Parent: [d4852f] (diff)

Child: [daa37c] (diff)

Download this file

rcldb.cpp    1540 lines (1417 with data), 48.2 kB

   1
   2
   3
   4
   5
   6
   7
   8
   9
  10
  11
  12
  13
  14
  15
  16
  17
  18
  19
  20
  21
  22
  23
  24
  25
  26
  27
  28
  29
  30
  31
  32
  33
  34
  35
  36
  37
  38
  39
  40
  41
  42
  43
  44
  45
  46
  47
  48
  49
  50
  51
  52
  53
  54
  55
  56
  57
  58
  59
  60
  61
  62
  63
  64
  65
  66
  67
  68
  69
  70
  71
  72
  73
  74
  75
  76
  77
  78
  79
  80
  81
  82
  83
  84
  85
  86
  87
  88
  89
  90
  91
  92
  93
  94
  95
  96
  97
  98
  99
 100
 101
 102
 103
 104
 105
 106
 107
 108
 109
 110
 111
 112
 113
 114
 115
 116
 117
 118
 119
 120
 121
 122
 123
 124
 125
 126
 127
 128
 129
 130
 131
 132
 133
 134
 135
 136
 137
 138
 139
 140
 141
 142
 143
 144
 145
 146
 147
 148
 149
 150
 151
 152
 153
 154
 155
 156
 157
 158
 159
 160
 161
 162
 163
 164
 165
 166
 167
 168
 169
 170
 171
 172
 173
 174
 175
 176
 177
 178
 179
 180
 181
 182
 183
 184
 185
 186
 187
 188
 189
 190
 191
 192
 193
 194
 195
 196
 197
 198
 199
 200
 201
 202
 203
 204
 205
 206
 207
 208
 209
 210
 211
 212
 213
 214
 215
 216
 217
 218
 219
 220
 221
 222
 223
 224
 225
 226
 227
 228
 229
 230
 231
 232
 233
 234
 235
 236
 237
 238
 239
 240
 241
 242
 243
 244
 245
 246
 247
 248
 249
 250
 251
 252
 253
 254
 255
 256
 257
 258
 259
 260
 261
 262
 263
 264
 265
 266
 267
 268
 269
 270
 271
 272
 273
 274
 275
 276
 277
 278
 279
 280
 281
 282
 283
 284
 285
 286
 287
 288
 289
 290
 291
 292
 293
 294
 295
 296
 297
 298
 299
 300
 301
 302
 303
 304
 305
 306
 307
 308
 309
 310
 311
 312
 313
 314
 315
 316
 317
 318
 319
 320
 321
 322
 323
 324
 325
 326
 327
 328
 329
 330
 331
 332
 333
 334
 335
 336
 337
 338
 339
 340
 341
 342
 343
 344
 345
 346
 347
 348
 349
 350
 351
 352
 353
 354
 355
 356
 357
 358
 359
 360
 361
 362
 363
 364
 365
 366
 367
 368
 369
 370
 371
 372
 373
 374
 375
 376
 377
 378
 379
 380
 381
 382
 383
 384
 385
 386
 387
 388
 389
 390
 391
 392
 393
 394
 395
 396
 397
 398
 399
 400
 401
 402
 403
 404
 405
 406
 407
 408
 409
 410
 411
 412
 413
 414
 415
 416
 417
 418
 419
 420
 421
 422
 423
 424
 425
 426
 427
 428
 429
 430
 431
 432
 433
 434
 435
 436
 437
 438
 439
 440
 441
 442
 443
 444
 445
 446
 447
 448
 449
 450
 451
 452
 453
 454
 455
 456
 457
 458
 459
 460
 461
 462
 463
 464
 465
 466
 467
 468
 469
 470
 471
 472
 473
 474
 475
 476
 477
 478
 479
 480
 481
 482
 483
 484
 485
 486
 487
 488
 489
 490
 491
 492
 493
 494
 495
 496
 497
 498
 499
 500
 501
 502
 503
 504
 505
 506
 507
 508
 509
 510
 511
 512
 513
 514
 515
 516
 517
 518
 519
 520
 521
 522
 523
 524
 525
 526
 527
 528
 529
 530
 531
 532
 533
 534
 535
 536
 537
 538
 539
 540
 541
 542
 543
 544
 545
 546
 547
 548
 549
 550
 551
 552
 553
 554
 555
 556
 557
 558
 559
 560
 561
 562
 563
 564
 565
 566
 567
 568
 569
 570
 571
 572
 573
 574
 575
 576
 577
 578
 579
 580
 581
 582
 583
 584
 585
 586
 587
 588
 589
 590
 591
 592
 593
 594
 595
 596
 597
 598
 599
 600
 601
 602
 603
 604
 605
 606
 607
 608
 609
 610
 611
 612
 613
 614
 615
 616
 617
 618
 619
 620
 621
 622
 623
 624
 625
 626
 627
 628
 629
 630
 631
 632
 633
 634
 635
 636
 637
 638
 639
 640
 641
 642
 643
 644
 645
 646
 647
 648
 649
 650
 651
 652
 653
 654
 655
 656
 657
 658
 659
 660
 661
 662
 663
 664
 665
 666
 667
 668
 669
 670
 671
 672
 673
 674
 675
 676
 677
 678
 679
 680
 681
 682
 683
 684
 685
 686
 687
 688
 689
 690
 691
 692
 693
 694
 695
 696
 697
 698
 699
 700
 701
 702
 703
 704
 705
 706
 707
 708
 709
 710
 711
 712
 713
 714
 715
 716
 717
 718
 719
 720
 721
 722
 723
 724
 725
 726
 727
 728
 729
 730
 731
 732
 733
 734
 735
 736
 737
 738
 739
 740
 741
 742
 743
 744
 745
 746
 747
 748
 749
 750
 751
 752
 753
 754
 755
 756
 757
 758
 759
 760
 761
 762
 763
 764
 765
 766
 767
 768
 769
 770
 771
 772
 773
 774
 775
 776
 777
 778
 779
 780
 781
 782
 783
 784
 785
 786
 787
 788
 789
 790
 791
 792
 793
 794
 795
 796
 797
 798
 799
 800
 801
 802
 803
 804
 805
 806
 807
 808
 809
 810
 811
 812
 813
 814
 815
 816
 817
 818
 819
 820
 821
 822
 823
 824
 825
 826
 827
 828
 829
 830
 831
 832
 833
 834
 835
 836
 837
 838
 839
 840
 841
 842
 843
 844
 845
 846
 847
 848
 849
 850
 851
 852
 853
 854
 855
 856
 857
 858
 859
 860
 861
 862
 863
 864
 865
 866
 867
 868
 869
 870
 871
 872
 873
 874
 875
 876
 877
 878
 879
 880
 881
 882
 883
 884
 885
 886
 887
 888
 889
 890
 891
 892
 893
 894
 895
 896
 897
 898
 899
 900
 901
 902
 903
 904
 905
 906
 907
 908
 909
 910
 911
 912
 913
 914
 915
 916
 917
 918
 919
 920
 921
 922
 923
 924
 925
 926
 927
 928
 929
 930
 931
 932
 933
 934
 935
 936
 937
 938
 939
 940
 941
 942
 943
 944
 945
 946
 947
 948
 949
 950
 951
 952
 953
 954
 955
 956
 957
 958
 959
 960
 961
 962
 963
 964
 965
 966
 967
 968
 969
 970
 971
 972
 973
 974
 975
 976
 977
 978
 979
 980
 981
 982
 983
 984
 985
 986
 987
 988
 989
 990
 991
 992
 993
 994
 995
 996
 997
 998
 999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
#ifndef lint
static char rcsid[] = "@(#$Id: rcldb.cpp,v 1.59 2006-03-29 11:18:14 dockes Exp $ (C) 2004 J.F.Dockes";
#endif
/*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the
* Free Software Foundation, Inc.,
* 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*/
#include <stdio.h>
#include <unistd.h>
#include <sys/stat.h>
#include <fnmatch.h>
#include <iostream>
#include <string>
#include <vector>
#include <algorithm>
#ifndef NO_NAMESPACES
using namespace std;
#endif /* NO_NAMESPACES */
#include "rcldb.h"
#include "textsplit.h"
#include "transcode.h"
#include "unacpp.h"
#include "conftree.h"
#include "debuglog.h"
#include "pathut.h"
#include "smallut.h"
#include "pathhash.h"
#include "utf8iter.h"
#include "wipedir.h"
#include "xapian.h"
#include <xapian/stem.h>
#ifndef MAX
#define MAX(A,B) (A>B?A:B)
#endif
#ifndef MIN
#define MIN(A,B) (A<B?A:B)
#endif
// This is how long an abstract we keep or build from beginning of text when
// indexing. It only has an influence on the size of the db as we are free
// to shorten it again when displaying
#define INDEX_ABSTRACT_SIZE 250
// This is the size of the abstract that we synthetize out of query
// term contexts at query time
#define MA_ABSTRACT_SIZE 250
// This is how many words (context size) we keep around query terms
// when building the abstract
#define MA_EXTRACT_WIDTH 4
// Data for a xapian database. There could actually be 2 different
// ones for indexing or query as there is not much in common.
