[3365] | 1 | /**************************************************************************
|
---|
| 2 | *
|
---|
| 3 | * Queryer.cpp -- simple interactive query program
|
---|
| 4 | * Copyright (C) 1999 Rodger McNab
|
---|
| 5 | *
|
---|
| 6 | * This program is free software; you can redistribute it and/or modify
|
---|
| 7 | * it under the terms of the GNU General Public License as published by
|
---|
| 8 | * the Free Software Foundation; either version 2 of the License, or
|
---|
| 9 | * (at your option) any later version.
|
---|
| 10 | *
|
---|
| 11 | * This program is distributed in the hope that it will be useful,
|
---|
| 12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
| 13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
| 14 | * GNU General Public License for more details.
|
---|
| 15 | *
|
---|
| 16 | * You should have received a copy of the GNU General Public License
|
---|
| 17 | * along with this program; if not, write to the Free Software
|
---|
| 18 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
| 19 | *
|
---|
| 20 | **************************************************************************/
|
---|
| 21 |
|
---|
| 22 | #define _XOPEN_SOURCE 1
|
---|
| 23 | #define _XOPEN_SOURCE_EXTENDED 1
|
---|
| 24 |
|
---|
| 25 | /* getopt is in posix.2, so cygwin should have it in unistd, but doesn't */
|
---|
| 26 | #if defined (__WIN32__) || defined (__CYGWIN__)
|
---|
| 27 | # include "getopt_old.h"
|
---|
| 28 | #else
|
---|
| 29 | # include <unistd.h>
|
---|
| 30 | #endif
|
---|
| 31 |
|
---|
| 32 | #include "MGQuery.h"
|
---|
| 33 | #include "TextGet.h"
|
---|
| 34 |
|
---|
| 35 | #include "messages.h"
|
---|
| 36 | #include "mg_files.h"
|
---|
| 37 |
|
---|
| 38 | #include "GSDLQueryParser.h"
|
---|
| 39 |
|
---|
| 40 | void printHelp() {
|
---|
| 41 |
|
---|
| 42 | cout << "commands available are:\n"
|
---|
| 43 | << "\t.q\t\tquit\n"
|
---|
| 44 | << "\t.h\t\tprint the help message\n"
|
---|
| 45 | << "\t.i\t\tchange the search level (enter the new level at the prompt)\n"
|
---|
| 46 | << "\t.l\t\tchange the result level ( \"\" \"\" )\n"
|
---|
| 47 | << "\t.b\t\tfull text browse (enter a word or fragment at the prompt)\n"
|
---|
| 48 | << "\t.r0/.r1\t\tranking off/on\n"
|
---|
| 49 | << "\t.t0/.t1\t\tquery type some/all\n"
|
---|
| 50 | << "\t.c0/.c1\t\tcasefolding off/on\n"
|
---|
| 51 | << "\t.s0/.s1\t\tstemming off/on\n"
|
---|
[13477] | 52 | #ifdef ENABLE_ACCENTFOLD
|
---|
| 53 | << "\t.a0/.a1\t\taccentfolding off/on\n"
|
---|
| 54 | #endif
|
---|
[12321] | 55 | << "\t.o0/.o1\t\tshort output off/on\n"
|
---|
| 56 | << "\t.m\t\tset maxnumeric (enter the number at the prompt)\n\n"
|
---|
[3365] | 57 | << "\t.p\t\tprint a document (enter the docnum at the prompt)\n"
|
---|
| 58 | << "otherwise just enter a query\n\n";
|
---|
| 59 |
|
---|
| 60 | }
|
---|
| 61 |
|
---|
| 62 | int main (int argc, char **argv) {
|
---|
| 63 | int ch;
|
---|
| 64 | char *textfilename = "";
|
---|
| 65 | char *indexfilename = "";
|
---|
| 66 | char *basePath = "";
|
---|
| 67 |
|
---|
| 68 | opterr = 0;
|
---|
| 69 | msg_prefix = argv[0];
|
---|
| 70 |
|
