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 | // This was added for Solaris, but it makes things worse on Solaris for me...
|
---|
25 | // #define _XOPEN_SOURCE_EXTENDED 1
|
---|
26 |
|
---|
27 | #if defined(WIN32) || defined(_WIN32) || defined(__WIN32__)
|
---|
28 | #pragma warning(disable:4786)
|
---|
29 | #endif
|
---|
30 |
|
---|
31 | // need this to avoid bizarre compiler problems under VC++ 6.0
|
---|
32 | #if defined (__WIN32__) && !defined (GSDL_USE_IOS_H)
|
---|
33 | # include <iostream>
|
---|
34 | #endif
|
---|
35 |
|
---|
36 | #include "sysfuncs.h"
|
---|
37 |
|
---|
38 | #ifdef HAVE_MALLINFO
|
---|
39 | # include <malloc.h>
|
---|
40 | #endif
|
---|
41 |
|
---|
42 | #if defined __WIN32__
|
---|
43 | # include <io.h>
|
---|
44 | # include "getopt_old.h"
|
---|
45 | # define close _close
|
---|
46 | # define open _open
|
---|
47 | #elif defined __CYGWIN__
|
---|
48 | #include "getopt_old.h"
|
---|
49 | #else
|
---|
50 | # include <unistd.h>
|
---|
51 | #endif
|
---|
52 |
|
---|
53 | #include "memlib.h"
|
---|
54 | #include "messages.h"
|
---|
55 | #include "longlong.h"
|
---|
56 | #include "mg_files.h"
|
---|
57 | #include "mg.h"
|
---|
58 | #include "build.h"
|
---|
59 | #include "text.h"
|
---|
60 | #include "stemmer.h"
|
---|
61 | #include "FileBuf.h"
|
---|
62 | #include "TextEl.h"
|
---|
63 | #include "TagInfo.h"
|
---|
64 |
|
---|
65 | #include "mgpp_passes_4jni.h"
|
---|
66 |
|
---|
67 | #define MAX_PASSES 5
|
---|
68 |
|
---|
69 | #define SPECIAL 1
|
---|
70 | #define TEXT_PASS_1 2
|
---|
71 | #define TEXT_PASS_2 4
|
---|
72 | #define IVF_PASS_1 8
|
---|
73 | #define IVF_PASS_2 16
|
---|
74 |
|
---|
75 | #define MIN_BUF 8192
|
---|
76 |
|
---|
77 |
|
---|
78 | mg_u_long invf_buffer_size = 5 * 1024 * 1024; /* 5Mb */
|
---|
79 |
|
---|
80 | static char Passes = 0;
|
---|
81 | static char **files = NULL;
|
---|
82 | static int num_files = 0;
|
---|
83 |
|
---|
84 | TagInfo tagInfo;
|
---|
85 | char *filename = NULL;
|
---|
86 |
|
---|
87 | mg_u_long numBytes = 0;
|
---|
88 | mg_u_long numDocs = 0;
|
---|
89 | int mgpp_passes_exit_value = 0;
|
---|
90 |
|
---|
91 | struct pass_data {
|
---|
92 | char *name;
|
---|
93 | int (*init) (const TagInfo &tagInfo, char *);
|
---|
94 | int (*process) (const TagInfo &tagInfo, const TextElArray &doc);
|
---|
95 | int (*done) (const TagInfo &tagInfo, char *);
|
---|
96 | };
|
---|
97 |
|
---|
98 |
|
---|
99 | static pass_data PassData[MAX_PASSES] = {
|
---|
100 | {(char*)"special", init_special, process_special, done_special},
|
---|
101 | {(char*)"text.pass1", init_text_1, process_text_1, done_text_1},
|
---|
102 | {(char*)"text.pass2", init_text_2, process_text_2, done_text_2},
|
---|
103 | {(char*)"ivf.pass1", init_ivf_1, process_ivf_1, done_ivf_1},
|
---|
104 | {(char*)"ivf.pass2", init_ivf_2, process_ivf_2, done_ivf_2},
|
---|
105 | };
|
---|
106 |
|
---|
107 |
|
---|
108 | /* clear all the settings from one mgpp_passes run to the next */
|
---|
109 | void clear_variables() {
|
---|
110 | tagInfo.Clear();
|
---|
111 | tagInfo.SetDocTag ("Document");
|
---|
112 | Passes = 0;
|
---|
113 | invf_buffer_size = 5 * 1024 * 1024;
|
---|
114 | numBytes = 0;
|
---|
115 | numDocs = 0;
|
---|
116 | mgpp_passes_exit_value = 0;
|
---|
117 | }
|
---|
118 |
|
---|
119 | /* ################################################## */
|
---|
120 | /* the following are methods to set all the variables that used to be
|
---|
121 | set by command line args */
|
---|
122 |
|
---|
