Parent: [a4d168] (diff)

Child: [08a65f] (diff)

Download this file

rcldb.cpp    1785 lines (1612 with data), 55.4 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
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
/* Copyright (C) 2004 J.F.Dockes
* 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 <cstring>
#include <unistd.h>
#include <fnmatch.h>
#include <regex.h>
#include <math.h>
#include <time.h>
#include <iostream>
#include <string>
#include <vector>
#include <algorithm>
#ifndef NO_NAMESPACES
using namespace std;
#endif /* NO_NAMESPACES */
#include "xapian.h"
#include "rclconfig.h"
#include "rcldb.h"
#include "rcldb_p.h"
#include "stemdb.h"
#include "textsplit.h"
#include "transcode.h"
#include "unacpp.h"
#include "conftree.h"
#include "debuglog.h"
#include "pathut.h"
#include "smallut.h"
#include "utf8iter.h"
#include "searchdata.h"
#include "rclquery.h"
#include "rclquery_p.h"
#include "md5.h"
#include "rclversion.h"
#include "cancelcheck.h"
#ifndef MAX
#define MAX(A,B) (A>B?A:B)
#endif
#ifndef MIN
#define MIN(A,B) (A<B?A:B)
#endif
// Recoll index format version is stored in user metadata. When this change,
// we can't open the db and will have to reindex.
static const string RCL_IDX_VERSION_KEY("RCL_IDX_VERSION_KEY");
static const string RCL_IDX_VERSION("1");
// This is the word position offset at which we index the body text
// (abstract, keywords, etc.. are stored before this)
static const unsigned int baseTextPosition = 100000;
#ifndef NO_NAMESPACES
namespace Rcl {
#endif
const string pathelt_prefix = "XP";
string version_string(){
return string("Recoll ") + string(rclversionstr) + string(" + Xapian ") +
string(Xapian::version_string());
}
// Synthetic abstract marker (to discriminate from abstract actually
// found in document)
static const string rclSyntAbs("?!#@");
// Compute the unique term used to link documents to their origin.
// "Q" + external udi
static inline string make_uniterm(const string& udi)
{
string uniterm("Q");
uniterm.append(udi);
return uniterm;
}
// Compute parent term used to link documents to their parent document (if any)
// "F" + parent external udi
static inline string make_parentterm(const string& udi)
{
// I prefer to be in possible conflict with omega than with
// user-defined fields (Xxxx) that we also allow. "F" is currently
// not used by omega (2008-07)
string pterm("F");
pterm.append(udi);
return pterm;
}
static inline void leftzeropad(string& s, unsigned len)
{
if (s.length() && s.length() < len)
s = s.insert(0, len - s.length(), '0');
}
/* See comment in class declaration: return all subdocuments of a
* document given by its unique id.
*/
bool Db::Native::subDocs(const string &udi, vector<Xapian::docid>& docids)
{
LOGDEB2(("subDocs: [%s]\n", uniterm.c_str()));
string pterm = make_parentterm(udi);
XAPTRY(docids.clear();
docids.insert(docids.begin(), xrdb.postlist_begin(pterm),
xrdb.postlist_end(pterm)),
xrdb, m_rcldb->m_reason);
if (!m_rcldb->m_reason.empty()) {
LOGERR(("Rcl::Db::subDocs: %s\n", m_rcldb->m_reason.c_str()));
return false;
} else {
LOGDEB0(("Db::Native::subDocs: returning %d ids\n", docids.size()));
return true;
}
}
// Only ONE field name inside the index data record differs from the
// Rcl::Doc ones: caption<->title, for a remnant of compatibility with
// omega
static const string keycap("caption");
// Turn data record from db into document fields
bool Db::Native::dbDataToRclDoc(Xapian::docid docid, std::string &data,
Doc &doc)
{
LOGDEB2(("Db::dbDataToRclDoc: data:\n%s\n", data.c_str()));
ConfSimple parms(data);
if (!parms.ok())
return false;
parms.get(Doc::keyurl, doc.url);
parms.get(Doc::keytp, doc.mimetype);
parms.get(Doc::keyfmt, doc.fmtime);
parms.get(Doc::keydmt, doc.dmtime);
parms.get(Doc::keyoc, doc.origcharset);
parms.get(keycap, doc.meta[Doc::keytt]);
parms.get(Doc::keykw, doc.meta[Doc::keykw]);
parms.get(Doc::keyabs, doc.meta[Doc::keyabs]);
// Possibly remove synthetic abstract indicator (if it's there, we
// used to index the beginning of the text as abstract).
doc.syntabs = false;
if (doc.meta[Doc::keyabs].find(rclSyntAbs) == 0) {
doc.meta[Doc::keyabs] = doc.meta[Doc::keyabs].substr(rclSyntAbs.length());
doc.syntabs = true;
}
parms.get(Doc::keyipt, doc.ipath);
parms.get(Doc::keyfs, doc.fbytes);
parms.get(Doc::keyds, doc.dbytes);
parms.get(Doc::keysig, doc.sig);
doc.xdocid = docid;
// Other, not predefined meta fields:
list<string> keys = parms.getNames(string());
for (list<string>::const_iterator it = keys.begin();
it != keys.end(); it++) {
if (doc.meta.find(*it) == doc.meta.end())
parms.get(*it, doc.meta[*it]);
}
doc.meta[Doc::keymt] = doc.dmtime.empty() ? doc.fmtime : doc.dmtime;
return true;
}
// Remove prefixes (caps) from a list of terms.
static void noPrefixList(const list<string>& in, list<string>& out)
{
for (list<string>::const_iterator qit = in.begin();
qit != in.end(); qit++) {
if ('A' <= qit->at(0) && qit->at(0) <= 'Z') {
string term = *qit;
while (term.length() && 'A' <= term.at(0) && term.at(0) <= 'Z')
term.erase(0, 1);
if (term.length())
out.push_back(term);
continue;
} else {
out.push_back(*qit);
}
}
}
//#define DEBUGABSTRACT 1
#ifdef DEBUGABSTRACT
#define LOGABS LOGDEB
#else
#define LOGABS LOGDEB2
#endif
#if 0
static void listList(const string& what, const list<string>&l)
{
string a;
for (list<string>::const_iterator it = l.begin(); it != l.end(); it++) {
a = a + *it + " ";
}
LOGDEB(("%s: %s\n", what.c_str(), a.c_str()));
}
#endif
// Build a document abstract by extracting text chunks around the query terms
// This uses the db termlists, not the original document.