---|
| 71 | // process the command line arguments
|
---|
| 72 | while ((ch = getopt (argc, argv, "f:t:d:h")) != -1) {
|
---|
| 73 | switch (ch) {
|
---|
| 74 | case 'f': /* input file */
|
---|
| 75 | indexfilename = optarg;
|
---|
| 76 | break;
|
---|
| 77 | case 't':
|
---|
| 78 | textfilename = optarg;
|
---|
| 79 | break;
|
---|
| 80 | case 'd':
|
---|
| 81 | basePath = optarg;
|
---|
| 82 | set_basepath (basePath);
|
---|
| 83 | break;
|
---|
| 84 | case 'h':
|
---|
| 85 | case '?':
|
---|
| 86 | fprintf (stderr, "usage: %s [-h] [-d directory] -f indexname -t textname\n", argv[0]);
|
---|
| 87 | exit (1);
|
---|
| 88 | }
|
---|
| 89 | }
|
---|
| 90 |
|
---|
| 91 | if (textfilename[0] == '\0' || indexfilename[0] == '\0') {
|
---|
| 92 | FatalError (1, "Index and text file names must be specified with -f and -t \n");
|
---|
| 93 | }
|
---|
| 94 |
|
---|
| 95 | // init the text system
|
---|
[13477] | 96 | TextData textData;
|
---|
[3365] | 97 | if (!textData.LoadData (basePath, textfilename)) {
|
---|
| 98 | FatalError (1, "Couldn't load text information for \"%s\"", textfilename);
|
---|
| 99 | }
|
---|
| 100 |
|
---|
| 101 | // init the query system
|
---|
| 102 | IndexData indexData;
|
---|
| 103 | if (!indexData.LoadData (basePath, indexfilename)) {
|
---|
| 104 | FatalError (1, "Couldn't load index information for \"%s\"", indexfilename);
|
---|
| 105 | }
|
---|
| 106 |
|
---|
[12321] | 107 | int maxnumeric = 4;
|
---|
| 108 |
|
---|
[3365] | 109 | // debug output
|
---|
| 110 | cerr << "num docs: "<<indexData.bdh.num_docs
|
---|
| 111 | << "\nnum frags: "<<indexData.bdh.num_frags
|
---|
| 112 | << "\nnum words: "<<indexData.bdh.num_words
|
---|
| 113 | << "\ntotal bytes: "<<indexData.bdh.total_bytes
|
---|
| 114 | << "\nindex string bytes: "<<indexData.bdh.index_string_bytes
|
---|
| 115 | << "\nnum levels: "<<indexData.bdh.num_levels<<endl;
|
---|
| 116 |
|
---|
| 117 | // do querying
|
---|
| 118 | QueryInfo queryInfo;
|
---|
[12321] | 119 | SetCStr (queryInfo.docLevel, "Doc", 3);
|
---|
[3365] | 120 | queryInfo.maxDocs = 50;
|
---|
| 121 | queryInfo.sortByRank = true;
|
---|
| 122 | queryInfo.exactWeights = false;
|
---|
| 123 | queryInfo.needRankInfo = true;
|
---|
| 124 | queryInfo.needTermFreqs = true;
|
---|
| 125 |
|
---|
| 126 | ExtQueryResult queryResult;
|
---|
| 127 | char query[2048];
|
---|
| 128 | UCArray queryArray;
|
---|
| 129 | QueryNode *queryTree = NULL;
|
---|
| 130 |
|
---|
| 131 |
|
---|
| 132 | UCArray docLevel;
|
---|
[12321] | 133 | SetCStr(docLevel, "Doc", 3);
|
---|
[3365] | 134 |
|
---|
| 135 | UCArray level;
|
---|
| 136 | UCArrayClear(level);
|
---|
| 137 | //SetCStr(level, "");
|
---|
| 138 |
|
---|
[13477] | 139 | int defaultStemMethod = 0; // uncasefolded, unstemmed, unaccentfolded
|
---|
[3365] | 140 | int defaultBoolCombine = 0; // OR
|
---|
| 141 | bool shortOutput = false;
|
---|
| 142 | BrowseQueryNode browseNode;
|
---|
| 143 | browseNode.startPosition = -10;
|
---|
| 144 | browseNode.numTerms = 40;
|
---|
| 145 |
|
---|
| 146 | BrowseQueryResult browseResult;
|
---|
| 147 |
|
---|
| 148 | while (true) {
|
---|