class Native {
public:
bool isopen;
bool iswritable;
string basedir;
// Indexing
Xapian::WritableDatabase wdb;
vector<bool> updated;
// Querying
Xapian::Database db;
Xapian::Query query; // query descriptor: terms and subqueries
// joined by operators (or/and etc...)
Xapian::Enquire *enquire;
Xapian::MSet mset;
string makeAbstract(Xapian::docid id, const list<string>& terms);
bool dbDataToRclDoc(std::string &data, Rcl::Doc &doc,
int qopts,
Xapian::docid docid,
const list<string>& terms);
Native() : isopen(false), iswritable(false), enquire(0) { }
~Native() {
delete enquire;
}
};
Rcl::Db::Db()
{
pdata = new Native;
m_qOpts = 0;
}
Rcl::Db::~Db()
{
LOGDEB1(("Rcl::Db::~Db\n"));
if (pdata == 0)
return;
Native *ndb = (Native *)pdata;
LOGDEB(("Db::~Db: isopen %d iswritable %d\n", ndb->isopen,
ndb->iswritable));
if (ndb->isopen == false)
return;
const char *ermsg = "Unknown error";
try {
LOGDEB(("Rcl::Db::~Db: closing native database\n"));
if (ndb->iswritable == true) {
ndb->wdb.flush();
}
delete ndb;
return;
} catch (const Xapian::Error &e) {
ermsg = e.get_msg().c_str();
} catch (const string &s) {
ermsg = s.c_str();
} catch (const char *s) {
ermsg = s;
} catch (...) {
ermsg = "Caught unknown exception";
}
LOGERR(("Rcl::Db::~Db: got exception: %s\n", ermsg));
}
bool Rcl::Db::open(const string& dir, OpenMode mode, int qops)
{
if (pdata == 0)
return false;
Native *ndb = (Native *)pdata;
LOGDEB(("Db::open: isopen %d iswritable %d\n", ndb->isopen,
ndb->iswritable));
m_qOpts = qops;
if (ndb->isopen) {
LOGERR(("Rcl::Db::open: already open\n"));
return false;
}
const char *ermsg = "Unknown";
try {
switch (mode) {
case DbUpd:
case DbTrunc:
{
int action = (mode == DbUpd) ? Xapian::DB_CREATE_OR_OPEN :
Xapian::DB_CREATE_OR_OVERWRITE;
ndb->wdb = Xapian::WritableDatabase(dir, action);
LOGDEB(("Rcl::Db::open: lastdocid: %d\n",
ndb->wdb.get_lastdocid()));
ndb->updated.resize(ndb->wdb.get_lastdocid() + 1);
for (unsigned int i = 0; i < ndb->updated.size(); i++)
ndb->updated[i] = false;
ndb->iswritable = true;
}
break;
case DbRO:
default:
ndb->iswritable = false;
ndb->db = Xapian::Database(dir);
break;
}
ndb->isopen = true;
ndb->basedir = dir;
return true;
} catch (const Xapian::Error &e) {
ermsg = e.get_msg().c_str();
} catch (const string &s) {
ermsg = s.c_str();
} catch (const char *s) {
ermsg = s;
} catch (...) {
ermsg = "Caught unknown exception";
}
LOGERR(("Rcl::Db::open: exception while opening [%s]: %s\n",
dir.c_str(), ermsg));
return false;
}
// Note: xapian has no close call, we delete and recreate the db
bool Rcl::Db::close()
{
if (pdata == 0)
return false;
Native *ndb = (Native *)pdata;
LOGDEB(("Db::close(): isopen %d iswritable %d\n", ndb->isopen,
ndb->iswritable));
if (ndb->isopen == false)
return true;
const char *ermsg = "Unknown";
try {
if (ndb->iswritable == true) {
ndb->wdb.flush();
LOGDEB(("Rcl:Db: Called xapian flush\n"));
}
delete ndb;
pdata = new Native;
if (pdata)
return true;
} catch (const Xapian::Error &e) {
ermsg = e.get_msg().c_str();
} catch (const string &s) {
ermsg = s.c_str();
} catch (const char *s) {
ermsg = s;
} catch (...) {
ermsg = "Caught unknown exception";
}
LOGERR(("Rcl::Db:close: exception while deleting db: %s\n", ermsg));
return false;
}
bool Rcl::Db::isopen()
{
if (pdata == 0)
return false;
Native *ndb = (Native *)pdata;
return ndb->isopen;
}
// A small class to hold state while splitting text
class mySplitterCB : public TextSplitCB {
public:
Xapian::Document &doc;
Xapian::termpos basepos; // Base for document section
Xapian::termpos curpos; // Last position sent to callback
mySplitterCB(Xapian::Document &d) : doc(d), basepos(1), curpos(0)
{}
bool takeword(const std::string &term, int pos, int, int);
};
// Callback for the document to word splitting class during indexation
bool mySplitterCB::takeword(const std::string &term, int pos, int, int)
{
#if 0
LOGDEB(("mySplitterCB::takeword:splitCb: [%s]\n", term.c_str()));
string printable;
if (transcode(term, printable, "UTF-8", "ISO-8859-1")) {
LOGDEB((" [%s]\n", printable.c_str()));
}
#endif
const char *ermsg;
try {
// Note: 1 is the within document frequency increment. It would
// be possible to assign different weigths to doc parts (ie title)
// by using a higher value
curpos = pos;
doc.add_posting(term, basepos + curpos, 1);
return true;
} catch (const Xapian::Error &e) {
ermsg = e.get_msg().c_str();
} catch (...) {
ermsg= "Unknown error";
}
LOGERR(("Rcl::Db: xapian add_posting error %s\n", ermsg));
return false;
}
// Unaccent and lowercase data, replace \n\r with spaces
// Removing crlfs is so that we can use the text in the document data fields.