//
// DatabaseModified and other general exceptions are catched and
// possibly retried by our caller
string Db::Native::makeAbstract(Xapian::docid docid, Query *query)
{
Chrono chron;
LOGDEB2(("makeAbstract:%d: maxlen %d wWidth %d\n", chron.ms(),
m_rcldb->m_synthAbsLen, m_rcldb->m_synthAbsWordCtxLen));
list<string> terms;
{
list<string> iterms;
query->getMatchTerms(docid, iterms);
noPrefixList(iterms, terms);
if (terms.empty()) {
LOGDEB(("makeAbstract::Empty term list\n"));
return string();
}
}
// listList("Match terms: ", terms);
// Retrieve db-wide frequencies for the query terms (we do this once per
// query, using all the query terms, not only the document match terms)
if (query->m_nq->termfreqs.empty()) {
list<string> iqterms, qterms;
query->getQueryTerms(iqterms);
noPrefixList(iqterms, qterms);
// listList("Query terms: ", qterms);
double doccnt = xrdb.get_doccount();
if (doccnt == 0) doccnt = 1;
for (list<string>::const_iterator qit = qterms.begin();
qit != qterms.end(); qit++) {
query->m_nq->termfreqs[*qit] = xrdb.get_termfreq(*qit) / doccnt;
LOGABS(("makeAbstract: [%s] db freq %.1e\n", qit->c_str(),
query->m_nq->termfreqs[*qit]));
}
LOGABS(("makeAbstract:%d: got termfreqs\n", chron.ms()));
}
// Compute a term quality coefficient by retrieving the term
// Within Document Frequencies and multiplying by overal term
// frequency, then using log-based thresholds. We are going to try
// and show text around the less common search terms.
map<string, double> termQcoefs;
double totalweight = 0;
double doclen = xrdb.get_doclength(docid);
if (doclen == 0) doclen = 1;
for (list<string>::const_iterator qit = terms.begin();
qit != terms.end(); qit++) {
Xapian::TermIterator term = xrdb.termlist_begin(docid);
term.skip_to(*qit);
if (term != xrdb.termlist_end(docid) && *term == *qit) {
double q = (term.get_wdf() / doclen) * query->m_nq->termfreqs[*qit];
q = -log10(q);
if (q < 3) {
q = 0.05;
} else if (q < 4) {
q = 0.3;
} else if (q < 5) {
q = 0.7;
} else if (q < 6) {
q = 0.8;
} else {
q = 1;
}
termQcoefs[*qit] = q;
totalweight += q;
}
}
LOGABS(("makeAbstract:%d: computed Qcoefs.\n", chron.ms()));
// Build a sorted by quality term list.
multimap<double, string> byQ;
for (list<string>::const_iterator qit = terms.begin();
qit != terms.end(); qit++) {
if (termQcoefs.find(*qit) != termQcoefs.end())
byQ.insert(pair<double,string>(termQcoefs[*qit], *qit));
}
#ifdef DEBUGABSTRACT
for (multimap<double, string>::reverse_iterator qit = byQ.rbegin();
qit != byQ.rend(); qit++) {
LOGDEB(("%.1e->[%s]\n", qit->first, qit->second.c_str()));
}
#endif
// For each of the query terms, ask xapian for its positions list
// in the document. For each position entry, remember it in
// qtermposs and insert it and its neighbours in the set of
// 'interesting' positions
// The terms 'array' that we partially populate with the document
// terms, at their positions around the search terms positions:
map<unsigned int, string> sparseDoc;
// All the chosen query term positions.
vector<unsigned int> qtermposs;
// Limit the total number of slots we populate. The 7 is taken as
// average word size. It was a mistake to have the user max
// abstract size parameter in characters, we basically only deal
// with words. We used to limit the character size at the end, but
// this damaged our careful selection of terms
const unsigned int maxtotaloccs =
m_rcldb->m_synthAbsLen /(7 * (m_rcldb->m_synthAbsWordCtxLen+1));
LOGABS(("makeAbstract:%d: mxttloccs %d\n", chron.ms(), maxtotaloccs));
// This can't happen, but would crash us
if (totalweight == 0.0) {
LOGERR(("makeAbstract: 0 totalweight!\n"));
return string();
}
// This is used to mark positions overlapped by a multi-word match term
const string occupiedmarker("?");
const string ellipsis("...");
// Let's go populate
for (multimap<double, string>::reverse_iterator qit = byQ.rbegin();
qit != byQ.rend(); qit++) {
string qterm = qit->second;
unsigned int maxoccs;
if (byQ.size() == 1) {
maxoccs = maxtotaloccs;
} else {
// We give more slots to the better terms
float q = qit->first / totalweight;
maxoccs = int(ceil(maxtotaloccs * q));
LOGABS(("makeAbstract: [%s] %d max occs (coef %.2f)\n",
qterm.c_str(), maxoccs, q));
}
// The match term may span several words
int qtrmwrdcnt = TextSplit::countWords(qterm, TextSplit::TXTS_NOSPANS);
Xapian::PositionIterator pos;
// There may be query terms not in this doc. This raises an
// exception when requesting the position list, we catch it.
string emptys;
try {
unsigned int occurrences = 0;
for (pos = xrdb.positionlist_begin(docid, qterm);
pos != xrdb.positionlist_end(docid, qterm); pos++) {
int ipos = *pos;
if (ipos < int(baseTextPosition)) // Not in text body
continue;
LOGABS(("makeAbstract: [%s] at %d occurrences %d maxoccs %d\n",
qterm.c_str(), ipos, occurrences, maxoccs));
// Remember the term position
qtermposs.push_back(ipos);
// Add adjacent slots to the set to populate at next
// step by inserting empty strings. Special provisions
// for adding ellipsis and for positions overlapped by
// the match term.
unsigned int sta = MAX(0, ipos-m_rcldb->m_synthAbsWordCtxLen);
unsigned int sto = ipos + qtrmwrdcnt-1 +
m_rcldb->m_synthAbsWordCtxLen;
for (unsigned int ii = sta; ii <= sto; ii++) {
if (ii == (unsigned int)ipos) {
sparseDoc[ii] = qterm;
} else if (ii > (unsigned int)ipos &&
ii < (unsigned int)ipos + qtrmwrdcnt) {
sparseDoc[ii] = occupiedmarker;
} else if (!sparseDoc[ii].compare(ellipsis)) {
// For an empty slot, the test has a side
// effect of inserting an empty string which
// is what we want
sparseDoc[ii] = emptys;
}
}
// Add ... at the end. This may be replaced later by
// an overlapping extract. Take care not to replace an
// empty string here, we really want an empty slot,
// use find()
if (sparseDoc.find(sto+1) == sparseDoc.end()) {
sparseDoc[sto+1] = ellipsis;
}
// Limit to allocated occurences and total size
if (++occurrences >= maxoccs ||
qtermposs.size() >= maxtotaloccs)
break;
}
} catch (...) {
// Term does not occur. No problem.
}
if (qtermposs.size() >= maxtotaloccs)
break;
}
LOGABS(("makeAbstract:%d:chosen number of positions %d\n",
chron.millis(), qtermposs.size()));
// This can happen if there are term occurences in the keywords
// etc. but not elsewhere ?
if (qtermposs.size() == 0)
return string();
// Walk all document's terms position lists 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 (missing words, potentially altering meaning),
// which is bad...