| 149 | cout << "> ";
|
---|
| 150 | cin.getline(query, 2048, '\n');
|
---|
[8692] | 151 | SetCStr (queryArray, query, strlen(query));
|
---|
[3365] | 152 |
|
---|
| 153 | // check for commands
|
---|
| 154 | if (queryArray.size() >= 2 && queryArray[0] == '.') {
|
---|
| 155 | if (queryArray[1] == 'q') break; // quit
|
---|
| 156 |
|
---|
| 157 | if (queryArray[1] == 'h') { // help
|
---|
| 158 | printHelp();
|
---|
[6116] | 159 | } else if (queryArray[1] == 'i') {
|
---|
[3365] | 160 | cout << "current index="<< queryInfo.docLevel << "\nchange to index:";
|
---|
| 161 | cin >> query;
|
---|
| 162 | UCArrayClear(queryInfo.docLevel);
|
---|
[8692] | 163 | SetCStr(queryInfo.docLevel, query, strlen(query));
|
---|
[3365] | 164 | cout << "index set to " << queryInfo.docLevel <<"\n";
|
---|
| 165 | cin.getline(query, 2048, '\n');
|
---|
[6116] | 166 | } else if (queryArray[1] == 'l') {
|
---|
[3365] | 167 | cout << "current level="<< level << "\nchange to level:";
|
---|
| 168 | cin >> query;
|
---|
| 169 | UCArrayClear(level);
|
---|
[8692] | 170 | SetCStr(level, query, strlen(query));
|
---|
[3365] | 171 | cout << "level set to " << level <<"\n";
|
---|
| 172 | cin.getline(query, 2048, '\n');
|
---|
| 173 | }
|
---|
| 174 |
|
---|
[12321] | 175 | else if (queryArray[1] == 'm') {
|
---|
| 176 | // maxnumeric
|
---|
| 177 | int m = 0;
|
---|
| 178 | cin >> m;
|
---|
| 179 | cin.getline(query, 2048, '\n'); // eat up return
|
---|
| 180 | if (4 < m < 512) {
|
---|
| 181 | maxnumeric = m;
|
---|
| 182 | }
|
---|
| 183 | }
|
---|
[3365] | 184 | else if (queryArray[1] == 'p') {
|
---|
| 185 | // print
|
---|
| 186 | UCArray docText;
|
---|
| 187 | unsigned long docNum = 0;
|
---|
| 188 | cin >> docNum;
|
---|
| 189 | cin.getline(query, 2048, '\n'); // eat up return
|
---|
| 190 |
|
---|
| 191 | if (!GetDocText (textData, queryInfo.docLevel, docNum, docText)) {
|
---|
| 192 | FatalError (1, "Error while trying to get document %u", docNum);
|
---|
| 193 | }
|
---|
| 194 |
|
---|
| 195 | cout << docText << "\n";
|
---|
| 196 | }
|
---|
| 197 | else if (queryArray[1] == 't') { // query type - all/some
|
---|
| 198 | if (queryArray[2] == '1') defaultBoolCombine = 1;
|
---|
| 199 | else if (queryArray[2] == '0') defaultBoolCombine = 0;
|
---|
| 200 | else {
|
---|
| 201 | cout << "Error: please enter .t0 (some) or .t1 (all)\n";
|
---|
| 202 | }
|
---|
| 203 | }
|
---|
| 204 | else if (queryArray[1] == 'r') { // ranking - on/off
|
---|
| 205 | if (queryArray[2] == '1') queryInfo.sortByRank = true;
|
---|
| 206 | else if (queryArray[2] == '0') queryInfo.sortByRank = false;
|
---|
| 207 | else {
|
---|
| 208 | cout << "Error: please enter .r0 (non-ranked) or .r1 (ranked)\n";
|
---|
| 209 | }
|
---|
| 210 | }
|
---|
| 211 | else if (queryArray[1] == 'c') { // casefolding - on/off
|
---|
[13477] | 212 | if (queryArray[2] == '1') defaultStemMethod |= STEM_CaseFolding;
|
---|
| 213 | else if (queryArray[2] == '0') defaultStemMethod &= (~STEM_CaseFolding);
|
---|
[3365] | 214 | else {
|
---|
| 215 | cout << "Error: please enter .c0 (case sensitive) or .c1 (casefolded)\n";
|
---|
| 216 | }