// Use unac (with folding extension) for removing accents and casefolding
//
// Note that we always return true (but set out to "" on error). We don't
// want to stop indexation because of a bad string
bool Rcl::dumb_string(const string &in, string &out)
{
out.erase();
if (in.empty())
return true;
string s1;
s1.reserve(in.length());
for (unsigned int i = 0; i < in.length(); i++) {
if (in[i] == '\n' || in[i] == '\r')
s1 += ' ';
else
s1 += in[i];
}
if (!unacmaybefold(s1, out, "UTF-8", true)) {
LOGERR(("dumb_string: unac failed for %s\n", in.c_str()));
out.erase();
// See comment at start of func
return true;
}
return true;
}
/* From omindex direct */
/* Truncate a string to a given maxlength, avoiding cutting off midword
* if reasonably possible. */
static string
truncate_to_word(string & input, string::size_type maxlen)
{
string output;
if (input.length() <= maxlen) {
output = input;
} else {
output = input.substr(0, maxlen);
const char *SEPAR = " \t\n\r-:.;,/[]{}";
string::size_type space = output.find_last_of(SEPAR);
// Original version only truncated at space if space was found after
// maxlen/2. But we HAVE to truncate at space, else we'd need to do
// utf8 stuff to avoid truncating at multibyte char. In any case,
// not finding space means that the text probably has no value.
// Except probably for Asian languages, so we may want to fix this
// one day
if (space == string::npos) {
output.erase();
} else {
output.erase(space);
}
output += " ...";
}
return output;
}
// remove some chars and replace them with spaces
static string stripchars(const string &str, string delims)
{
string out;
string::size_type startPos, pos;
for (pos = 0;;) {
// Skip initial delims, break if this eats all.
if ((startPos = str.find_first_not_of(delims, pos)) == string::npos)
break;
// Find next delimiter or end of string (end of token)
pos = str.find_first_of(delims, startPos);
// Add token to the vector. Note: token cant be empty here
if (pos == string::npos) {
out += str.substr(startPos) + " ";
} else {
out += str.substr(startPos, pos - startPos) + " ";
}
}
return out;
}
// Truncate longer path and uniquize with hash . The goal for this is
// to avoid xapian max term length limitations, not to gain space (we
// gain very little even with very short maxlens like 30)
#define PATHHASHLEN 150
const static string rclSyntAbs = "?!#@";
// Add document in internal form to the database: index the terms in
// the title abstract and body and add special terms for file name,
// date, mime type ... , create the document data record (more
// metadata), and update database
bool Rcl::Db::add(const string &fn, const Rcl::Doc &idoc,
const struct stat *stp)
{
LOGDEB1(("Rcl::Db::add: fn %s\n", fn.c_str()));
if (pdata == 0)
return false;
Native *ndb = (Native *)pdata;
Rcl::Doc doc = idoc;
// Truncate abstract, title and keywords to reasonable lengths. If
// abstract is currently empty, we make up one with the beginning
// of the document.
if (doc.abstract.empty()) {
doc.abstract = rclSyntAbs +
truncate_to_word(doc.text, INDEX_ABSTRACT_SIZE);
} else {
doc.abstract = truncate_to_word(doc.abstract, INDEX_ABSTRACT_SIZE);
}
doc.abstract = stripchars(doc.abstract, "\n\r");
doc.title = truncate_to_word(doc.title, 100);
doc.keywords = truncate_to_word(doc.keywords, 300);
Xapian::Document newdocument;
mySplitterCB splitData(newdocument);
TextSplit splitter(&splitData);
// /////// Split and index terms in document body and auxiliary fields
string noacc;
// Split and index file name as document term(s)
if (dumb_string(doc.utf8fn, noacc)) {
splitter.text_to_words(noacc);
splitData.basepos += splitData.curpos + 100;
}
// Split and index title
if (!dumb_string(doc.title, noacc)) {
LOGERR(("Rcl::Db::add: dumb_string failed\n"));
return false;
}
splitter.text_to_words(noacc);
splitData.basepos += splitData.curpos + 100;
// Split and index body
if (!dumb_string(doc.text, noacc)) {
LOGERR(("Rcl::Db::add: dumb_string failed\n"));
return false;
}
splitter.text_to_words(noacc);
splitData.basepos += splitData.curpos + 100;
// Split and index keywords
if (!dumb_string(doc.keywords, noacc)) {
LOGERR(("Rcl::Db::add: dumb_string failed\n"));
return false;
}
splitter.text_to_words(noacc);
splitData.basepos += splitData.curpos + 100;
// Split and index abstract
if (!dumb_string(doc.abstract, noacc)) {
LOGERR(("Rcl::Db::add: dumb_string failed\n"));
return false;
}
splitter.text_to_words(noacc);
splitData.basepos += splitData.curpos + 100;
////// Special terms for metadata
// Mime type
newdocument.add_term("T" + doc.mimetype);
// Path name term. This is used for existence/uptodate checks
string hash;
pathHash(fn, hash, PATHHASHLEN);
LOGDEB2(("Rcl::Db::add: pathhash [%s]\n", hash.c_str()));
string pathterm = "P" + hash;
newdocument.add_term(pathterm);
// Simple file name. This is used for file name searches only. We index
// it with a term prefix. utf8fn used to be the full path, but it's now
// the simple file name.
if (dumb_string(doc.utf8fn, noacc) && !noacc.empty()) {
noacc = string("XSFN") + noacc;
newdocument.add_term(noacc);
}
// Internal path: with path, makes unique identifier for documents
// inside multidocument files.
string uniterm;
if (!doc.ipath.empty()) {
uniterm = "Q" + hash + "|" + doc.ipath;
newdocument.add_term(uniterm);
}
// Dates etc...
time_t mtime = atol(doc.dmtime.empty() ? doc.fmtime.c_str() :
doc.dmtime.c_str());
struct tm *tm = localtime(&mtime);
char buf[9];
sprintf(buf, "%04d%02d%02d",tm->tm_year+1900, tm->tm_mon + 1, tm->tm_mday);
newdocument.add_term("D" + string(buf)); // Date (YYYYMMDD)
buf[7] = '\0';
if (buf[6] == '3') buf[6] = '2';
newdocument.add_term("W" + string(buf)); // "Weak" - 10ish day interval
buf[6] = '\0';
newdocument.add_term("M" + string(buf)); // Month (YYYYMM)
buf[4] = '\0';
newdocument.add_term("Y" + string(buf)); // Year (YYYY)
// Document data record. omindex has the following nl separated fields:
// - url
// - sample
// - caption (title limited to 100 chars)
// - mime type
string record = "url=file://" + fn;
record += "\nmtype=" + doc.mimetype;
record += "\nfmtime=" + doc.fmtime;
if (!doc.dmtime.empty()) {
record += "\ndmtime=" + doc.dmtime;
}
record += "\norigcharset=" + doc.origcharset;
char sizebuf[20];
sizebuf[0] = 0;
if (stp)
sprintf(sizebuf, "%ld", (long)stp->st_size);
if (sizebuf[0])
record += string("\nfbytes=") + sizebuf;
sprintf(sizebuf, "%u", (unsigned int)doc.text.length());
record += string("\ndbytes=") + sizebuf;
if (!doc.ipath.empty()) {
record += "\nipath=" + doc.ipath;
}
record += "\ncaption=" + doc.title;
record += "\nkeywords=" + doc.keywords;
record += "\nabstract=" + doc.abstract;
record += "\n";
LOGDEB1(("Newdocument data: %s\n", record.c_str()));
newdocument.set_data(record);
const char *fnc = fn.c_str();
// Add db entry or update existing entry:
try {
Xapian::docid did =
ndb->wdb.replace_document(uniterm.empty() ? pathterm : uniterm,
newdocument);
if (did < ndb->updated.size()) {
ndb->updated[did] = true;
LOGDEB(("Rcl::Db::add: docid %d updated [%s , %s]\n", did, fnc,
doc.ipath.c_str()));
} else {
LOGDEB(("Rcl::Db::add: docid %d added [%s , %s]\n", did, fnc,
doc.ipath.c_str()));
}
} catch (...) {
// FIXME: is this ever actually needed?