{
Xapian::TermIterator term;
int cutoff = 500 * 1000;
for (term = xrdb.termlist_begin(docid);
term != xrdb.termlist_end(docid); term++) {
// Ignore prefixed terms
if ('A' <= (*term).at(0) && (*term).at(0) <= 'Z')
continue;
if (cutoff-- < 0) {
LOGDEB0(("makeAbstract: max term count cutoff\n"));
break;
}
Xapian::PositionIterator pos;
for (pos = xrdb.positionlist_begin(docid, *term);
pos != xrdb.positionlist_end(docid, *term); pos++) {
if (cutoff-- < 0) {
LOGDEB0(("makeAbstract: max term count cutoff\n"));
break;
}
map<unsigned int, string>::iterator vit;
if ((vit=sparseDoc.find(*pos)) != sparseDoc.end()) {
// Don't replace a term: the terms list is in
// alphabetic order, and we may have several terms
// at the same position, we want to keep only the
// first one (ie: dockes and dockes@wanadoo.fr)
if (vit->second.empty()) {
LOGABS(("makeAbstract: populating: [%s] at %d\n",
(*term).c_str(), *pos));
sparseDoc[*pos] = *term;
}
}
}
}
}
#if 0
// Debug only: output the full term[position] vector
bool epty = false;
int ipos = 0;
for (map<unsigned int, string>::iterator it = sparseDoc.begin();
it != sparseDoc.end();
it++, ipos++) {
if (it->empty()) {
if (!epty)
LOGDEB(("makeAbstract:vec[%d]: [%s]\n", ipos, it->c_str()));
epty=true;
} else {
epty = false;
LOGDEB(("makeAbstract:vec[%d]: [%s]\n", ipos, it->c_str()));
}
}
#endif
LOGABS(("makeAbstract:%d: extracting\n", chron.millis()));
// Finally build the abstract by walking the map (in order of position)
string abstract;
abstract.reserve(sparseDoc.size() * 10);
bool incjk = false;
for (map<unsigned int, string>::const_iterator it = sparseDoc.begin();
it != sparseDoc.end(); it++) {
LOGDEB2(("Abtract:output %u -> [%s]\n", it->first,it->second.c_str()));
if (!occupiedmarker.compare(it->second))
continue;
Utf8Iter uit(it->second);
bool newcjk = false;
if (TextSplit::isCJK(*uit))
newcjk = true;
if (!incjk || (incjk && !newcjk))
abstract += " ";
incjk = newcjk;
abstract += it->second;
}
// This happens for docs with no terms (only filename) indexed? I'll fix
// one day (yeah)
if (!abstract.compare("... "))
abstract.clear();
LOGDEB2(("makeAbtract: done in %d mS\n", chron.millis()));
return abstract;
}
/* Rcl::Db methods ///////////////////////////////// */
Db::Db(RclConfig *cfp)
: m_ndb(0), m_config(cfp), m_idxAbsTruncLen(250), m_synthAbsLen(250),
m_synthAbsWordCtxLen(4), m_flushMb(-1),
m_curtxtsz(0), m_flushtxtsz(0), m_occtxtsz(0), m_occFirstCheck(1),
m_maxFsOccupPc(0), m_mode(Db::DbRO)
{
m_ndb = new Native(this);
if (m_config) {
m_config->getConfParam("maxfsoccuppc", &m_maxFsOccupPc);
m_config->getConfParam("idxflushmb", &m_flushMb);
}
}
Db::~Db()
{
LOGDEB2(("Db::~Db\n"));
if (m_ndb == 0)
return;
LOGDEB(("Db::~Db: isopen %d m_iswritable %d\n", m_ndb->m_isopen,
m_ndb->m_iswritable));
i_close(true);
}
list<string> Db::getStemmerNames()
{
list<string> res;
stringToStrings(Xapian::Stem::get_available_languages(), res);
return res;
}
bool Db::open(OpenMode mode, OpenError *error)
{
if (error)
*error = DbOpenMainDb;
if (m_ndb == 0 || m_config == 0) {
m_reason = "Null configuration or Xapian Db";
return false;
}
LOGDEB(("Db::open: m_isopen %d m_iswritable %d\n", m_ndb->m_isopen,
m_ndb->m_iswritable));
if (m_ndb->m_isopen) {
// We used to return an error here but I see no reason to
if (!close())
return false;
}
if (!m_config->getStopfile().empty())
m_stops.setFile(m_config->getStopfile());
string dir = m_config->getDbDir();
string ermsg;
try {
switch (mode) {
case DbUpd:
case DbTrunc:
{
int action = (mode == DbUpd) ? Xapian::DB_CREATE_OR_OPEN :
Xapian::DB_CREATE_OR_OVERWRITE;
m_ndb->xwdb = Xapian::WritableDatabase(dir, action);
// If db is empty, write the data format version at once
// to avoid stupid error messages:
if (m_ndb->xwdb.get_doccount() == 0)
m_ndb->xwdb.set_metadata(RCL_IDX_VERSION_KEY,
RCL_IDX_VERSION);
m_ndb->m_iswritable = true;
// We open a readonly object in all cases (possibly in
// addition to the r/w one) because some operations
// are faster when performed through a Database: no
// forced flushes on allterms_begin(), ie, used in
// subDocs()
m_ndb->xrdb = Xapian::Database(dir);
LOGDEB(("Db::open: lastdocid: %d\n",
m_ndb->xwdb.get_lastdocid()));
LOGDEB2(("Db::open: resetting updated\n"));
updated.resize(m_ndb->xwdb.get_lastdocid() + 1);
for (unsigned int i = 0; i < updated.size(); i++)
updated[i] = false;
}
break;
case DbRO:
default:
m_ndb->m_iswritable = false;
m_ndb->xrdb = Xapian::Database(dir);
for (list<string>::iterator it = m_extraDbs.begin();
it != m_extraDbs.end(); it++) {
if (error)
*error = DbOpenExtraDb;
LOGDEB(("Db::Open: adding query db [%s]\n", it->c_str()));
// An error here used to be non-fatal (1.13 and older)
// but I can't see why
m_ndb->xrdb.add_database(Xapian::Database(*it));
}
break;
}
if (error)
*error = DbOpenMainDb;
// Check index format version. Must not try to check a just created or
// truncated db
if (mode != DbTrunc && m_ndb->xdb().get_doccount() > 0) {
string version = m_ndb->xdb().get_metadata(RCL_IDX_VERSION_KEY);
if (version.compare(RCL_IDX_VERSION)) {
m_ndb->m_noversionwrite = true;
LOGERR(("Rcl::Db::open: file index [%s], software [%s]\n",
version.c_str(), RCL_IDX_VERSION.c_str()));
throw Xapian::DatabaseError("Recoll index version mismatch",
"", "");
}
}
m_mode = mode;
m_ndb->m_isopen = true;
m_basedir = dir;
if (error)
*error = DbOpenNoError;
return true;
} XCATCHERROR(ermsg);
m_reason = ermsg;
LOGERR(("Db::open: exception while opening [%s]: %s\n",
dir.c_str(), ermsg.c_str()));
return false;
}
// Note: xapian has no close call, we delete and recreate the db
bool Db::close()
{
LOGDEB2(("Db::close()\n"));
return i_close(false);
}
bool Db::i_close(bool final)
{
if (m_ndb == 0)
return false;
LOGDEB(("Db::i_close(%d): m_isopen %d m_iswritable %d\n", final,
m_ndb->m_isopen, m_ndb->m_iswritable));
if (m_ndb->m_isopen == false && !final)
return true;
string ermsg;
try {
bool w = m_ndb->m_iswritable;
if (w) {
if (!m_ndb->m_noversionwrite)
m_ndb->xwdb.set_metadata(RCL_IDX_VERSION_KEY, RCL_IDX_VERSION);
LOGDEB(("Rcl::Db:close: xapian will close. May take some time\n"));
}
// Used to do a flush here. Cant see why it should be necessary.
