[14424] | 1 | /*
|
---|
| 2 | * MGRetrieveWrapperImpl.c
|
---|
| 3 | * Copyright (C) 2002 New Zealand Digital Library, http://www.nzdl.org
|
---|
| 4 | *
|
---|
| 5 | * This program is free software; you can redistribute it and/or modify
|
---|
| 6 | * it under the terms of the GNU General Public License as published by
|
---|
| 7 | * the Free Software Foundation; either version 2 of the License, or
|
---|
| 8 | * (at your option) any later version.
|
---|
| 9 | *
|
---|
| 10 | * This program is distributed in the hope that it will be useful,
|
---|
| 11 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
| 12 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
| 13 | * GNU General Public License for more details.
|
---|
| 14 | *
|
---|
| 15 | * You should have received a copy of the GNU General Public License
|
---|
| 16 | * along with this program; if not, write to the Free Software
|
---|
| 17 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
| 18 | */
|
---|
| 19 |
|
---|
| 20 |
|
---|
| 21 | #include "MGWrapperImpl.h"
|
---|
| 22 | #include <jni.h>
|
---|
| 23 | #include "org_greenstone_mg_MGRetrieveWrapper.h"
|
---|
| 24 |
|
---|
| 25 | #include "backend.h"
|
---|
| 26 | #include "environment.h"
|
---|
| 27 | #include "text_get.h"
|
---|
| 28 | #include "stemmer.h"
|
---|
| 29 |
|
---|
| 30 | #define MAX_INDEXES_CACHED 3
|
---|
| 31 |
|
---|
| 32 | /* copied from mgquery, needed to convert paragraph numbers to document numbers
|
---|
| 33 | for greenstone */
|
---|
| 34 | #if defined(PARADOCNUM) || defined(NZDL)
|
---|
| 35 | static int GetDocNumFromParaNum(query_data *qd, int paranum) {
|
---|
| 36 | int Documents = qd->td->cth.num_of_docs;
|
---|
| 37 | int *Paragraph = qd->paragraph;
|
---|
| 38 | int low = 1, high = Documents;
|
---|
| 39 | int mid = (low+high)/2;
|
---|
| 40 |
|
---|
| 41 | while ((mid = (low+high)/2) >=1 && mid <= Documents)
|
---|
| 42 | {
|
---|
| 43 | if (paranum > Paragraph[mid])
|
---|
| 44 | low = mid+1;
|
---|
| 45 | else if (paranum <= Paragraph[mid-1])
|
---|
| 46 | high = mid-1;
|
---|
| 47 | else
|
---|
| 48 | return mid;
|
---|
| 49 | }
|
---|
| 50 | FatalError(1, "Bad paragraph number.\n");
|
---|
| 51 | return 0;
|
---|
| 52 | }
|
---|
| 53 | #endif
|
---|
| 54 |
|
---|
| 55 |
|
---|
| 56 | /*********************************************
|
---|
| 57 | initialisation stuff
|
---|
| 58 | *********************************************/
|
---|
| 59 |
|
---|
| 60 | /* cached ids for java stuff */
|
---|
| 61 | jfieldID FID_mg_data = NULL; /* MGWrapperData */
|
---|
| 62 | jfieldID FID_query_result = NULL; /* MGQueryResult */
|
---|
| 63 | jmethodID MID_addDoc = NULL; /* MGQueryResult.addDoc() */
|
---|
| 64 | jmethodID MID_addTerm = NULL; /* MGQueryResult.addTerm() */
|
---|
| 65 | jmethodID MID_addEquivTerm = NULL; /* MGQueryResult.addEquivTerm() */
|
---|
| 66 | jmethodID MID_setTotalDocs = NULL; /* MGQueryResult.setTotalDocs() */
|
---|
| 67 | jmethodID MID_clearResult = NULL; /* MGQueryResult.clear() */
|
---|
| 68 |
|
---|
| 69 |
|
---|
| 70 | /*******************************************
|
---|
| 71 | Index caching
|
---|
| 72 | *******************************************/
|
---|
| 73 |
|
---|
| 74 | query_data* cached_indexes[MAX_INDEXES_CACHED] = { NULL };
|
---|
| 75 |
|
---|
| 76 |
|
---|
| 77 | /* Get the index data necessary to perform a query or document retrieval */
|
---|
| 78 | query_data*
|
---|
| 79 | loadIndexData(char* base_dir, char* index_path, char* text_path)
|
---|
| 80 | {
|
---|