try {
ndb->wdb.add_document(newdocument);
LOGDEB(("Rcl::Db::add: %s added (failed re-seek for duplicate)\n",
fnc));
} catch (...) {
LOGERR(("Rcl::Db::add: failed again after replace_document\n"));
return false;
}
}
return true;
}
// Test if given filename has changed since last indexed:
bool Rcl::Db::needUpdate(const string &filename, const struct stat *stp)
{
if (pdata == 0)
return false;
Native *ndb = (Native *)pdata;
// If no document exist with this path, we do need update
string hash;
pathHash(filename, hash, PATHHASHLEN);
string pathterm = "P" + hash;
const char *ermsg;
// Look for all documents with this path. We need to look at all
// to set their existence flag. We check the update time on the
// fmtime field which will be identical for all docs inside a
// multi-document file (we currently always reindex all if the
// file changed)
Xapian::PostingIterator doc;
try {
if (!ndb->wdb.term_exists(pathterm)) {
LOGDEB1(("Db::needUpdate: no such path: %s\n", pathterm.c_str()));
return true;
}
Xapian::PostingIterator docid0 = ndb->wdb.postlist_begin(pathterm);
for (Xapian::PostingIterator docid = docid0;
docid != ndb->wdb.postlist_end(pathterm); docid++) {
Xapian::Document doc = ndb->wdb.get_document(*docid);
// Check the date once. no need to look at the others if
// the db needs updating. Note that the fmtime used to be
// called mtime, and we're keeping compat
if (docid == docid0) {
string data = doc.get_data();
const char *cp = strstr(data.c_str(), "fmtime=");
if (cp) {
cp += 7;
} else {
cp = strstr(data.c_str(), "mtime=");
if (cp)
cp+= 6;
}
long mtime = cp ? atol(cp) : 0;
if (mtime < stp->st_mtime) {
LOGDEB2(("Db::needUpdate: yes: mtime: Db %ld file %ld\n",
(long)mtime, (long)stp->st_mtime));
// Db is not up to date. Let's index the file
return true;
}
}
// Db is up to date. Make a note that this document exists.
if (*docid < ndb->updated.size())
ndb->updated[*docid] = true;
}
return false;
} catch (const Xapian::Error &e) {
ermsg = e.get_msg().c_str();
} catch (...) {
ermsg= "Unknown error";
}
LOGERR(("Db::needUpdate: error while checking existence: %s\n", ermsg));
return true;
}
const static string stemdirstem = "stem_";
/// Compute name of stem db for given base database and language
static string stemdbname(const string& basename, string lang)
{
string nm = path_cat(basename, stemdirstem + lang);
return nm;
}
// Deciding if we try to stem the term. If it has numerals or capitals
// we don't
inline static bool
p_notlowerorutf(unsigned int c)
{
if (c < 'a' || (c > 'z' && c < 128))
return true;
return false;
}
/**
* Delete stem db for given language
*/
bool Rcl::Db::deleteStemDb(const string& lang)
{
LOGDEB(("Rcl::Db::deleteStemDb(%s)\n", lang.c_str()));
if (pdata == 0)
return false;
Native *ndb = (Native *)pdata;
if (ndb->isopen == false)
return false;
string dir = stemdbname(ndb->basedir, lang);
if (wipedir(dir) == 0 && rmdir(dir.c_str()) == 0)
return true;
return false;
}
/**
* Create database of stem to parents associations for a given language.
* We walk the list of all terms, stem them, and create another Xapian db
* with documents indexed by a single term (the stem), and with the list of
* parent terms in the document data.
*/
bool Rcl::Db::createStemDb(const string& lang)
{
LOGDEB(("Rcl::Db::createStemDb(%s)\n", lang.c_str()));
if (pdata == 0)
return false;
Native *ndb = (Native *)pdata;
if (ndb->isopen == false)
return false;
// First build the in-memory stem database:
// We walk the list of all terms, and stem each.
// If the stem is identical to the term, no need to create an entry
// Else, we add an entry to the multimap.
// At the end, we only save stem-terms associations with several terms, the
// others are not useful
multimap<string, string> assocs;
// Statistics
int nostem=0; // Dont even try: not-alphanum (incomplete for now)
int stemconst=0; // Stem == term
int stemdiff=0; // Count of all different stems
int stemmultiple = 0; // Count of stems with multiple derivatives
try {
Xapian::Stem stemmer(lang);
Xapian::TermIterator it;
for (it = ndb->wdb.allterms_begin();
it != ndb->wdb.allterms_end(); it++) {
// If it has any non-lowercase 7bit char, cant be stemmable
string::iterator sit = (*it).begin(), eit = sit + (*it).length();
if ((sit = find_if(sit, eit, p_notlowerorutf)) != eit) {
++nostem;
// LOGDEB(("stemskipped: [%s], because of 0x%x\n",
// (*it).c_str(), *sit));
continue;
}
string stem = stemmer.stem_word(*it);
//cerr << "word " << *it << " stem " << stem << endl;
if (stem == *it) {
++stemconst;
continue;
}
assocs.insert(pair<string,string>(stem, *it));
}
} catch (const Xapian::Error &e) {
LOGERR(("Db::createStemDb: build failed: %s\n", e.get_msg().c_str()));
return false;
} catch (...) {
LOGERR(("Db::createStemDb: build failed: no stemmer for %s ? \n",
lang.c_str()));
return false;
}
class DirWiper {
public:
string dir;
bool do_it;
DirWiper(string d) : dir(d), do_it(true) {}
~DirWiper() {
if (do_it) {
wipedir(dir);
rmdir(dir.c_str());
}
}
};
// Create xapian database for stem relations
string stemdbdir = stemdbname(ndb->basedir, lang);
// We want to get rid of the db dir in case of error. This gets disarmed
// just before success return.