deleteZ(m_ndb);
if (w)
LOGDEB(("Rcl::Db:close() xapian close done.\n"));
if (final) {
return true;
}
m_ndb = new Native(this);
if (m_ndb) {
return true;
}
LOGERR(("Rcl::Db::close(): cant recreate db object\n"));
return false;
} XCATCHERROR(ermsg);
LOGERR(("Db:close: exception while deleting db: %s\n", ermsg.c_str()));
return false;
}
// Reopen the db with a changed list of additional dbs
bool Db::adjustdbs()
{
if (m_mode != DbRO) {
LOGERR(("Db::adjustdbs: mode not RO\n"));
return false;
}
if (m_ndb && m_ndb->m_isopen) {
if (!close())
return false;
if (!open(m_mode)) {
return false;
}
}
return true;
}
int Db::docCnt()
{
int res = -1;
if (!m_ndb || !m_ndb->m_isopen)
return -1;
XAPTRY(res = m_ndb->xdb().get_doccount(), m_ndb->xrdb, m_reason);
if (!m_reason.empty()) {
LOGERR(("Db::docCnt: got error: %s\n", m_reason.c_str()));
return -1;
}
return res;
}
bool Db::addQueryDb(const string &dir)
{
LOGDEB(("Db::addQueryDb: ndb %p iswritable %d db [%s]\n", m_ndb,
(m_ndb)?m_ndb->m_iswritable:0, dir.c_str()));
if (!m_ndb)
return false;
if (m_ndb->m_iswritable)
return false;
if (find(m_extraDbs.begin(), m_extraDbs.end(), dir) == m_extraDbs.end()) {
m_extraDbs.push_back(dir);
}
return adjustdbs();
}
bool Db::rmQueryDb(const string &dir)
{
if (!m_ndb)
return false;
if (m_ndb->m_iswritable)
return false;
if (dir.empty()) {
m_extraDbs.clear();
} else {
list<string>::iterator it = find(m_extraDbs.begin(),
m_extraDbs.end(), dir);
if (it != m_extraDbs.end()) {
m_extraDbs.erase(it);
}
}
return adjustdbs();
}
bool Db::testDbDir(const string &dir)
{
string aerr;
LOGDEB(("Db::testDbDir: [%s]\n", dir.c_str()));
try {
Xapian::Database db(dir);
} XCATCHERROR(aerr);
if (!aerr.empty()) {
LOGERR(("Db::Open: error while trying to open database "
"from [%s]: %s\n", dir.c_str(), aerr.c_str()));
return false;
}
return true;
}
bool Db::isopen()
{
if (m_ndb == 0)
return false;
return m_ndb->m_isopen;
}
// Try to translate field specification into field prefix. We have a
// default table used if translations are not in the config for some
// reason (old config not updated ?). We use it only if the config
// translation fails. Also we add in there fields which should be
// indexed with no prefix (ie: abstract)
// Default table. We prefer the data from rclconfig if available. Note
// that it is logically const after initialization. This would be
// simpler with c0xx initializer lists.
static map<string, string> fldToPrefs;
class InitFldToPrefs {
public:
InitFldToPrefs()
{
fldToPrefs[Doc::keyabs] = string();
fldToPrefs["ext"] = "XE";
fldToPrefs[Doc::keyfn] = "XSFN";
fldToPrefs[keycap] = "S";
fldToPrefs[Doc::keytt] = "S";
fldToPrefs["subject"] = "S";
fldToPrefs[Doc::keyau] = "A";
fldToPrefs["creator"] = "A";
fldToPrefs["from"] = "A";
fldToPrefs[Doc::keykw] = "K";
fldToPrefs["keyword"] = "K";
fldToPrefs["tag"] = "K";
fldToPrefs["tags"] = "K";
fldToPrefs["xapyear"] = "Y";
fldToPrefs["xapyearmon"] = "M";
fldToPrefs["xapdate"] = "D";
}
};
static InitFldToPrefs IFTP;
bool Db::fieldToPrefix(const string& fld, string &pfx)
{
if (m_config && m_config->getFieldPrefix(fld, pfx))
return true;
// No data in rclconfig? Check default values
map<string, string>::const_iterator it = fldToPrefs.find(fld);
if (it != fldToPrefs.end()) {
pfx = it->second;
return true;
}
return false;
}
// The splitter breaks text into words and adds postings to the Xapian document.
class TextSplitDb : public TextSplit {
public:
Xapian::Document &doc; // Xapian document
Xapian::termpos basepos; // Base for document section
Xapian::termpos curpos; // Current position. Used to set basepos for the
// following section
StopList &stops;
TextSplitDb(Xapian::Document &d, StopList &_stops)
: doc(d), basepos(1), curpos(0), stops(_stops)
{}
bool takeword(const std::string &term, int pos, int, int);
void setprefix(const string& pref) {prefix = pref;}
private:
// If prefix is set, we also add a posting for the prefixed terms
// (ie: for titles, add postings for both "term" and "Sterm")
string prefix;
};
// Get one term from the doc, remove accents and lowercase, then add posting
bool TextSplitDb::takeword(const std::string &_term, int pos, int, int)
{
LOGDEB2(("TextSplitDb::takeword: [%s]\n", _term.c_str()));
string term;
if (!unacmaybefold(_term, term, "UTF-8", true)) {
LOGINFO(("Db::splitter::takeword: unac failed for [%s]\n",
_term.c_str()));
term.clear();
// We don't generate a fatal error because of a bad term
return true;
}
if (stops.hasStops() && stops.isStop(term)) {
LOGDEB1(("Db: takeword [%s] in stop list\n", term.c_str()));
return true;
}
// Compute absolute position (pos is relative to current segment),
// and remember relative.
curpos = pos;
pos += basepos;
string 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
doc.add_posting(term, pos, 1);
if (!prefix.empty()) {
doc.add_posting(prefix + term, pos, 1);
}
return true;
} XCATCHERROR(ermsg);
LOGERR(("Db: xapian add_posting error %s\n", ermsg.c_str()));
return false;
}
// Let our user set the parameters for abstract processing
void Db::setAbstractParams(int idxtrunc, int syntlen, int syntctxlen)
{
LOGDEB1(("Db::setAbstractParams: trunc %d syntlen %d ctxlen %d\n",
idxtrunc, syntlen, syntctxlen));
if (idxtrunc > 0)
m_idxAbsTruncLen = idxtrunc;
if (syntlen > 0)
m_synthAbsLen = syntlen;
if (syntctxlen > 0)
m_synthAbsWordCtxLen = syntctxlen;
}
static const int MB = 1024 * 1024;
static const string nc("\n\r\x0c");
#define RECORD_APPEND(R, NM, VAL) {R += NM + "=" + VAL + "\n";}
// 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 Db::addOrUpdate(const string &udi, const string &parent_udi,
const Doc &idoc)
{
LOGDEB(("Db::add: udi [%s] parent [%s]\n",
udi.c_str(), parent_udi.c_str()));
if (m_ndb == 0)
return false;
// Check file system full every mbyte of indexed text.
