1 | /**********************************************************************
|
---|
2 | *
|
---|
3 | * lucenesearch.cpp --
|
---|
4 | * Copyright (C) 1999-2002 The New Zealand Digital Library Project
|
---|
5 | *
|
---|
6 | * A component of the Greenstone digital library software
|
---|
7 | * from the New Zealand Digital Library Project at the
|
---|
8 | * University of Waikato, New Zealand.
|
---|
9 | *
|
---|
10 | * This program is free software; you can redistribute it and/or modify
|
---|
11 | * it under the terms of the GNU General Public License as published by
|
---|
12 | * the Free Software Foundation; either version 2 of the License, or
|
---|
13 | * (at your option) any later version.
|
---|
14 | *
|
---|
15 | * This program is distributed in the hope that it will be useful,
|
---|
16 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
17 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
18 | * GNU General Public License for more details.
|
---|
19 | *
|
---|
20 | * You should have received a copy of the GNU General Public License
|
---|
21 | * along with this program; if not, write to the Free Software
|
---|
22 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
23 | *
|
---|
24 | *********************************************************************/
|
---|
25 |
|
---|
26 |
|
---|
27 | #if defined(GSDL_USE_OBJECTSPACE)
|
---|
28 | # include <ospace\std\iostream>
|
---|
29 | #elif defined(GSDL_USE_IOS_H)
|
---|
30 | # include <iostream.h>
|
---|
31 | #else
|
---|
32 | # include <iostream>
|
---|
33 | #endif
|
---|
34 |
|
---|
35 | #include <stdio.h>
|
---|
36 | #include <time.h>
|
---|
37 |
|
---|
38 | #include "gsdlconf.h"
|
---|
39 | #include "gsdltools.h"
|
---|
40 | #include "lucenesearch.h"
|
---|
41 | #include "fileutil.h"
|
---|
42 | #include "queryinfo.h"
|
---|
43 | #include "gsdlunicode.h"
|
---|
44 |
|
---|
45 | #include "expat_resultset.h"
|
---|
46 |
|
---|
47 | text_t lucenesearchclass::getindexsuffix(const queryparamclass &qp) {
|
---|
48 | text_t indexsuffix = "index";
|
---|
49 | // get the first char of the level to be the start of the index name
|
---|
50 | text_t suffix = substr(qp.level.begin(), qp.level.begin()+1);
|
---|
51 | lc(suffix);
|
---|
52 | text_t ind = qp.index;
|
---|
53 | text_t sub = qp.subcollection;
|
---|
54 | text_t lang = qp.language;
|
---|
55 |
|
---|
56 | // collection name not added for Lucene
|
---|
57 | indexsuffix = filename_cat(indexsuffix, suffix +ind + sub + lang);
|
---|
58 | return indexsuffix;
|
---|
59 |
|
---|
60 | }
|
---|
61 |
|
---|
62 | ////////////////////
|
---|
63 | // lucenesearch class //
|
---|
64 | ////////////////////
|
---|
65 |
|
---|
66 | lucenesearchclass::lucenesearchclass ()
|
---|
67 | : searchclass() {
|
---|
68 |
|
---|
69 | textlevel = "Doc";
|
---|
70 | }
|
---|
71 |
|
---|
72 | lucenesearchclass::~lucenesearchclass ()
|
---|
73 | {
|
---|
74 | if (cache != NULL)
|
---|
75 | {
|
---|
76 | delete cache;
|
---|
77 | cache = NULL;
|
---|
78 | }
|
---|
79 | }
|
---|
80 |
|
---|
81 | void lucenesearchclass::set_text_level(const text_t &textlevel_arg)
|
---|
82 | {
|
---|
83 | textlevel = textlevel_arg;
|
---|
84 | }
|
---|
85 |
|
---|
86 |
|
---|
87 | bool lucenesearchclass::search(const queryparamclass &queryparams,
|
---|
88 | queryresultsclass &queryresult) {
|
---|
89 |
|
---|
90 | #ifdef __WIN32__
|
---|
91 | char basepath[]="";
|
---|
92 | #else
|
---|
93 | char basepath[] = "/";
|
---|
94 | #endif
|
---|
95 |
|
---|
96 | char *indexname = (filename_cat(collectdir, getindexsuffix(queryparams))).getcstr();
|
---|
97 |
|
---|
98 | // set default stem method from values originally set on prefs page