| 81 | char* index_path_name;
|
---|
| 82 | char* text_path_name;
|
---|
| 83 | query_data* qd;
|
---|
| 84 | int i = 0;
|
---|
| 85 |
|
---|
| 86 | /* Form the path name of the desired indexes */
|
---|
| 87 | index_path_name = (char*) malloc(strlen(base_dir) + strlen(index_path) + 1);
|
---|
| 88 | assert(index_path_name != NULL);
|
---|
| 89 | strcpy(index_path_name, base_dir);
|
---|
| 90 | strcat(index_path_name, index_path);
|
---|
| 91 | printf("Index pathname: %s\n", index_path_name);
|
---|
| 92 |
|
---|
| 93 | text_path_name = (char*) malloc(strlen(base_dir) + strlen(text_path) + 1);
|
---|
| 94 | assert(text_path_name != NULL);
|
---|
| 95 | strcpy(text_path_name, base_dir);
|
---|
| 96 | strcat(text_path_name, text_path);
|
---|
| 97 | printf("Text pathname: %s\n", text_path_name);
|
---|
| 98 |
|
---|
| 99 | /* Search through the cached indexes for the desired one */
|
---|
| 100 | while (i < MAX_INDEXES_CACHED && cached_indexes[i] != NULL) {
|
---|
| 101 | printf("(Cached) Pathname: %s\n", cached_indexes[i]->pathname);
|
---|
| 102 | printf("(Cached) Textpathname: %s\n", cached_indexes[i]->textpathname);
|
---|
| 103 |
|
---|
| 104 | /* Check if the index has already been loaded */
|
---|
| 105 | if ((strcmp(index_path_name, cached_indexes[i]->pathname) == 0) &&
|
---|
| 106 | (strcmp(text_path_name, cached_indexes[i]->textpathname) == 0)) {
|
---|
| 107 | /* Index has already been loaded and cached, so return it */
|
---|
| 108 | printf("Found index!\n");
|
---|
| 109 | free(index_path_name);
|
---|
| 110 | free(text_path_name);
|
---|
| 111 | return cached_indexes[i];
|
---|
| 112 | }
|
---|
| 113 |
|
---|
| 114 | i++;
|
---|
| 115 | }
|
---|
| 116 |
|
---|
| 117 | /* Text strings no longer needed */
|
---|
| 118 | free(index_path_name);
|
---|
| 119 | free(text_path_name);
|
---|
| 120 |
|
---|
| 121 | /* The index is not cached, so load it now */
|
---|
| 122 | qd = InitQuerySystem(base_dir, index_path, text_path, NULL);
|
---|
| 123 | if (!qd) {
|
---|
| 124 | printf("Error: Could not InitQuerySystem()...\n");
|
---|
| 125 | return NULL;
|
---|
| 126 | }
|
---|
| 127 |
|
---|
| 128 | /* The index loaded OK, so cache it */
|
---|
| 129 | /* This could be a little more sophisticated, eg. replace least frequently used index */
|
---|
| 130 | if (i >= MAX_INDEXES_CACHED)
|
---|
| 131 | i = MAX_INDEXES_CACHED - 1;
|
---|
| 132 |
|
---|
| 133 | /* Free the index being replaced */
|
---|
| 134 | if (cached_indexes[i] != NULL)
|
---|
| 135 | FinishQuerySystem(cached_indexes[i]);
|
---|
| 136 |
|
---|
| 137 | /* Cache the loaded index, and return it */
|
---|
| 138 | cached_indexes[i] = qd;
|
---|
| 139 | return cached_indexes[i];
|
---|
| 140 | }
|
---|
| 141 |
|
---|
| 142 | /*
|
---|
| 143 | * Class: org_greenstone_mg_MGRetrieveWrapper
|
---|
| 144 | * Method: initIDs
|
---|
| 145 | * Signature: ()V
|
---|
| 146 | */
|
---|
| 147 | JNIEXPORT void JNICALL
|
---|
| 148 | Java_org_greenstone_mg_MGRetrieveWrapper_initIDs(JNIEnv *j_env, jclass j_cls)
|
---|
| 149 | {
|
---|
| 150 | /* a long-"J" */
|
---|
| 151 | FID_mg_data = (*j_env)->GetFieldID(j_env, j_cls, "mg_data_ptr_", "J");
|
---|
| 152 | assert(FID_mg_data != NULL);
|
---|
| 153 |
|
---|
| 154 | }
|
---|
| 155 |
|
---|
| 156 | /*
|
---|
| 157 | * Class: org_greenstone_mg_MGRetrieveWrapper
|
---|
| 158 | * Method: initCSide
|
---|
| 159 | * Signature: ()Z
|
---|
| 160 | */
|
---|
| 161 | JNIEXPORT jboolean JNICALL Java_org_greenstone_mg_MGRetrieveWrapper_initCSide
|
---|