if (m_maxFsOccupPc > 0 &&
(m_occFirstCheck || (m_curtxtsz - m_occtxtsz) / MB >= 1)) {
LOGDEB(("Db::add: checking file system usage\n"));
int pc;
m_occFirstCheck = 0;
if (fsocc(m_basedir, &pc) && pc >= m_maxFsOccupPc) {
LOGERR(("Db::add: stop indexing: file system "
"%d%% full > max %d%%\n", pc, m_maxFsOccupPc));
return false;
}
m_occtxtsz = m_curtxtsz;
}
Doc doc = idoc;
Xapian::Document newdocument;
TextSplitDb splitter(newdocument, m_stops);
// Split and index file name as document term(s)
LOGDEB2(("Db::add: split file name [%s]\n", fn.c_str()));
if (!splitter.text_to_words(doc.utf8fn))
LOGDEB(("Db::addOrUpdate: split failed for file name\n"));
splitter.basepos += splitter.curpos + 100;
// If the ipath is like a path, index the last element. This is
// for compound documents like zip and chm for which the filter
// uses the file path as ipath.
if (!doc.ipath.empty() &&
doc.ipath.find_first_not_of("0123456789") != string::npos) {
string utf8ipathlast;
// There is no way in hell we could have an idea of the
// charset here, so let's hope it's ascii or utf-8. We call
// transcode to strip the bad chars and pray
if (transcode(path_getsimple(doc.ipath), utf8ipathlast,
"UTF-8", "UTF-8")) {
splitter.text_to_words(utf8ipathlast);
splitter.basepos += splitter.curpos + 100;
}
}
// Split and index the path from the url for path-based filtering
{
string path = url_gpath(doc.url);
vector<string> vpath;
stringToTokens(path, vpath, "/");
splitter.curpos = 0;
newdocument.add_posting(pathelt_prefix,
splitter.basepos + splitter.curpos++);
for (vector<string>::iterator it = vpath.begin();
it != vpath.end(); it++){
if (it->length() > 230) {
// Just truncate it. May still be useful because of wildcards
*it = it->substr(0, 230);
}
newdocument.add_posting(pathelt_prefix + *it,
splitter.basepos + splitter.curpos++);
}
}
splitter.basepos += splitter.curpos + 100;
// Index textual metadata. These are all indexed as text with
// positions, as we may want to do phrase searches with them (this
// makes no sense for keywords by the way).
//
// The order has no importance, and we set a position gap of 100
// between fields to avoid false proximity matches.
map<string,string>::iterator meta_it;
string pfx;
for (meta_it = doc.meta.begin(); meta_it != doc.meta.end(); meta_it++) {
if (!meta_it->second.empty()) {
if (!fieldToPrefix(meta_it->first, pfx)) {
LOGDEB0(("Db::add: no prefix for field [%s], no indexing\n",
meta_it->first.c_str()));
continue;
}
LOGDEB0(("Db::add: field [%s] pfx [%s]: [%s]\n",
meta_it->first.c_str(), pfx.c_str(),
meta_it->second.c_str()));
splitter.setprefix(pfx); // Subject
if (!splitter.text_to_words(meta_it->second))
LOGDEB(("Db::addOrUpdate: split failed for %s\n",
meta_it->first.c_str()));
splitter.setprefix(string());
splitter.basepos += splitter.curpos + 100;
}
}
if (splitter.curpos < baseTextPosition)
splitter.basepos = baseTextPosition;
else
splitter.basepos += splitter.curpos + 100;
// Split and index body text
LOGDEB2(("Db::add: split body\n"));
if (!splitter.text_to_words(doc.text))
LOGDEB(("Db::addOrUpdate: split failed for main text\n"));
////// Special terms for other metadata. No positions for these.
// Mime type
newdocument.add_term("T" + doc.mimetype);
// Simple file name indexed for file name searches with a term prefix
// We also add a term for the filename extension if any.
if (!doc.utf8fn.empty()) {
string fn;
if (unacmaybefold(doc.utf8fn, fn, "UTF-8", true)) {
// We should truncate after extracting the extension, but this is
// a pathological case anyway
if (fn.size() > 230)
utf8truncate(fn, 230);
string::size_type pos = fn.rfind('.');
if (pos != string::npos && pos != fn.length() - 1) {
newdocument.add_term(string("XE") + fn.substr(pos + 1));
}
fn = string("XSFN") + fn;
newdocument.add_term(fn);
}
// Store utf8fn inside the metadata array as keyfn
// (="filename") so that it can be accessed by the "stored"
// processing below, without special-casing it. We only do it
// if keyfn is currently empty, because there could be a value
// already (ie for a mail attachment with a file name
// attribute)
if (doc.meta[Doc::keyfn].empty()) {
doc.meta[Doc::keyfn] = doc.utf8fn;
}
}
// Udi unique term: this is used for file existence/uptodate
// checks, and unique id for the replace_document() call.
string uniterm = make_uniterm(udi);
newdocument.add_term(uniterm);
// Parent term. This is used to find all descendents, mostly to delete them
// when the parent goes away
if (!parent_udi.empty()) {
newdocument.add_term(make_parentterm(parent_udi));
}
// 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];
snprintf(buf, 9, "%04d%02d%02d",
tm->tm_year+1900, tm->tm_mon + 1, tm->tm_mday);
newdocument.add_term("D" + string(buf)); // Date (YYYYMMDD)
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
//
// The title, author, abstract and keywords fields are special,
// they always get stored in the document data
// record. Configurable other fields can be, too.
//
// We truncate stored fields abstract, title and keywords to
// reasonable lengths and suppress newlines (so that the data
// record can keep a simple syntax)
string record;
RECORD_APPEND(record, Doc::keyurl, doc.url);
RECORD_APPEND(record, Doc::keytp, doc.mimetype);
// We left-zero-pad the times so that they are lexico-sortable
leftzeropad(doc.fmtime, 11);
RECORD_APPEND(record, Doc::keyfmt, doc.fmtime);
if (!doc.dmtime.empty()) {
leftzeropad(doc.dmtime, 11);
RECORD_APPEND(record, Doc::keydmt, doc.dmtime);
}
RECORD_APPEND(record, Doc::keyoc, doc.origcharset);
if (!doc.fbytes.empty())
RECORD_APPEND(record, Doc::keyfs, doc.fbytes);
// Note that we add the signature both as a value and in the data record
if (!doc.sig.empty())
RECORD_APPEND(record, Doc::keysig, doc.sig);
newdocument.add_value(VALUE_SIG, doc.sig);
char sizebuf[30];
sprintf(sizebuf, "%u", (unsigned int)doc.text.length());
RECORD_APPEND(record, Doc::keyds, sizebuf);
if (!doc.ipath.empty())
RECORD_APPEND(record, Doc::keyipt, doc.ipath);
if (doc.meta[Doc::keytt].empty())
doc.meta[Doc::keytt] = doc.utf8fn;
doc.meta[Doc::keytt] =
neutchars(truncate_to_word(doc.meta[Doc::keytt], 150), nc);
if (!doc.meta[Doc::keytt].empty())
RECORD_APPEND(record, keycap, doc.meta[Doc::keytt]);
doc.meta[Doc::keykw] =
neutchars(truncate_to_word(doc.meta[Doc::keykw], 300), nc);
if (!doc.meta[Doc::keykw].empty())
RECORD_APPEND(record, Doc::keykw, doc.meta[Doc::keykw]);
// If abstract is empty, we make up one with the beginning of the
// document. This is then not indexed, but part of the doc data so
// that we can return it to a query without having to decode the
// original file.