DirWiper wiper(stemdbdir);
const char *ermsg = "NOERROR";
Xapian::WritableDatabase sdb;
try {
sdb = Xapian::WritableDatabase(stemdbdir,
Xapian::DB_CREATE_OR_OVERWRITE);
} catch (const Xapian::Error &e) {
ermsg = e.get_msg().c_str();
} catch (const string &s) {
ermsg = s.c_str();
} catch (const char *s) {
ermsg = s;
} catch (...) {
ermsg = "Caught unknown exception";
}
if (ermsg != "NOERROR") {
LOGERR(("Rcl::Db::createstemdb: exception while opening [%s]: %s\n",
stemdbdir.c_str(), ermsg));
return false;
}
// Enter pseud-docs in db. Walk the multimap, only enter
// associations where there are several parent terms
string stem;
list<string> derivs;
for (multimap<string,string>::const_iterator it = assocs.begin();
it != assocs.end(); it++) {
if (stem == it->first) {
// Staying with same stem
derivs.push_back(it->second);
// cerr << " " << it->second << endl;
} else {
// Changing stems
++stemdiff;
if (derivs.size() > 1) {
// Previous stem has multiple derivatives. Enter in db
++stemmultiple;
Xapian::Document newdocument;
newdocument.add_term(stem);
// The doc data is just parents=blank-separated-list
string record = "parents=";
for (list<string>::const_iterator it = derivs.begin();
it != derivs.end(); it++) {
record += *it + " ";
}
record += "\n";
LOGDEB1(("stemdocument data: %s\n", record.c_str()));
newdocument.set_data(record);
try {
sdb.replace_document(stem, newdocument);
} catch (...) {
LOGERR(("Rcl::Db::createstemdb: replace failed\n"));
return false;
}
}
derivs.clear();
stem = it->first;
derivs.push_back(it->second);
// cerr << "\n" << stem << " " << it->second;
}
}
LOGDEB(("Stem map size: %d stems %d mult %d no %d const %d\n",
assocs.size(), stemdiff, stemmultiple, nostem, stemconst));
wiper.do_it = false;
return true;
}
list<string> Rcl::Db::getStemLangs()
{
list<string> dirs;
LOGDEB(("Rcl::Db::getStemLang\n"));
if (pdata == 0)
return dirs;
Native *ndb = (Native *)pdata;
string pattern = stemdirstem + "*";
dirs = path_dirglob(ndb->basedir, pattern);
for (list<string>::iterator it = dirs.begin(); it != dirs.end(); it++) {
*it = path_basename(*it);
*it = it->substr(stemdirstem.length(), string::npos);
}
return dirs;
}
/**
* This is called at the end of an indexing session, to delete the
* documents for files that are no longer there. We also build the
* stem database while we are at it.
*/
bool Rcl::Db::purge()
{
LOGDEB(("Rcl::Db::purge\n"));
if (pdata == 0)
return false;
Native *ndb = (Native *)pdata;
LOGDEB(("Rcl::Db::purge: isopen %d iswritable %d\n", ndb->isopen,
ndb->iswritable));
if (ndb->isopen == false || ndb->iswritable == false)
return false;
// There seems to be problems with the document delete code, when
// we do this, the database is not actually updated. Especially,
// if we delete a bunch of docs, so that there is a hole in the
// docids at the beginning, we can't add anything (appears to work
// and does nothing). Maybe related to the exceptions below when
// trying to delete an unexistant document ?
// Flushing before trying the deletes seeems to work around the problem
try {
ndb->wdb.flush();
} catch (...) {
LOGDEB(("Rcl::Db::purge: 1st flush failed\n"));
}
for (Xapian::docid docid = 1; docid < ndb->updated.size(); ++docid) {
if (!ndb->updated[docid]) {
try {
ndb->wdb.delete_document(docid);
LOGDEB(("Rcl::Db::purge: deleted document #%d\n", docid));
} catch (const Xapian::DocNotFoundError &) {
LOGDEB(("Rcl::Db::purge: document #%d not found\n", docid));
}
}
}
try {
ndb->wdb.flush();
} catch (...) {
LOGDEB(("Rcl::Db::purge: 2nd flush failed\n"));
}
return true;
}
/**
* Expand term to list of all terms which stem to the same term.
*/
static list<string> stemexpand(Native *ndb, string term, const string& lang)
{
list<string> explist;
try {
Xapian::Stem stemmer(lang);
string stem = stemmer.stem_word(term);
LOGDEB(("stemexpand: [%s] stem-> [%s]\n", term.c_str(), stem.c_str()));
// Try to fetch the doc from the stem db
string stemdbdir = stemdbname(ndb->basedir, lang);
Xapian::Database sdb(stemdbdir);
LOGDEB1(("stemexpand: %s lastdocid: %d\n",
stemdbdir.c_str(), sdb.get_lastdocid()));
if (!sdb.term_exists(stem)) {
LOGDEB1(("Rcl::Db::stemexpand: no term for %s\n", stem.c_str()));
explist.push_back(term);
return explist;
}
Xapian::PostingIterator did = sdb.postlist_begin(stem);
if (did == sdb.postlist_end(stem)) {
LOGDEB1(("stemexpand: no term(1) for %s\n",stem.c_str()));
explist.push_back(term);
return explist;
}
Xapian::Document doc = sdb.get_document(*did);
string data = doc.get_data();
// No need for a conftree, but we need to massage the data a little
string::size_type pos = data.find_first_of("=");
++pos;
string::size_type pos1 = data.find_last_of("\n");
if (pos == string::npos || pos1 == string::npos ||pos1 <= pos) { // ??
explist.push_back(term);
return explist;
}
stringToStrings(data.substr(pos, pos1-pos), explist);
if (find(explist.begin(), explist.end(), term) == explist.end()) {
explist.push_back(term);
}
LOGDEB(("stemexpand: %s -> %s\n", stem.c_str(),
stringlistdisp(explist).c_str()));
} catch (...) {
LOGERR(("stemexpand: error accessing stem db\n"));
explist.push_back(term);
return explist;
}
return explist;
}
// Splitter callback for breaking query into terms
class wsQData : public TextSplitCB {
public:
vector<string> terms;
string catterms() {
string s;
for (unsigned int i=0;i<terms.size();i++) {
s += "[" + terms[i] + "] ";
}
return s;
}
bool takeword(const std::string &term, int , int, int) {
LOGDEB1(("wsQData::takeword: %s\n", term.c_str()));
terms.push_back(term);
return true;
}
void dumball() {
for (vector<string>::iterator it=terms.begin(); it !=terms.end();it++){
string dumb;
Rcl::dumb_string(*it, dumb);
*it = dumb;
}
}
};
// Turn string into list of xapian queries. There is little
// interpretation done on the string (no +term -term or filename:term
// stuff). We just separate words and phrases, and interpret
// capitalized terms as wanting no stem expansion.
// The final list contains one query for each term or phrase
// - Elements corresponding to a stem-expanded part are an OP_OR
// composition of the stem-expanded terms (or a single term query).