| 162 | (JNIEnv *j_env, jobject j_obj)
|
---|
| 163 | {
|
---|
| 164 | /* Allocate a MGWrapperData object to store query parameters */
|
---|
| 165 | MGWrapperData* data = (MGWrapperData*) malloc(sizeof(MGWrapperData));
|
---|
| 166 | assert(data != NULL);
|
---|
| 167 |
|
---|
| 168 | /* Set default values - no stemming, no case-folding, boolean OR queries */
|
---|
| 169 | data->defaultStemMethod = 0;
|
---|
| 170 | data->defaultBoolCombine = 0;
|
---|
| 171 |
|
---|
| 172 | /* Allocate a QueryInfo object to store more query parameters */
|
---|
| 173 | data->queryInfo = (QueryInfo*) malloc(sizeof(QueryInfo));
|
---|
| 174 | assert(data->queryInfo != NULL);
|
---|
| 175 |
|
---|
| 176 | /* Set default values - 50 documents max, return term freqs, sort by rank */
|
---|
| 177 | data->queryInfo->index = NULL;
|
---|
| 178 | data->queryInfo->maxDocs = 50;
|
---|
| 179 | data->queryInfo->needTermFreqs = 1;
|
---|
| 180 |
|
---|
| 181 | /* Save the object on the Java side */
|
---|
[25147] | 182 | (*j_env)->SetIntField(j_env, j_obj, FID_mg_data, (mg_s_long) data);
|
---|
[14424] | 183 |
|
---|
| 184 | /* Initialise MG environment variables */
|
---|
| 185 | InitEnv();
|
---|
| 186 | SetEnv("expert", "true", NULL);
|
---|
| 187 | SetEnv("mode", "docnums", NULL);
|
---|
| 188 |
|
---|
| 189 | return 1; /* true - no errors */
|
---|
| 190 | }
|
---|
| 191 |
|
---|
| 192 | /*
|
---|
| 193 | * Class: org_greenstone_mg_MGRetrieveWrapper
|
---|
| 194 | * Method: unloadIndexData
|
---|
| 195 | * Signature: ()Z
|
---|
| 196 | */
|
---|
| 197 | JNIEXPORT jboolean JNICALL Java_org_greenstone_mg_MGRetrieveWrapper_unloadIndexData
|
---|
| 198 | (JNIEnv* j_env, jobject j_obj)
|
---|
| 199 | {
|
---|
| 200 | /* Free all the loaded indexes */
|
---|
| 201 | int i = 0;
|
---|
| 202 | while (i < MAX_INDEXES_CACHED && cached_indexes[i] != NULL) {
|
---|
| 203 | FinishQuerySystem(cached_indexes[i]);
|
---|
| 204 | cached_indexes[i] = NULL;
|
---|
| 205 | i++;
|
---|
| 206 | }
|
---|
| 207 |
|
---|
| 208 | return 1; /* true - no errors */
|
---|
| 209 | }
|
---|
| 210 | /* Choose MG index to search */
|
---|
| 211 | JNIEXPORT void JNICALL
|
---|
| 212 | Java_org_greenstone_mg_MGRetrieveWrapper_setIndex(JNIEnv *j_env, jobject j_obj,
|
---|
| 213 | jstring j_index)
|
---|
| 214 | {
|
---|
| 215 | MGWrapperData* data = (MGWrapperData*) (*j_env)->GetIntField(j_env, j_obj, FID_mg_data);
|
---|
| 216 |
|
---|
| 217 | /* Get the index name as a C string */
|
---|
| 218 | const char* index = (*j_env)->GetStringUTFChars(j_env, j_index, NULL);
|
---|
| 219 | assert(index != NULL);
|
---|
| 220 | printf("Choosing index %s...\n", index);
|
---|
| 221 |
|
---|
| 222 | /* Free the previous index name */
|
---|
| 223 | if (data->queryInfo->index)
|
---|
| 224 | free(data->queryInfo->index);
|
---|
| 225 |
|
---|
| 226 | /* Allocate memory for the index name, and fill it */
|
---|
| 227 | data->queryInfo->index = (char*) malloc(strlen(index) + 1);
|
---|
| 228 | assert(data->queryInfo->index != NULL);
|
---|
| 229 | strcpy(data->queryInfo->index, index);
|
---|
| 230 |
|
---|
| 231 | /* Release the index string */
|
---|
| 232 | (*j_env)->ReleaseStringUTFChars(j_env, j_index, index);
|
---|
| 233 | }
|
---|
| 234 |
|
---|
| 235 | /*
|
---|
| 236 | * Class: org_greenstone_mg_MGRetrieveWrapper
|
---|
| 237 | * Method: getDocument
|
---|
| 238 | * Signature: (Ljava/lang/String;Ljava/lang/String;J)Ljava/lang/String;
|
---|
| 239 | */
|
---|
| 240 | JNIEXPORT jstring JNICALL Java_org_greenstone_mg_MGRetrieveWrapper_getDocument