|
---|
| 217 | }
|
---|
| 218 | else if (queryArray[1] == 's') { // stemming - on/off
|
---|
[13477] | 219 | if (queryArray[2] == '1') defaultStemMethod |= STEM_Stemming;
|
---|
| 220 | else if (queryArray[2] == '0') defaultStemMethod &= (~STEM_Stemming);
|
---|
[3365] | 221 | else {
|
---|
| 222 | cout << "Error: please enter .s0 (unstemmed) or .s1 (stemmed)\n";
|
---|
| 223 | }
|
---|
| 224 | }
|
---|
[13477] | 225 | #ifdef ENABLE_ACCENTFOLD
|
---|
| 226 | else if (queryArray[1] == 'a') { // accentfolding - on/off
|
---|
| 227 | if (queryArray[2] == '1') defaultStemMethod |= STEM_AccentFolding;
|
---|
| 228 | else if (queryArray[2] == '0') defaultStemMethod &= (~STEM_AccentFolding);
|
---|
| 229 | else {
|
---|
| 230 | cout << "Error: please enter .a0 (accent sensitive) or .a1 (accentfolded)\n";
|
---|
| 231 | }
|
---|
| 232 | }
|
---|
| 233 | #endif
|
---|
[3365] | 234 | else if (queryArray[1] == 'o') { // output - short/long
|
---|
| 235 | if (queryArray[2] == '1') shortOutput = true;
|
---|
| 236 | else if (queryArray[2] == '0') shortOutput = false;
|
---|
| 237 | else {
|
---|
| 238 | cout << "Error: please enter .o0 (long output) or .o1 (short output)\n";
|
---|
| 239 | }
|
---|
| 240 | }
|
---|
| 241 | else if (queryArray[1] == 'b') {
|
---|
| 242 | // full text browse
|
---|
| 243 | cout<<"enter a few letters to start browsing from:";
|
---|
| 244 | cin>>query;
|
---|
| 245 | UCArrayClear(browseNode.term);
|
---|
[8692] | 246 | SetCStr(browseNode.term, query, strlen(query));
|
---|
[3365] | 247 | cin.getline(query, 2048, '\n'); // get rest of line
|
---|
| 248 |
|
---|
| 249 | // print the query
|
---|
| 250 | PrintNode (cout, &browseNode);
|
---|
| 251 |
|
---|
| 252 | MGBrowseQuery(indexData, docLevel, browseNode, browseResult);
|
---|
| 253 | cout << browseResult;
|
---|
| 254 | cout << "\n";
|
---|
| 255 |
|
---|
| 256 | }
|
---|
| 257 | else { // bad option
|
---|
| 258 | cout << "bad command\n\n";
|
---|
| 259 | printHelp();
|
---|
| 260 | }
|
---|
| 261 | } // if a .x query
|
---|
| 262 | else {
|
---|
| 263 | // regular query
|
---|
[12321] | 264 | queryTree = ParseQuery (queryArray, defaultBoolCombine, defaultStemMethod, maxnumeric);
|
---|
[4210] | 265 | if (queryTree == NULL) {
|
---|
| 266 | cout << "invalid syntax\n";
|
---|
| 267 | } else {
|
---|
| 268 | // print the query
|
---|
| 269 | PrintNode (cout, queryTree);
|
---|
| 270 |
|
---|
| 271 | MGQuery (indexData, queryInfo, queryTree, queryResult, level);
|
---|
| 272 | if (shortOutput) {
|
---|
| 273 | queryResult.printShort(cout);
|
---|
| 274 | cout << "\n";
|
---|
| 275 | } else {
|
---|
| 276 | cout << queryResult;
|
---|
| 277 | cout << "\n";
|
---|
| 278 | }
|
---|
| 279 | // delete the query
|
---|
| 280 | delete queryTree;
|
---|
| 281 | queryTree = NULL;
|
---|
[3365] | 282 | }
|
---|
| 283 | }
|
---|
| 284 | }
|
---|
| 285 |
|
---|
| 286 |
|
---|
| 287 | // clean up, everybody clean up
|
---|
[13477] | 288 | textData.UnloadData ();
|
---|
[3365] | 289 | indexData.UnloadData ();
|
---|
| 290 |
|
---|
| 291 | return (0);
|
---|
| 292 | }
|
---|
| 293 |
|
---|
| 294 |
|
---|