123 | /* -S, -T1, -T2, -I1, -I2, args to mg_passes */
|
---|
124 | void add_pass (char pass_type, char pass_num) {
|
---|
125 |
|
---|
126 | switch(pass_type) {
|
---|
127 | case 'S':
|
---|
128 | Passes |= SPECIAL;
|
---|
129 | break;
|
---|
130 | case 'I':
|
---|
131 | if (pass_num == '1')
|
---|
132 | Passes |= IVF_PASS_1;
|
---|
133 | else if (pass_num == '2')
|
---|
134 | Passes |= IVF_PASS_2;
|
---|
135 | else
|
---|
136 | fprintf(stderr, "Invalid pass number %c for pass type %c\n", pass_num, pass_type);
|
---|
137 | break;
|
---|
138 | case 'T':
|
---|
139 | if (pass_num == '1')
|
---|
140 | Passes |= TEXT_PASS_1;
|
---|
141 | else if (pass_num == '2')
|
---|
142 | Passes |= TEXT_PASS_2;
|
---|
143 | else
|
---|
144 | fprintf(stderr, "Invalid pass number %c for pass type %c\n", pass_num, pass_type);
|
---|
145 | break;
|
---|
146 | }
|
---|
147 | }
|
---|
148 |
|
---|
149 | /* -m arg to mgpp_passes */
|
---|
150 | void set_inversion_limit(int limit) {
|
---|
151 | invf_buffer_size = limit * 1024 * 1024;
|
---|
152 | }
|
---|
153 |
|
---|
154 | /* -J arg to mgpp_passes */
|
---|
155 | void set_document_tag(const char *tag_name) {
|
---|
156 | tagInfo.SetDocTag (tag_name);
|
---|
157 |
|
---|
158 | // a doc tag is also a level tag
|
---|
159 | tagInfo.AddLevelTag(tag_name);
|
---|
160 |
|
---|
161 | }
|
---|
162 |
|
---|
163 | /* -K arg to mgpp_passes */
|
---|
164 | void add_level_tag(const char *tag_nam) {
|
---|
165 | tagInfo.AddLevelTag(tag_nam);
|
---|
166 | }
|
---|
167 |
|
---|
168 | /* -L arg to mgpp_passes */
|
---|
169 | void set_index_level(const char *tag_name) {
|
---|
170 | tagInfo.SetIndexLevel(tag_name);
|
---|
171 | }
|
---|
172 |
|
---|
173 | // is this enough???
|
---|
174 | /* -f arg to mgpp_passes */
|
---|
175 | void set_filename(const char * filen) {
|
---|
176 | if (filename) {
|
---|
177 | Xfree (filename);
|
---|
178 | filename = NULL;
|
---|
179 | }
|
---|
180 | filename = Xstrdup (filen);
|
---|
181 | }
|
---|
182 |
|
---|
183 |
|
---|
184 | /* ############################################### */
|
---|
185 | /* The old driver method has been split into 3:
|
---|
186 | init_driver, process_document (called numdocs times),
|
---|
187 | finalise_driver.
|
---|
188 | The above set vars methods should all be called before init_driver.
|
---|
189 | */
|
---|
190 |
|
---|
191 | void init_driver () {
|
---|
192 |
|
---|
193 | if (!filename || *filename == '\0') {
|
---|
194 | mgpp_passes_exit_value = 1;
|
---|
195 | FatalError (1, "A document collection name must be specified.");
|
---|
196 | }
|
---|
197 | if ((Passes & (IVF_PASS_1 | IVF_PASS_2)) == (IVF_PASS_1 | IVF_PASS_2)) {
|
---|
198 | mgpp_passes_exit_value = 1;
|
---|
199 | FatalError (1, "I1 and I2 cannot be done simultaneously.");
|
---|
200 | }
|
---|
201 | if ((Passes & (TEXT_PASS_1 | TEXT_PASS_2)) == (TEXT_PASS_1 | TEXT_PASS_2)) {
|
---|
202 | mgpp_passes_exit_value = 1;
|
---|
203 | FatalError (1, "T1 and T2 cannot be done simultaneously.");
|
---|
204 | }
|
---|
205 | if (!Passes) {
|
---|
206 | mgpp_passes_exit_value = 1;
|
---|
207 | FatalError (1, "S, T1, T2, I1 or I2 must be specified.");
|
---|
208 | }
|
---|
209 |
|
---|
210 | // initialise all the passes
|
---|
211 | for (int pass = 0; pass < MAX_PASSES; pass++) {
|
---|
212 | if (Passes & (1 << pass)) {
|
---|
213 | if (PassData[pass].init (tagInfo, filename) == COMPERROR) {
|
---|
214 | mgpp_passes_exit_value = 1;
|
---|
215 | FatalError (1, "Error during init of \"%s\"", PassData[pass].name);
