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