// - Elements corresponding to a phrase are an OP_PHRASE composition of the
// phrase terms (no stem expansion in this case)
static void stringToXapianQueries(const string &iq,
const string& stemlang,
Native *ndb,
list<Xapian::Query> &pqueries,
Rcl::Db::QueryOpts opts = Rcl::Db::QO_NONE)
{
string qstring = iq;
// Split into (possibly single word) phrases ("this is a phrase"):
list<string> phrases;
stringToStrings(qstring, phrases);
// Then process each phrase: split into terms and transform into
// appropriate Xapian Query
for (list<string>::iterator it=phrases.begin(); it !=phrases.end(); it++) {
LOGDEB(("strToXapianQ: phrase or word: [%s]\n", it->c_str()));
wsQData splitData;
TextSplit splitter(&splitData, true);
splitter.text_to_words(*it);
LOGDEB1(("strToXapianQ: splitter term count: %d\n",
splitData.terms.size()));
switch(splitData.terms.size()) {
case 0: continue;// ??
case 1: // Not a real phrase: one term
{
string term = splitData.terms.front();
bool nostemexp = false;
// Check if the first letter is a majuscule in which
// case we do not want to do stem expansion. Note that
// the test is convoluted and possibly problematic
if (term.length() > 0) {
string noacterm,noaclowterm;
if (unacmaybefold(term, noacterm, "UTF-8", false) &&
unacmaybefold(noacterm, noaclowterm, "UTF-8", true)) {
Utf8Iter it1(noacterm);
Utf8Iter it2(noaclowterm);
if (*it1 != *it2)
nostemexp = true;
}
}
LOGDEB1(("Term: %s stem expansion: %s\n",
term.c_str(), nostemexp?"no":"yes"));
list<string> exp;
string term1;
Rcl::dumb_string(term, term1);
// Possibly perform stem compression/expansion
if (!nostemexp && (opts & Rcl::Db::QO_STEM)) {
exp = stemexpand(ndb, term1, stemlang);
} else {
exp.push_back(term1);
}
// Push either term or OR of stem-expanded set
pqueries.push_back(Xapian::Query(Xapian::Query::OP_OR,
exp.begin(), exp.end()));
}
break;
default:
// Phrase: no stem expansion
splitData.dumball();
LOGDEB(("Pushing phrase: [%s]\n", splitData.catterms().c_str()));
pqueries.push_back(Xapian::Query(Xapian::Query::OP_PHRASE,
splitData.terms.begin(),
splitData.terms.end()));
}
}
}
// Prepare query out of simple query string
bool Rcl::Db::setQuery(const std::string &iqstring, QueryOpts opts,
const string& stemlang)
{
LOGDEB(("Rcl::Db::setQuery: q: [%s], opts 0x%x, stemlang %s\n",
iqstring.c_str(), (unsigned int)opts, stemlang.c_str()));
Native *ndb = (Native *)pdata;
if (!ndb)
return false;
m_asdata.erase();
dbindices.clear();
list<Xapian::Query> pqueries;
stringToXapianQueries(iqstring, stemlang, ndb, pqueries, opts);
ndb->query = Xapian::Query(Xapian::Query::OP_OR, pqueries.begin(),
pqueries.end());
delete ndb->enquire;
ndb->enquire = new Xapian::Enquire(ndb->db);
ndb->enquire->set_query(ndb->query);
ndb->mset = Xapian::MSet();
return true;
}
// Prepare query out of "advanced search" data
bool Rcl::Db::setQuery(AdvSearchData &sdata, QueryOpts opts,
const string& stemlang)
{
LOGDEB(("Rcl::Db::setQuery: adv:\n"));
LOGDEB((" allwords: %s\n", sdata.allwords.c_str()));
LOGDEB((" phrase: %s\n", sdata.phrase.c_str()));
LOGDEB((" orwords: %s\n", sdata.orwords.c_str()));
LOGDEB((" nowords: %s\n", sdata.nowords.c_str()));
string ft;
for (list<string>::iterator it = sdata.filetypes.begin();
it != sdata.filetypes.end(); it++) {ft += *it + " ";}
if (!ft.empty())
LOGDEB((" searched file types: %s\n", ft.c_str()));
if (!sdata.topdir.empty())
LOGDEB((" restricted to: %s\n", sdata.topdir.c_str()));
m_asdata = sdata;
dbindices.clear();
Native *ndb = (Native *)pdata;
if (!ndb)
return false;
list<Xapian::Query> pqueries;
Xapian::Query xq;
if (!sdata.filename.empty()) {
LOGDEB((" filename search\n"));
// File name search, with possible wildcards.
// We expand wildcards by scanning the filename terms (prefixed
// with XSFN) from the database.
// We build an OR query with the expanded values if any.
string pattern;
dumb_string(sdata.filename, pattern);
// If pattern is not quoted, and has no wildcards, we add * at
// each end: match any substring
if (pattern[0] == '"' && pattern[pattern.size()-1] == '"') {
pattern = pattern.substr(1, pattern.size() -2);
} else if (pattern.find_first_of("*?[") == string::npos) {
pattern = "*" + pattern + "*";
} // else let it be
LOGDEB((" pattern: [%s]\n", pattern.c_str()));
// Match pattern against all file names in the db
Xapian::TermIterator it = ndb->db.allterms_begin();
it.skip_to("XSFN");
list<string> names;
for (;it != ndb->db.allterms_end(); it++) {
if ((*it).find("XSFN") != 0)
break;
string fn = (*it).substr(4);
LOGDEB2(("Matching [%s] and [%s]\n", pattern.c_str(), fn.c_str()));
if (fnmatch(pattern.c_str(), fn.c_str(), 0) != FNM_NOMATCH) {
names.push_back((*it).c_str());
}
// Limit the match count
if (names.size() > 1000) {
LOGERR(("Rcl::Db::SetQuery: too many matched file names\n"));
break;
}
}
if (names.empty()) {
// Build an impossible query: we know its impossible because we
// control the prefixes!
names.push_back("XIMPOSSIBLE");
}
// Build a query out of the matching file name terms.
xq = Xapian::Query(Xapian::Query::OP_OR, names.begin(), names.end());
}
if (!sdata.allwords.empty()) {
stringToXapianQueries(sdata.allwords, stemlang, ndb, pqueries, opts);
if (!pqueries.empty()) {
Xapian::Query nq =
Xapian::Query(Xapian::Query::OP_AND, pqueries.begin(),
pqueries.end());
xq = xq.empty() ? nq :
Xapian::Query(Xapian::Query::OP_AND, xq, nq);
pqueries.clear();
}
}
if (!sdata.orwords.empty()) {
stringToXapianQueries(sdata.orwords, stemlang, ndb, pqueries, opts);
if (!pqueries.empty()) {
Xapian::Query nq =
Xapian::Query(Xapian::Query::OP_OR, pqueries.begin(),
pqueries.end());
xq = xq.empty() ? nq :
Xapian::Query(Xapian::Query::OP_AND_MAYBE, xq, nq);
pqueries.clear();
}
}
// We do no stem expansion on 'No' words. Should we ?
if (!sdata.nowords.empty()) {
stringToXapianQueries(sdata.nowords, stemlang, ndb, pqueries);
if (!pqueries.empty()) {
Xapian::Query nq;
nq = Xapian::Query(Xapian::Query::OP_OR, pqueries.begin(),
pqueries.end());
xq = xq.empty() ? nq :
Xapian::Query(Xapian::Query::OP_AND_NOT, xq, nq);
pqueries.clear();
}
}
if (!sdata.phrase.empty()) {
Xapian::Query nq;
string s = string("\"") + sdata.phrase + string("\"");
stringToXapianQueries(s, stemlang, ndb, pqueries);
if (!pqueries.empty()) {
// There should be a single list element phrase query.
