
---|
| 254 | if (!Passes)
|
---|
| 255 | FatalError (1, "S, T1, T2, I1 or I2 must be specified.");
|
---|
| 256 |
|
---|
| 257 | if (optind < argc) {
|
---|
| 258 | if ((in_fd = open (argv[optind], O_RDONLY)) == -1)
|
---|
| 259 | FatalError (1, "Cannot open %s", argv[optind]);
|
---|
| 260 | files = &argv[optind + 1];
|
---|
| 261 | num_files = argc - (optind + 1);
|
---|
| 262 |
|
---|
| 263 | } else in_fd = 0; // stdin
|
---|
| 264 |
|
---|
| 265 |
|
---|
| 266 | if (compatMode) tagInfo.SetDocTag ("Document");
|
---|
| 267 |
|
---|
| 268 | // a document tag is also a level tag
|
---|
| 269 | tagInfo.levelTags.insert (tagInfo.docTag);
|
---|
| 270 |
|
---|
| 271 | Driver (in_fd, filename, tagInfo, compatMode);
|
---|
| 272 |
|
---|
| 273 | return (0);
|
---|
| 274 | }
|
---|