|
---|
216 | }
|
---|
217 | }
|
---|
218 | }
|
---|
219 |
|
---|
220 | }
|
---|
221 |
|
---|
222 | void process_document(u_char *buffer, int len) {
|
---|
223 |
|
---|
224 | TextElArray doc;
|
---|
225 | mg_u_long doc_len = 0;
|
---|
226 |
|
---|
227 | while(ReadDoc(&buffer, tagInfo.docTag, doc, doc_len, false)) {
|
---|
228 | // give this document to each pass
|
---|
229 | for (int pass = 0; pass < MAX_PASSES; pass++) {
|
---|
230 | if (Passes & (1 << pass)) {
|
---|
231 | if (PassData[pass].process (tagInfo, doc) == COMPERROR) {
|
---|
232 | mgpp_passes_exit_value = 1;
|
---|
233 | FatalError(1,"Error during processing of \"%s\"",PassData[pass].name);
|
---|
234 | }
|
---|
235 | }
|
---|
236 | }
|
---|
237 |
|
---|
238 | // another document has been processed
|
---|
239 | numBytes += doc_len;
|
---|
240 | numDocs++;
|
---|
241 | cout << "doc_len = "<<doc_len<<endl;
|
---|
242 | }
|
---|
243 | }
|
---|
244 |
|
---|
245 | void finalise_driver () {
|
---|
246 |
|
---|
247 | // do done for each pass
|
---|
248 | for (int pass = 0; pass < MAX_PASSES; pass++) {
|
---|
249 | if (Passes & (1 << pass)) {
|
---|
250 | if (PassData[pass].done (tagInfo, filename) == COMPERROR) {
|
---|
251 | mgpp_passes_exit_value = 1;
|
---|
252 | FatalError (1, "Error during done of \"%s\"", PassData[pass].name);
|
---|
253 | }
|
---|
254 | }
|
---|
255 | }
|
---|
256 | }
|
---|
257 |
|
---|
258 | int get_exit_value() {
|
---|
259 | return mgpp_passes_exit_value;
|
---|
260 | }
|
---|
261 | /*
|
---|
262 | static void Driver (int in_fd, char *file_name,
|
---|
263 | const TagInfo &tagInfo, bool compatMode) {
|
---|
264 | // cout << tagInfo;
|
---|
265 |
|
---|
266 | int pass;
|
---|
267 |
|
---|
268 | mg_u_long numBytes = 0;
|
---|
269 | mg_u_long numDocs = 0;
|
---|
270 |
|
---|
271 | // initialise all the passes
|
---|
272 | for (pass = 0; pass < MAX_PASSES; pass++) {
|
---|
273 | if (Passes & (1 << pass)) {
|
---|
274 | if (PassData[pass].init (tagInfo, file_name) == COMPERROR)
|
---|
275 | FatalError (1, "Error during init of \"%s\"", PassData[pass].name);
|
---|
276 | }
|
---|
277 | }
|
---|
278 |
|
---|
279 |
|
---|
280 | // set up various variables
|
---|
281 | FileBuf buf;
|
---|
282 | TextElArray doc;
|
---|
283 | mg_u_long docLen = 0;
|
---|
284 |
|
---|
285 | // read and process each file (start with an open file)
|
---|
286 | do {
|
---|
287 |
|
---|
288 | // read and process each document in this file
|
---|
289 | buf.SetFD (in_fd);
|
---|
290 | while (ReadDoc (buf, tagInfo.docTag, doc, docLen, compatMode)) {
|
---|
291 |
|
---|
292 | // give this document to each pass
|
---|
293 | for (pass = 0; pass < MAX_PASSES; pass++) {
|
---|
294 | if (Passes & (1 << pass)) {
|
---|
295 | if (PassData[pass].process (tagInfo, doc) == COMPERROR)
|
---|
296 | FatalError(1,"Error during processing of \"%s\"",PassData[pass].name);
|
---|
297 | }
|
---|
298 | }
|
---|
299 |
|
---|
300 | // another document has been processed
|
---|
301 | numBytes += docLen;
|
---|
302 | numDocs++;
|
---|
303 | }
|
---|
304 |
|
---|
305 | } while ((in_fd = OpenNextFile (in_fd)) > 0);
|
---|
306 |
|
---|
307 |
|
---|
308 | // do done for each pass
|
---|
309 | for (pass = 0; pass < MAX_PASSES; pass++) {
|
---|
310 | if (Passes & (1 << pass)) {
|
---|
311 | if (PassData[pass].done (tagInfo, file_name) == COMPERROR)
|
---|
312 | FatalError (1, "Error during done of \"%s\"", PassData[pass].name);
|
---|
313 | }
|
---|
314 | }
|
---|
315 | }
|
---|
316 |
|
---|
317 | */
|
---|