bool syntabs = false;
// Note that the map accesses by operator[] create empty entries if they
// don't exist yet.
if (doc.meta[Doc::keyabs].empty()) {
syntabs = true;
if (!doc.text.empty())
doc.meta[Doc::keyabs] = rclSyntAbs +
neutchars(truncate_to_word(doc.text, m_idxAbsTruncLen), nc);
} else {
doc.meta[Doc::keyabs] =
neutchars(truncate_to_word(doc.meta[Doc::keyabs], m_idxAbsTruncLen),
nc);
}
if (!doc.meta[Doc::keyabs].empty())
RECORD_APPEND(record, Doc::keyabs, doc.meta[Doc::keyabs]);
const set<string>& stored = m_config->getStoredFields();
for (set<string>::const_iterator it = stored.begin();
it != stored.end(); it++) {
string nm = m_config->fieldCanon(*it);
if (!doc.meta[*it].empty()) {
string value =
neutchars(truncate_to_word(doc.meta[*it], 150), nc);
RECORD_APPEND(record, nm, value);
}
}
// If the file's md5 was computed, add value. This is optionally
// used for query result duplicate elimination.
string& md5 = doc.meta[Doc::keymd5];
if (!md5.empty()) {
string digest;
MD5HexScan(md5, digest);
newdocument.add_value(VALUE_MD5, digest);
}
LOGDEB0(("Rcl::Db::add: new doc record:\n%s\n", record.c_str()));
newdocument.set_data(record);
const char *fnc = udi.c_str();
string ermsg;
// Add db entry or update existing entry:
try {
Xapian::docid did =
m_ndb->xwdb.replace_document(uniterm, newdocument);
if (did < updated.size()) {
updated[did] = true;
LOGDEB(("Db::add: docid %d updated [%s]\n", did, fnc));
} else {
LOGDEB(("Db::add: docid %d added [%s]\n", did, fnc));
}
} XCATCHERROR(ermsg);
if (!ermsg.empty()) {
LOGERR(("Db::add: replace_document failed: %s\n", ermsg.c_str()));
ermsg.erase();
// FIXME: is this ever actually needed?
try {
m_ndb->xwdb.add_document(newdocument);
LOGDEB(("Db::add: %s added (failed re-seek for duplicate)\n",
fnc));
} XCATCHERROR(ermsg);
if (!ermsg.empty()) {
LOGERR(("Db::add: add_document failed: %s\n", ermsg.c_str()));
return false;
}
}
// Test if we're over the flush threshold (limit memory usage):
m_curtxtsz += doc.text.length();
if (m_flushMb > 0) {
if ((m_curtxtsz - m_flushtxtsz) / MB >= m_flushMb) {
ermsg.erase();
LOGDEB(("Db::add: text size >= %d Mb, flushing\n", m_flushMb));
try {
m_ndb->xwdb.flush();
} XCATCHERROR(ermsg);
if (!ermsg.empty()) {
LOGERR(("Db::add: flush() failed: %s\n", ermsg.c_str()));
return false;
}
m_flushtxtsz = m_curtxtsz;
}
}
return true;
}
// Test if doc given by udi has changed since last indexed (test sigs)
bool Db::needUpdate(const string &udi, const string& sig)
{
if (m_ndb == 0)
return false;
string uniterm = make_uniterm(udi);
string ermsg;
// We look up the document indexed by the uniterm. This is either
// the actual document file, or, for a multi-document file, the
// pseudo-doc we create to stand for the file itself.
// We try twice in case database needs to be reopened.
for (int tries = 0; tries < 2; tries++) {
try {
// Get the doc or pseudo-doc
Xapian::PostingIterator docid =m_ndb->xrdb.postlist_begin(uniterm);
if (docid == m_ndb->xrdb.postlist_end(uniterm)) {
// If no document exist with this path, we do need update
LOGDEB(("Db::needUpdate:yes (new): [%s]\n", uniterm.c_str()));
return true;
}
Xapian::Document doc = m_ndb->xrdb.get_document(*docid);
// Retrieve old file/doc signature from value
string osig = doc.get_value(VALUE_SIG);
LOGDEB2(("Db::needUpdate: oldsig [%s] new [%s]\n",
osig.c_str(), sig.c_str()));
// Compare new/old sig
if (sig != osig) {
LOGDEB(("Db::needUpdate:yes: olsig [%s] new [%s]\n",
osig.c_str(), sig.c_str()));
// Db is not up to date. Let's index the file
return true;
}
LOGDEB(("Db::needUpdate:no: [%s]\n", uniterm.c_str()));
// Up to date.
// Set the uptodate flag for doc / pseudo doc
updated[*docid] = true;
// Set the existence flag for all the subdocs (if any)
vector<Xapian::docid> docids;
if (!m_ndb->subDocs(udi, docids)) {
LOGERR(("Rcl::Db::needUpdate: can't get subdocs list\n"));
return true;
}
for (vector<Xapian::docid>::iterator it = docids.begin();
it != docids.end(); it++) {
if (*it < updated.size()) {
LOGDEB2(("Db::needUpdate: set flag for docid %d\n", *it));
updated[*it] = true;
}
}
return false;
} catch (const Xapian::DatabaseModifiedError &e) {
LOGDEB(("Db::needUpdate: got modified error. reopen/retry\n"));
m_reason = e.get_msg();
m_ndb->xrdb.reopen();
continue;
} XCATCHERROR(m_reason);
break;
}
LOGERR(("Db::needUpdate: error while checking existence: %s\n",
m_reason.c_str()));
return true;
}
// Return list of existing stem db languages
list<string> Db::getStemLangs()
{
LOGDEB(("Db::getStemLang\n"));
list<string> dirs;
if (m_ndb == 0 || m_ndb->m_isopen == false)
return dirs;
dirs = StemDb::getLangs(m_basedir);
return dirs;
}
/**
* Delete stem db for given language
*/
bool Db::deleteStemDb(const string& lang)
{
LOGDEB(("Db::deleteStemDb(%s)\n", lang.c_str()));
if (m_ndb == 0 || m_ndb->m_isopen == false)
return false;
return StemDb::deleteDb(m_basedir, lang);
}
/**
* 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 Db::createStemDb(const string& lang)
{
LOGDEB(("Db::createStemDb(%s)\n", lang.c_str()));
if (m_ndb == 0 || m_ndb->m_isopen == false)
return false;
return StemDb::createDb(m_ndb->xdb(), m_basedir, lang);
}
/**
* This is called at the end of an indexing session, to delete the
* documents for files that are no longer there. This can ONLY be called
* after a full file-system tree walk, else the file existence flags will
* be wrong.