|
---|
| 241 | (JNIEnv *j_env, jobject j_obj,
|
---|
| 242 | jstring j_base_dir, jstring j_text_path,
|
---|
| 243 | jlong j_docnum)
|
---|
| 244 | {
|
---|
| 245 | MGWrapperData* data = (MGWrapperData*) (*j_env)->GetIntField(j_env, j_obj, FID_mg_data);
|
---|
| 246 |
|
---|
| 247 | char* index_path;
|
---|
| 248 | const char* base_dir;
|
---|
| 249 | const char* text_path;
|
---|
| 250 | query_data* qd;
|
---|
| 251 |
|
---|
[25147] | 252 | mg_u_long pos, len;
|
---|
[14424] | 253 | u_char* c_buffer = NULL;
|
---|
| 254 | u_char* uc_buffer = NULL;
|
---|
| 255 | int ULen;
|
---|
| 256 |
|
---|
| 257 | jstring result;
|
---|
| 258 |
|
---|
| 259 | /* Make sure an index has been specified */
|
---|
| 260 | index_path = data->queryInfo->index;
|
---|
| 261 | assert(index_path != NULL);
|
---|
| 262 |
|
---|
| 263 | /* Obtain C versions of the two string parameters */
|
---|
| 264 | base_dir = (*j_env)->GetStringUTFChars(j_env, j_base_dir, NULL);
|
---|
| 265 | if (base_dir == NULL) {
|
---|
| 266 | return NULL;
|
---|
| 267 | }
|
---|
| 268 | text_path = (*j_env)->GetStringUTFChars(j_env, j_text_path, NULL);
|
---|
| 269 | if (text_path == NULL) {
|
---|
| 270 | (*j_env)->ReleaseStringUTFChars(j_env, j_base_dir, base_dir);
|
---|
| 271 | return NULL;
|
---|
| 272 | }
|
---|
| 273 |
|
---|
| 274 | /* Load the appropriate index for satisfying this request */
|
---|
| 275 | printf("Document retrieval, index path: %s\n", index_path);
|
---|
| 276 | qd = loadIndexData((char*) base_dir, (char*) index_path, (char*) text_path);
|
---|
| 277 |
|
---|
| 278 | /* The C text strings are no longer needed */
|
---|
| 279 | (*j_env)->ReleaseStringUTFChars(j_env, j_base_dir, base_dir);
|
---|
| 280 | (*j_env)->ReleaseStringUTFChars(j_env, j_text_path, text_path);
|
---|
| 281 |
|
---|
| 282 | /* Check that the index was loaded successfully */
|
---|
| 283 | if (qd==NULL) {
|
---|
| 284 | return NULL;
|
---|
| 285 | }
|
---|
| 286 | /*assert(qd != NULL);*/
|
---|
| 287 |
|
---|
| 288 | /* Get the document position and length in the text file */
|
---|
[25147] | 289 | printf("Fetching document number %d...\n", (mg_u_long) j_docnum);
|
---|
| 290 | FetchDocStart(qd, (mg_u_long) j_docnum, &pos, &len);
|
---|
| 291 | printf("Fetched document start. Pos: %d, Len: %d\n", pos, len);
|
---|
[14424] | 292 |
|
---|
| 293 | /* Allocate memory for the document text (from mg/src/text/mgquery.c:RawDocOutput()) */
|
---|
| 294 | c_buffer = (u_char*) malloc(len);
|
---|
| 295 | assert(c_buffer != NULL);
|
---|
| 296 | uc_buffer = (u_char*) malloc((int) (qd->td->cth.ratio * 1.01 * len) + 100);
|
---|
| 297 | assert(uc_buffer != NULL);
|
---|
| 298 |
|
---|
| 299 | /* Seek to the correct position in the file and read the document text */
|
---|
| 300 | Fseek (qd->td->TextFile, pos, 0);
|
---|
| 301 | Fread (c_buffer, 1, len, qd->td->TextFile);
|
---|
| 302 |
|
---|
| 303 | /* Decompress the document text into another buffer, and terminate it */
|
---|
| 304 | DecodeText (qd->cd, c_buffer, len, uc_buffer, &ULen);
|
---|
| 305 | uc_buffer[ULen] = '\0';
|
---|
| 306 |
|
---|
| 307 | /* Load the document text into a Java string */
|
---|
| 308 | result = (*j_env)->NewStringUTF(j_env, uc_buffer);
|
---|
| 309 | assert(result != NULL);
|
---|
| 310 |
|
---|
| 311 | /* Free C buffers */
|
---|
| 312 | free(c_buffer);
|
---|
| 313 | free(uc_buffer);
|
---|
| 314 |
|
---|
| 315 | /* Return the document text */
|
---|
| 316 | return result;
|
---|
| 317 | }
|
---|
| 318 |
|
---|