xq = xq.empty() ? *pqueries.begin() :
Xapian::Query(Xapian::Query::OP_AND, xq, *pqueries.begin());
pqueries.clear();
}
}
if (!sdata.filetypes.empty()) {
Xapian::Query tq;
for (list<string>::iterator it = sdata.filetypes.begin();
it != sdata.filetypes.end(); it++) {
string term = "T" + *it;
LOGDEB(("Adding file type term: [%s]\n", term.c_str()));
tq = tq.empty() ? Xapian::Query(term) :
Xapian::Query(Xapian::Query::OP_OR, tq, Xapian::Query(term));
}
xq = xq.empty() ? tq : Xapian::Query(Xapian::Query::OP_FILTER, xq, tq);
}
ndb->query = xq;
delete ndb->enquire;
ndb->enquire = new Xapian::Enquire(ndb->db);
ndb->enquire->set_query(ndb->query);
ndb->mset = Xapian::MSet();
// Get the query description and trim the "Xapian::Query"
sdata.description = ndb->query.get_description();
if (sdata.description.find("Xapian::Query") == 0)
sdata.description = sdata.description.substr(strlen("Xapian::Query"));
LOGDEB(("Rcl::Db::SetQuery: Q: %s\n", sdata.description.c_str()));
return true;
}
bool Rcl::Db::getQueryTerms(list<string>& terms)
{
Native *ndb = (Native *)pdata;
if (!ndb)
return false;
terms.clear();
Xapian::TermIterator it;
for (it = ndb->query.get_terms_begin(); it != ndb->query.get_terms_end();
it++) {
terms.push_back(*it);
}
return true;
}
static const int qquantum = 30;
int Rcl::Db::getResCnt()
{
Native *ndb = (Native *)pdata;
if (!ndb || !ndb->enquire) {
LOGERR(("Rcl::Db::getResCnt: no query opened\n"));
return -1;
}
if (ndb->mset.size() <= 0) {
try {
ndb->mset = ndb->enquire->get_mset(0, qquantum);
} catch (const Xapian::DatabaseModifiedError &error) {
ndb->db.reopen();
ndb->mset = ndb->enquire->get_mset(0, qquantum);
} catch (const Xapian::Error & error) {
LOGERR(("enquire->get_mset: exception: %s\n",
error.get_msg().c_str()));
return -1;
}
}
return ndb->mset.get_matches_lower_bound();
}
// This class (friend to RclDb) exists so that we can have functions that
// access private RclDb data and have Xapian-specific parameters (so that we
// don't want them to appear in the public rcldb.h).
class Rcl::DbPops {
public:
static bool filterMatch(Rcl::Db *rdb, Xapian::Document &xdoc) {
// Parse xapian document's data and populate doc fields
string data = xdoc.get_data();
ConfSimple parms(&data);
// The only filtering for now is on file path (subtree)
string url;
parms.get(string("url"), url);
url = url.substr(7);
if (url.find(rdb->m_asdata.topdir) == 0)
return true;
return false;
}
};
bool Native::dbDataToRclDoc(std::string &data, Rcl::Doc &doc,
int qopts,
Xapian::docid docid, const list<string>& terms)
{
LOGDEB1(("Rcl::Db::dbDataToRclDoc: data: %s\n", data.c_str()));
ConfSimple parms(&data);
if (!parms.ok())
return false;
parms.get(string("url"), doc.url);
parms.get(string("mtype"), doc.mimetype);
parms.get(string("fmtime"), doc.fmtime);
parms.get(string("dmtime"), doc.dmtime);
parms.get(string("origcharset"), doc.origcharset);
parms.get(string("caption"), doc.title);
parms.get(string("keywords"), doc.keywords);
parms.get(string("abstract"), doc.abstract);
bool syntabs = false;
if (doc.abstract.find(rclSyntAbs) == 0) {
doc.abstract = doc.abstract.substr(rclSyntAbs.length());
syntabs = true;
}
if ((qopts && Rcl::Db::QO_BUILD_ABSTRACT) && !terms.empty()) {
LOGDEB1(("dbDataToRclDoc:: building abstract from position data\n"));
if (doc.abstract.empty() || syntabs ||
(qopts & Rcl::Db::QO_REPLACE_ABSTRACT))
doc.abstract = makeAbstract(docid, terms);
}
parms.get(string("ipath"), doc.ipath);
parms.get(string("fbytes"), doc.fbytes);
parms.get(string("dbytes"), doc.dbytes);
return true;
}
// Get document at rank i in query (i is the index in the whole result
// set, as in the enquire class. We check if the current mset has the
// doc, else ask for an other one. We use msets of 10 documents. Don't
// know if the whole thing makes sense at all but it seems to work.
//
// If there is a postquery filter (ie: file names), we have to
// maintain a correspondance from the sequential external index
// sequence to the internal Xapian hole-y one (the holes being the documents
// that dont match the filter).
bool Rcl::Db::getDoc(int exti, Doc &doc, int *percent)
{
LOGDEB1(("Rcl::Db::getDoc: exti %d\n", exti));
Native *ndb = (Native *)pdata;
if (!ndb || !ndb->enquire) {
LOGERR(("Rcl::Db::getDoc: no query opened\n"));
return false;
}
// For now the only post-query filter is on dir subtree
bool postqfilter = !m_asdata.topdir.empty();
LOGDEB1(("Topdir %s postqflt %d\n", m_asdata.topdir.c_str(), postqfilter));
int xapi;
if (postqfilter) {
// There is a postquery filter, does this fall in already known area ?
if (exti >= (int)dbindices.size()) {
// Have to fetch xapian docs and filter until we get
// enough or fail
dbindices.reserve(exti+1);
// First xapian doc we fetch is the one after last stored
int first = dbindices.size() > 0 ? dbindices.back() + 1 : 0;
// Loop until we get enough docs
while (exti >= (int)dbindices.size()) {
LOGDEB(("Rcl::Db::getDoc: fetching %d starting at %d\n",
qquantum, first));
try {
ndb->mset = ndb->enquire->get_mset(first, qquantum);
} catch (const Xapian::DatabaseModifiedError &error) {
ndb->db.reopen();
ndb->mset = ndb->enquire->get_mset(first, qquantum);
} catch (const Xapian::Error & error) {
LOGERR(("enquire->get_mset: exception: %s\n",
error.get_msg().c_str()));
abort();
}
if (ndb->mset.empty()) {
LOGDEB(("Rcl::Db::getDoc: got empty mset\n"));
return false;
}
first = ndb->mset.get_firstitem();
for (unsigned int i = 0; i < ndb->mset.size() ; i++) {
LOGDEB(("Rcl::Db::getDoc: [%d]\n", i));
Xapian::Document xdoc = ndb->mset[i].get_document();
if (Rcl::DbPops::filterMatch(this, xdoc)) {
dbindices.push_back(first + i);
}
}
first = first + ndb->mset.size();
}
}
xapi = dbindices[exti];
} else {
xapi = exti;
}
// From there on, we work with a xapian enquire item number. Fetch it
int first = ndb->mset.get_firstitem();
int last = first + ndb->mset.size() -1;
if (!(xapi >= first && xapi <= last)) {
LOGDEB(("Fetching for first %d, count %d\n", xapi, qquantum));
try {
ndb->mset = ndb->enquire->get_mset(xapi, qquantum);
} catch (const Xapian::DatabaseModifiedError &error) {
ndb->db.reopen();
ndb->mset = ndb->enquire->get_mset(xapi, qquantum);
} catch (const Xapian::Error & error) {
LOGERR(("enquire->get_mset: exception: %s\n",
error.get_msg().c_str()));
abort();
}
if (ndb->mset.empty())
return false;
first = ndb->mset.get_firstitem();
last = first + ndb->mset.size() -1;
}
LOGDEB1(("Rcl::Db::getDoc: Qry [%s] win [%d-%d] Estimated results: %d",
ndb->query.get_description().c_str(),
first, last,
ndb->mset.get_matches_lower_bound()));
Xapian::Document xdoc = ndb->mset[xapi-first].get_document();
Xapian::docid docid = *(ndb->mset[xapi-first]);
if (percent)
*percent = ndb->mset.convert_to_percent(ndb->mset[xapi-first]);
// Parse xapian document's data and populate doc fields
string data = xdoc.get_data();
list<string> terms;
getQueryTerms(terms);
return ndb->dbDataToRclDoc(data, doc, m_qOpts, docid, terms);
}
// Retrieve document defined by file name and internal path. Very inefficient,
// used only for history display. We'd need to enter path+ipath terms in the
// db if we wanted to make this more efficient.