|
---|
99 | int defaultStemMethod = 0;
|
---|
100 | if (queryparams.casefolding) {
|
---|
101 | defaultStemMethod |= 1;
|
---|
102 | }
|
---|
103 | if (queryparams.stemming) {
|
---|
104 | defaultStemMethod |= 2;
|
---|
105 | }
|
---|
106 |
|
---|
107 | text_t utf8querystring = to_utf8(queryparams.querystring);
|
---|
108 |
|
---|
109 | text_t escaped_utf8querystring = "";
|
---|
110 | text_t::const_iterator here = utf8querystring.begin();
|
---|
111 | while (here != utf8querystring.end()) {
|
---|
112 | if (*here == '"') escaped_utf8querystring.push_back('\\');
|
---|
113 | escaped_utf8querystring.push_back(*here);
|
---|
114 | ++here;
|
---|
115 | }
|
---|
116 |
|
---|
117 | // Use this to skip lucene_query.pl and access GS2LuceneQuery directly (Java must be on path)
|
---|
118 | // text_t cmd = "java -classpath \"" + filename_cat(gsdlhome, "bin", "java", "LuceneWrapper.jar") + "\" org.greenstone.LuceneWrapper.GS2LuceneQuery ";
|
---|
119 | text_t cmd = "\"" + filename_cat(gsdlhome, "bin", "script", "lucene_query.pl") + "\"";
|
---|
120 | cmd += (text_t)" \""+indexname + (text_t)"\" \"" + escaped_utf8querystring + (text_t)"\"";
|
---|
121 |
|
---|
122 | if (!queryparams.filterstring.empty()) {
|
---|
123 | cmd += " -filter \"" + queryparams.filterstring + "\"";
|
---|
124 | }
|
---|
125 | if (!queryparams.sortfield.empty()) {
|
---|
126 | cmd += " -sort \"" + queryparams.sortfield + "\"";
|
---|
127 | }
|
---|
128 | if (!queryparams.fuzziness.empty()) {
|
---|
129 | cmd += " -fuzziness " + queryparams.fuzziness;
|
---|
130 | }
|
---|
131 |
|
---|
132 | // New code to support configuration of the default conjuction operator
|
---|
133 | // set default Boolean combiner from all/some setting
|
---|
134 | // if match_mode == 1, ie all, default=1 ie AND
|
---|
135 | // if match_mode == 0, ie some, default=0, ie OR
|
---|
136 | if (queryparams.match_mode)
|
---|
137 | {
|
---|
138 | cmd += " -dco AND";
|
---|
139 | }
|
---|
140 |
|
---|
141 | // New code to allow Lucene to do paging of search results. This should
|
---|
142 | // substantially improve performance as we don't have to return all 12000
|
---|
143 | // hits if we only need the first 20!
|
---|
144 | if (queryparams.startresults && queryparams.endresults)
|
---|
145 | {
|
---|
146 | cmd += (text_t)" -startresults " + queryparams.startresults;
|
---|
147 | cmd += (text_t)" -endresults " + queryparams.endresults;
|
---|
148 | }
|
---|
149 |
|
---|
150 | text_t xml_text = "";
|
---|
151 |
|
---|
152 | // I don't want to do this, but I have to.
|
---|
153 | text_t gsdlhome_env = "GSDLHOME=" + gsdlhome;
|
---|
154 | putenv(gsdlhome_env.getcstr());
|
---|
155 |
|
---|
156 | #ifdef __WIN32__
|
---|
157 | putenv("GSDLOS=windows");
|
---|
158 |
|
---|
159 | //FILE *PIN = _popen(cmd.getcstr(), "r"); // didn't seem to work
|
---|
160 | cmd = (text_t)"perl -S "+cmd;
|
---|
161 | // we write the result to a file
|
---|
162 | clock_t this_time = clock();
|
---|
163 | text_t filename = "luc";
|
---|
164 | filename.append(this_time);
|
---|
165 | filename.append(".txt");
|
---|
166 |
|
---|
167 | // Used concat 'collectdir' with 'filename' but this ran into problems
|
---|
168 | // when run from a CD-ROM.
|
---|
169 | // Since this is Windows only code, now changed to use %TEMP% directory,
|
---|
170 | // a location designed to be writable
|
---|
171 |
|
---|
172 | // Need to set the Apache server to pass the TEMP envvar (or TMP) with
|
---|
173 | // passenv. Little point using _tempnam() function that generates a
|
---|
174 | // filename and where you need to pass in the dir if *TMP* doesn't exist.