*/
bool Db::purge()
{
LOGDEB(("Db::purge\n"));
if (m_ndb == 0)
return false;
LOGDEB(("Db::purge: m_isopen %d m_iswritable %d\n", m_ndb->m_isopen,
m_ndb->m_iswritable));
if (m_ndb->m_isopen == false || m_ndb->m_iswritable == false)
return false;
// For xapian versions up to 1.0.1, deleting a non-existant
// document would trigger an exception that would discard any
// pending update. This could lose both previous added documents
// or deletions. Adding the flush before the delete pass ensured
// that any added document would go to the index. Kept here
// because it doesn't really hurt.
try {
m_ndb->xwdb.flush();
} catch (...) {
LOGERR(("Db::purge: 1st flush failed\n"));
}
// Walk the document array and delete any xapian document whose
// flag is not set (we did not see its source during indexing).
int purgecount = 0;
for (Xapian::docid docid = 1; docid < updated.size(); ++docid) {
if (!updated[docid]) {
if ((purgecount+1) % 100 == 0) {
try {
CancelCheck::instance().checkCancel();
} catch(CancelExcept) {
LOGINFO(("Db::purge: partially cancelled\n"));
break;
}
}
try {
m_ndb->xwdb.delete_document(docid);
LOGDEB(("Db::purge: deleted document #%d\n", docid));
} catch (const Xapian::DocNotFoundError &) {
LOGDEB0(("Db::purge: document #%d not found\n", docid));
} catch (const Xapian::Error &e) {
LOGERR(("Db::purge: document #%d: %s\n", docid, e.get_msg().c_str()));
} catch (...) {
LOGERR(("Db::purge: document #%d: unknown error\n", docid));
}
purgecount++;
}
}
try {
m_ndb->xwdb.flush();
} catch (...) {
LOGERR(("Db::purge: 2nd flush failed\n"));
}
return true;
}
/* Delete document(s) for given unique identifier (doc and descendents) */
bool Db::purgeFile(const string &udi, bool *existed)
{
LOGDEB(("Db:purgeFile: [%s]\n", udi.c_str()));
if (m_ndb == 0 || !m_ndb->m_iswritable)
return false;
Xapian::WritableDatabase db = m_ndb->xwdb;
string uniterm = make_uniterm(udi);
string ermsg;
try {
Xapian::PostingIterator docid = db.postlist_begin(uniterm);
if (docid == db.postlist_end(uniterm)) {
if (existed)
*existed = false;
return true;
}
*existed = true;
LOGDEB(("purgeFile: delete docid %d\n", *docid));
db.delete_document(*docid);
vector<Xapian::docid> docids;
m_ndb->subDocs(udi, docids);
LOGDEB(("purgeFile: subdocs cnt %d\n", docids.size()));
for (vector<Xapian::docid>::iterator it = docids.begin();
it != docids.end(); it++) {
LOGDEB(("Db::purgeFile: delete subdoc %d\n", *it));
db.delete_document(*it);
}
return true;
} XCATCHERROR(ermsg);
if (!ermsg.empty()) {
LOGERR(("Db::purgeFile: %s\n", ermsg.c_str()));
}
return false;
}
// File name wild card expansion. This is a specialisation ot termMatch
bool Db::filenameWildExp(const string& fnexp, list<string>& names)
{
string pattern = fnexp;
names.clear();
// If pattern is not capitalized, not quoted (quoted pattern can't
// get here currently anyway), 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 &&
!unaciscapital(pattern)) {
pattern = "*" + pattern + "*";
} // else let it be
LOGDEB(("Rcl::Db::filenameWildExp: pattern: [%s]\n", pattern.c_str()));
TermMatchResult result;
if (!termMatch(ET_WILD, string(), pattern, result, 1000, Doc::keyfn))
return false;
for (list<TermMatchEntry>::const_iterator it = result.entries.begin();
it != result.entries.end(); it++)
names.push_back(it->term);
if (names.empty()) {
// Build an impossible query: we know its impossible because we
// control the prefixes!
names.push_back("XNONENoMatchingTerms");
}
return true;
}
// Walk the Y terms and return min/max
bool Db::maxYearSpan(int *minyear, int *maxyear)
{
*minyear = 1000000;
*maxyear = -1000000;
TermMatchResult result;
if (!termMatch(ET_WILD, string(), "*", result, 5000, "xapyear"))
return false;
for (list<TermMatchEntry>::const_iterator it = result.entries.begin();
it != result.entries.end(); it++) {
if (!it->term.empty()) {
int year = atoi(it->term.c_str()+1);
if (year < *minyear)
*minyear = year;
if (year > *maxyear)
*maxyear = year;
}
}
return true;
}
class TermMatchCmpByWcf {
public:
int operator()(const TermMatchEntry& l, const TermMatchEntry& r) {
return r.wcf - l.wcf < 0;
}
};
class TermMatchCmpByTerm {
public:
int operator()(const TermMatchEntry& l, const TermMatchEntry& r) {
return l.term.compare(r.term) > 0;
}
};
class TermMatchTermEqual {
public:
int operator()(const TermMatchEntry& l, const TermMatchEntry& r) {
return !l.term.compare(r.term);
}
};
bool Db::stemExpand(const string &lang, const string &term,
TermMatchResult& result, int max)
{
list<string> dirs = m_extraDbs;
dirs.push_front(m_basedir);
for (list<string>::iterator it = dirs.begin(); it != dirs.end(); it++) {
list<string> more;
StemDb::stemExpand(*it, lang, term, more);
LOGDEB1(("Db::stemExpand: Got %d from %s\n",
more.size(), it->c_str()));
result.entries.insert(result.entries.end(), more.begin(), more.end());
if (result.entries.size() >= (unsigned int)max)
break;
}
LOGDEB1(("Db:::stemExpand: final count %d \n", result.size()));
return true;
}
/** Add prefix to all strings in list */
static void addPrefix(list<TermMatchEntry>& terms, const string& prefix)
{
if (prefix.empty())
return;
for (list<TermMatchEntry>::iterator it = terms.begin();
it != terms.end(); it++)
it->term.insert(0, prefix);
}
// Characters that can begin a wildcard or regexp expression. We use skipto
// to begin the allterms search with terms that begin with the portion of
// the input string prior to these chars.