bool Rcl::Db::getDoc(const string &fn, const string &ipath, Doc &doc, int *pc)
{
LOGDEB(("Rcl::Db:getDoc: [%s] (%d) [%s]\n", fn.c_str(), fn.length(),
ipath.c_str()));
if (pdata == 0)
return false;
Native *ndb = (Native *)pdata;
// Initialize what we can in any case. If this is history, caller
// will make partial display in case of error
doc.ipath = ipath;
doc.url = string("file://") + fn;
if (*pc)
*pc = 100;
string hash;
pathHash(fn, hash, PATHHASHLEN);
string pathterm = "P" + hash;
// Look for all documents with this path, searching for the one
// with the appropriate ipath. This is very inefficient.
const char *ermsg = "";
try {
if (!ndb->db.term_exists(pathterm)) {
// Document found in history no longer in the database.
// We return true (because their might be other ok docs further)
// but indicate the error with pc = -1
if (*pc)
*pc = -1;
LOGINFO(("Rcl::Db:getDoc: path inexistant: [%s] length %d\n",
pathterm.c_str(), pathterm.length()));
return true;
}
for (Xapian::PostingIterator docid =
ndb->db.postlist_begin(pathterm);
docid != ndb->db.postlist_end(pathterm); docid++) {
Xapian::Document xdoc = ndb->db.get_document(*docid);
string data = xdoc.get_data();
list<string> terms;
if (ndb->dbDataToRclDoc(data, doc, QO_NONE, *docid, terms)
&& doc.ipath == ipath)
return true;
}
} catch (const Xapian::Error &e) {
ermsg = e.get_msg().c_str();
} catch (const string &s) {
ermsg = s.c_str();
} catch (const char *s) {
ermsg = s;
} catch (...) {
ermsg = "Caught unknown exception";
}
if (*ermsg) {
LOGERR(("Rcl::Db::getDoc: %s\n", ermsg));
}
return false;
}
// Width of a sample extract around a query term
//
// We build a possibly full size but sparsely populated (only around
// the search term) reconstruction of the document. It would be
// possible to compress the array, by having only multiple chunks
// around the terms, but this would seriously complicate the data
// structure.
string Native::makeAbstract(Xapian::docid docid, const list<string>& terms)
{
Chrono chron;
// A buffer that we populate with the document terms, at their position
vector<string> buf;
// Go through the list of query terms. For each entry in each
// position list, populate the slot in the document buffer, and
// remember the position and its neigbours
vector<unsigned int> qtermposs; // The term positions
set<unsigned int> chunkposs; // All the positions we shall populate
for (list<string>::const_iterator qit = terms.begin(); qit != terms.end();
qit++) {
Xapian::PositionIterator pos;
// There may be query terms not in this doc. This raises an
// exception when requesting the position list, we just catch it.
try {
unsigned int occurrences = 0;
for (pos = db.positionlist_begin(docid, *qit);
pos != db.positionlist_end(docid, *qit); pos++) {
unsigned int ipos = *pos;
LOGDEB1(("Abstract: [%s] at %d\n", qit->c_str(), ipos));
// Possibly extend the array. Do it in big chunks
if (ipos + MA_EXTRACT_WIDTH >= buf.size()) {
buf.resize(ipos + MA_EXTRACT_WIDTH + 1000);
}
buf[ipos] = *qit;
// Remember the term position
qtermposs.push_back(ipos);
// Add adjacent slots to the set to populate at next step
for (unsigned int ii = MAX(0, ipos-MA_EXTRACT_WIDTH);
ii <= MIN(ipos+MA_EXTRACT_WIDTH, buf.size()-1); ii++) {
chunkposs.insert(ii);
}
// Limit the number of occurences we keep for each
// term. The abstract has a finite length anyway !
if (occurrences++ > 10)
break;
}
} catch (...) {
}
}
LOGDEB1(("Abstract:%d:chosen number of positions %d. Populating\n",
chron.millis(), qtermposs.size()));
// Walk the full document position list and populate slots around
// the query terms. We arbitrarily truncate the list to avoid
// taking forever. If we do cutoff, the abstract may be
// inconsistant, which is bad...
{ Xapian::TermIterator term;
int cutoff = 500 * 1000;
for (term = db.termlist_begin(docid);
term != db.termlist_end(docid); term++) {
Xapian::PositionIterator pos;
for (pos = db.positionlist_begin(docid, *term);
pos != db.positionlist_end(docid, *term); pos++) {
if (cutoff-- < 0)
break;
unsigned int ipos = *pos;
if (chunkposs.find(ipos) != chunkposs.end()) {
buf[ipos] = *term;
}
}
if (cutoff-- < 0)
break;
}
}
LOGDEB1(("Abstract:%d: randomizing and extracting\n", chron.millis()));
// We randomize the selection of term positions, from which we
// shall pull, starting at the beginning, until the abstract is
// big enough. The abstract is finally built in correct position
// order, thanks to the position map.
random_shuffle(qtermposs.begin(), qtermposs.end());
map<unsigned int, string> mabs;
unsigned int abslen = 0;
LOGDEB1(("Abstract:%d: extracting\n", chron.millis()));
// Extract data around the first (in random order) term positions,
// and store the chunks in the map
for (vector<unsigned int>::const_iterator it = qtermposs.begin();
it != qtermposs.end(); it++) {
unsigned int ipos = *it;
unsigned int start = MAX(0, ipos-MA_EXTRACT_WIDTH);
unsigned int end = MIN(ipos+MA_EXTRACT_WIDTH, buf.size()-1);
string chunk;
for (unsigned int ii = start; ii <= end; ii++) {
if (!buf[ii].empty()) {
chunk += buf[ii] + " ";
abslen += buf[ii].length();
}
if (abslen > MA_ABSTRACT_SIZE)
break;
}
if (end != buf.size()-1)
chunk += "... ";
mabs[ipos] = chunk;
if (abslen > MA_ABSTRACT_SIZE)
break;
}
// Build the abstract by walking the map (in order of position)
string abstract;
for (map<unsigned int, string>::const_iterator it = mabs.begin();
it != mabs.end(); it++) {
abstract += (*it).second;
}
LOGDEB(("Abtract: done in %d mS\n", chron.millis()));
return abstract;
}