|
---|
175 | // http://msdn.microsoft.com/en-us/library/hs3e7355%28v=vs.80%29.aspx
|
---|
176 | // dir parameter: "The path used in the file name if there is no TMP
|
---|
177 | // environment variable, or if TMP is not a valid directory." Seems
|
---|
178 | // the inverse of what we want: to use TMP/TEMP if collectdir is invalid.
|
---|
179 |
|
---|
180 | text_t out_file = filename_cat(getenv("TEMP"),filename); // TMP or TEMP
|
---|
181 |
|
---|
182 | cmd += " -out \"" + out_file + "\"";
|
---|
183 | int rv = gsdl_system(cmd, true, cerr);
|
---|
184 | if (rv != 0) {
|
---|
185 | cerr << "tried to run command \""<<cmd<<"\", but it failed\n";
|
---|
186 | } else {
|
---|
187 | read_file(out_file, xml_text);
|
---|
188 | remove(out_file.getcstr()); // now delete it
|
---|
189 | }
|
---|
190 | #else
|
---|
191 |
|
---|
192 | #if defined(MACOS)
|
---|
193 | putenv("GSDLOS=darwin");
|
---|
194 | #else
|
---|
195 | putenv("GSDLOS=linux");
|
---|
196 | #endif
|
---|
197 |
|
---|
198 | // What about Solaris etc ????
|
---|
199 |
|
---|
200 |
|
---|
201 | FILE *PIN = popen(cmd.getcstr(), "r");
|
---|
202 |
|
---|
203 | if (PIN==NULL) {
|
---|
204 | perror("PIPE");
|
---|
205 | cerr << "Error: unable to open pipe to " << cmd << endl;
|
---|
206 |
|
---|
207 | return false;
|
---|
208 | }
|
---|
209 | while (!feof(PIN)) {
|
---|
210 | char buffer[256];
|
---|
211 | int num_bytes = fread(buffer,1,256,PIN);
|
---|
212 | xml_text.appendcarr(buffer,num_bytes);
|
---|
213 | }
|
---|
214 |
|
---|
215 | #endif
|
---|
216 |
|
---|
217 | expat_resultset(xml_text,queryresult);
|
---|
218 |
|
---|
219 | #ifdef __WIN32__
|
---|
220 | // _pclose(PIN);
|
---|
221 | #else
|
---|
222 | pclose(PIN);
|
---|
223 | #endif
|
---|
224 |
|
---|
225 | return true;
|
---|
226 | }
|
---|
227 |
|
---|
228 |
|
---|
229 | bool lucenesearchclass::browse_search(const queryparamclass &queryparams,
|
---|
230 | int start, int numDocs,
|
---|
231 | queryresultsclass &queryresult) {
|
---|
232 |
|
---|
233 | cerr << "**** Not sure what this function does!" << endl;
|
---|
234 | return false;
|
---|
235 | }
|
---|
236 |
|
---|
237 | // the document text for 'docnum' is placed in 'output'
|
---|
238 | // docTargetDocument returns 'true' if it was able to
|
---|
239 | // try to get a document
|
---|
240 | // collection is needed to see if an index from the
|
---|
241 | // collection is loaded. THe default index bits are just there cos
|
---|
242 | // the mg version needs them
|
---|
243 |
|
---|
244 | bool lucenesearchclass::docTargetDocument(const text_t &/*defaultindex*/,
|
---|
245 | const text_t &/*defaultsubcollection*/,
|
---|
246 | const text_t &/*defaultlanguage*/,
|
---|
247 | const text_t &collection,
|
---|
248 | int docnum,
|
---|
249 | text_t &output) {
|
---|
250 |
|
---|
251 | // we now get the document directly by lucenesource, so don't use this
|
---|
252 | // method
|
---|
253 | return false;
|
---|
254 | }
|
---|
255 |
|
---|
256 | // used to clear any cached databases for persistent versions of
|
---|
257 | // Greenstone like the Windows local library
|
---|
258 | void lucenesearchclass::unload_database () {
|
---|
259 | }
|
---|
260 |
|
---|
261 | void lucenesearchclass::set_gsdlhome (const text_t &gh)
|
---|
262 | {
|
---|
263 | gsdlhome = gh;
|
---|
264 | }
|
---|