const string wildSpecChars = "*?[";
const string regSpecChars = "(.[{";
// Find all index terms that match a wildcard or regular expression
bool Db::termMatch(MatchType typ, const string &lang,
const string &root,
TermMatchResult& res,
int max,
const string& field,
string *prefixp
)
{
if (!m_ndb || !m_ndb->m_isopen)
return false;
Xapian::Database xdb = m_ndb->xdb();
res.clear();
XAPTRY(res.dbdoccount = xdb.get_doccount();
res.dbavgdoclen = xdb.get_avlength(), xdb, m_reason);
if (!m_reason.empty())
return false;
// Get rid of capitals and accents
string droot;
if (!unacmaybefold(root, droot, "UTF-8", true)) {
LOGERR(("Db::termMatch: unac failed for [%s]\n", root.c_str()));
return false;
}
string nochars = typ == ET_WILD ? wildSpecChars : regSpecChars;
string prefix;
if (!field.empty()) {
(void)fieldToPrefix(field, prefix);
if (prefix.empty()) {
LOGDEB(("Db::termMatch: field is not indexed (no prefix): [%s]\n",
field.c_str()));
}
if (prefixp)
*prefixp = prefix;
}
if (typ == ET_STEM) {
if (!stemExpand(lang, root, res, max))
return false;
res.entries.sort();
res.entries.unique();
for (list<TermMatchEntry>::iterator it = res.entries.begin();
it != res.entries.end(); it++) {
XAPTRY(it->wcf = xdb.get_collection_freq(it->term);
it->docs = xdb.get_termfreq(it->term),
xdb, m_reason);
if (!m_reason.empty())
return false;
LOGDEB1(("termMatch: %d [%s]\n", it->wcf, it->term.c_str()));
}
if (!prefix.empty())
addPrefix(res.entries, prefix);
} else {
regex_t reg;
int errcode;
if (typ == ET_REGEXP) {
string mroot = droot;
if ((errcode = regcomp(&reg, mroot.c_str(),
REG_EXTENDED|REG_NOSUB))) {
char errbuf[200];
regerror(errcode, &reg, errbuf, 199);
LOGERR(("termMatch: regcomp failed: %s\n", errbuf));
res.entries.push_back(string(errbuf));
regfree(&reg);
return false;
}
}
// Find the initial section before any special char
string::size_type es = droot.find_first_of(nochars);
string is;
switch (es) {
case string::npos: is = prefix + droot; break;
case 0: is = prefix; break;
default: is = prefix + droot.substr(0, es); break;
}
LOGDEB(("termMatch: initsec: [%s]\n", is.c_str()));
for (int tries = 0; tries < 2; tries++) {
try {
Xapian::TermIterator it = xdb.allterms_begin();
if (!is.empty())
it.skip_to(is.c_str());
for (int n = 0; it != xdb.allterms_end(); it++) {
// If we're beyond the terms matching the initial
// string, end
if (!is.empty() && (*it).find(is) != 0)
break;
string term;
if (!prefix.empty())
term = (*it).substr(prefix.length());
else
term = *it;
if (typ == ET_WILD) {
if (fnmatch(droot.c_str(), term.c_str(), 0) ==
FNM_NOMATCH)
continue;
} else {
if (regexec(&reg, term.c_str(), 0, 0, 0))
continue;
}
// Do we want stem expansion here? We don't do it for now
res.entries.push_back(TermMatchEntry(*it,
xdb.get_collection_freq(*it),
it.get_termfreq()));
++n;
}
m_reason.erase();
break;
} catch (const Xapian::DatabaseModifiedError &e) {
m_reason = e.get_msg();
xdb.reopen();
continue;
} XCATCHERROR(m_reason);
break;
}
if (!m_reason.empty()) {
LOGERR(("termMatch: %s\n", m_reason.c_str()));
return false;
}
if (typ == ET_REGEXP) {
regfree(&reg);
}
}
TermMatchCmpByTerm tcmp;
res.entries.sort(tcmp);
TermMatchTermEqual teq;
res.entries.unique(teq);
TermMatchCmpByWcf wcmp;
res.entries.sort(wcmp);
if (max > 0) {
res.entries.resize(MIN(res.entries.size(), (unsigned int)max));
}
return true;
}
/** Term list walking. */
class TermIter {
public:
Xapian::TermIterator it;
Xapian::Database db;
};
TermIter *Db::termWalkOpen()
{
if (!m_ndb || !m_ndb->m_isopen)
return 0;
TermIter *tit = new TermIter;
if (tit) {
tit->db = m_ndb->xdb();
XAPTRY(tit->it = tit->db.allterms_begin(), tit->db, m_reason);
if (!m_reason.empty()) {
LOGERR(("Db::termWalkOpen: xapian error: %s\n", m_reason.c_str()));
return 0;
}
}
return tit;
}
bool Db::termWalkNext(TermIter *tit, string &term)
{
XAPTRY(
if (tit && tit->it != tit->db.allterms_end()) {
term = *(tit->it)++;
return true;
}
, tit->db, m_reason);
if (!m_reason.empty()) {
LOGERR(("Db::termWalkOpen: xapian error: %s\n", m_reason.c_str()));
}
return false;
}
void Db::termWalkClose(TermIter *tit)
{
try {
delete tit;
} catch (...) {}
}
bool Db::termExists(const string& word)
{
if (!m_ndb || !m_ndb->m_isopen)
return 0;
XAPTRY(if (!m_ndb->xdb().term_exists(word)) return false,
m_ndb->xrdb, m_reason);
if (!m_reason.empty()) {
LOGERR(("Db::termWalkOpen: xapian error: %s\n", m_reason.c_str()));
return false;
}
return true;
}
bool Db::stemDiffers(const string& lang, const string& word,
const string& base)
{
Xapian::Stem stemmer(lang);
if (!stemmer(word).compare(stemmer(base))) {
LOGDEB2(("Rcl::Db::stemDiffers: same for %s and %s\n",
word.c_str(), base.c_str()));
return false;
}
return true;
}
bool Db::makeDocAbstract(Doc &doc, Query *query, string& abstract)
{
LOGDEB1(("Db::makeDocAbstract: exti %d\n", exti));
if (!m_ndb || !m_ndb->m_isopen) {
LOGERR(("Db::makeDocAbstract: no db\n"));
return false;
}
XAPTRY(abstract = m_ndb->makeAbstract(doc.xdocid, query),
m_ndb->xrdb, m_reason);
return m_reason.empty() ? true : false;
}
// Retrieve document defined by Unique doc identifier. This is mainly used
// by the GUI history feature
bool Db::getDoc(const string &udi, Doc &doc)
{
LOGDEB(("Db:getDoc: [%s]\n", udi.c_str()));
if (m_ndb == 0)
return false;
// Initialize what we can in any case. If this is history, caller
// will make partial display in case of error
doc.meta[Rcl::Doc::keyrr] = "100%";
doc.pc = 100;
string uniterm = make_uniterm(udi);
for (int tries = 0; tries < 2; tries++) {
try {
if (!m_ndb->xrdb.term_exists(uniterm)) {
// 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
doc.pc = -1;
LOGINFO(("Db:getDoc: no such doc in index: [%s] (len %d)\n",
uniterm.c_str(), uniterm.length()));
return true;
}
Xapian::PostingIterator docid =
m_ndb->xrdb.postlist_begin(uniterm);
Xapian::Document xdoc = m_ndb->xrdb.get_document(*docid);
string data = xdoc.get_data();
doc.meta[Rcl::Doc::keyudi] = udi;
return m_ndb->dbDataToRclDoc(*docid, data, doc);
} catch (const Xapian::DatabaseModifiedError &e) {
m_reason = e.get_msg();
m_ndb->xrdb.reopen();
continue;
} XCATCHERROR(m_reason);
break;
}
LOGERR(("Db::getDoc: %s\n", m_reason.c_str()));
return false;
}
#ifndef NO_NAMESPACES
}
#endif