Parent: [525fff] (diff)

Child: [bde991] (diff)

Download this file

rcldb.cpp    2210 lines (1999 with data), 69.9 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
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
/* 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 "autoconfig.h"
#include <stdio.h>
#include <cstring>
#include "safeunistd.h"
#include <math.h>
#include <time.h>
#include <string>
#include <vector>
#include <algorithm>
#include <sstream>
#include <iostream>
using namespace std;
#include "xapian.h"
#include "rclconfig.h"
#include "log.h"
#include "rcldb.h"
#include "rcldb_p.h"
#include "stemdb.h"
#include "textsplit.h"
#include "transcode.h"
#include "unacpp.h"
#include "conftree.h"
#include "pathut.h"
#include "rclutil.h"
#include "smallut.h"
#include "chrono.h"
#include "utf8iter.h"
#include "searchdata.h"
#include "rclquery.h"
#include "rclquery_p.h"
#include "md5ut.h"
#include "rclversion.h"
#include "cancelcheck.h"
#include "termproc.h"
#include "expansiondbs.h"
#include "rclinit.h"
#include "internfile.h"
#include "utf8fn.h"
// 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 cstr_RCL_IDX_VERSION_KEY("RCL_IDX_VERSION_KEY");
static const string cstr_RCL_IDX_VERSION("1");
static const string cstr_mbreaks("rclmbreaks");
namespace Rcl {
// Some prefixes that we could get from the fields file, but are not going
// to ever change.
static const string fileext_prefix = "XE";
const string mimetype_prefix = "T";
static const string xapday_prefix = "D";
static const string xapmonth_prefix = "M";
static const string xapyear_prefix = "Y";
const string pathelt_prefix = "XP";
static const string udi_prefix("Q");
static const string parent_prefix("F");
// Special terms to mark begin/end of field (for anchored searches), and
// page breaks
string start_of_field_term;
string end_of_field_term;
const string page_break_term = "XXPG/";
// Special term to mark documents with children.
const string has_children_term("XXC/");
// Field name for the unsplit file name. Has to exist in the field file
// because of usage in termmatch()
const string unsplitFilenameFieldName = "rclUnsplitFN";
static const string unsplitfilename_prefix = "XSFS";
// Empty string md5s
static const string cstr_md5empty("d41d8cd98f00b204e9800998ecf8427e");
static const int MB = 1024 * 1024;
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 cstr_syntAbs("?!#@");
// Compute the unique term used to link documents to their origin.
// "Q" + external udi
static inline string make_uniterm(const string& udi)
{
string uniterm(wrap_prefix(udi_prefix));
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(wrap_prefix(parent_prefix));
pterm.append(udi);
return pterm;
}
static void utf8truncate(string& s, int maxlen)
{
if (s.size() <= string::size_type(maxlen)) {
return;
}
Utf8Iter iter(s);
string::size_type pos = 0;
while (iter++ != string::npos)
if (iter.getBpos() < string::size_type(maxlen)) {
pos = iter.getBpos();
}
s.erase(pos);
}
Db::Native::Native(Db *db)
: m_rcldb(db), m_isopen(false), m_iswritable(false),
m_noversionwrite(false)
#ifdef IDX_THREADS
, m_wqueue("DbUpd",
m_rcldb->m_config->getThrConf(RclConfig::ThrDbWrite).first),
m_totalworkns(0LL), m_havewriteq(false)
#endif // IDX_THREADS
{
LOGDEB1("Native::Native: me " << (this) << "\n" );
}
Db::Native::~Native()
{
LOGDEB1("Native::~Native: me " << (this) << "\n" );
#ifdef IDX_THREADS
if (m_havewriteq) {
void *status = m_wqueue.setTerminateAndWait();
if (status) {
LOGDEB1("Native::~Native: worker status " << status << "\n");
}
}
#endif // IDX_THREADS
}
#ifdef IDX_THREADS
void *DbUpdWorker(void* vdbp)
{
recoll_threadinit();
Db::Native *ndbp = (Db::Native *)vdbp;
WorkQueue<DbUpdTask*> *tqp = &(ndbp->m_wqueue);
DbUpdTask *tsk = 0;
for (;;) {
size_t qsz = -1;
if (!tqp->take(&tsk, &qsz)) {
tqp->workerExit();
return (void*)1;
}
bool status = false;
switch (tsk->op) {
case DbUpdTask::AddOrUpdate:
LOGDEB("DbUpdWorker: got add/update task, ql " << (int(qsz)) << "\n" );
status = ndbp->addOrUpdateWrite(tsk->udi, tsk->uniterm,
tsk->doc, tsk->txtlen);
break;
case DbUpdTask::Delete:
LOGDEB("DbUpdWorker: got delete task, ql " << (int(qsz)) << "\n" );
status = ndbp->purgeFileWrite(false, tsk->udi, tsk->uniterm);
break;
case DbUpdTask::PurgeOrphans:
LOGDEB("DbUpdWorker: got orphans purge task, ql " << (int(qsz)) << "\n" );
status = ndbp->purgeFileWrite(true, tsk->udi, tsk->uniterm);
break;
default:
LOGERR("DbUpdWorker: unknown op " << (tsk->op) << " !!\n" );
break;
}
if (!status) {
LOGERR("DbUpdWorker: xxWrite failed\n" );
tqp->workerExit();
delete tsk;
return (void*)0;
}
delete tsk;
}
}
void Db::Native::maybeStartThreads()
{
m_havewriteq = false;
const RclConfig *cnf = m_rcldb->m_config;
int writeqlen = cnf->getThrConf(RclConfig::ThrDbWrite).first;
int writethreads = cnf->getThrConf(RclConfig::ThrDbWrite).second;
if (writethreads > 1) {
LOGINFO("RclDb: write threads count was forced down to 1\n" );
writethreads = 1;
}
if (writeqlen >= 0 && writethreads > 0) {
if (!m_wqueue.start(writethreads, DbUpdWorker, this)) {
LOGERR("Db::Db: Worker start failed\n" );
return;
}
m_havewriteq = true;
}
LOGDEB("RclDb:: threads: haveWriteQ " << (m_havewriteq) << ", wqlen " << (writeqlen) << " wqts " << (writethreads) << "\n" );
}
#endif // IDX_THREADS
/* See comment in class declaration: return all subdocuments of a
* document given by its unique id.
*/
bool Db::Native::subDocs(const string &udi, int idxi,
vector<Xapian::docid>& docids)
{
LOGDEB2("subDocs: [" << (uniterm) << "]\n" );
string pterm = make_parentterm(udi);
vector<Xapian::docid> candidates;
XAPTRY(docids.clear();
candidates.insert(candidates.begin(), xrdb.postlist_begin(pterm),
xrdb.postlist_end(pterm)),
xrdb, m_rcldb->m_reason);
if (!m_rcldb->m_reason.empty()) {
LOGERR("Rcl::Db::subDocs: " << (m_rcldb->m_reason) << "\n" );
return false;
} else {
for (unsigned int i = 0; i < candidates.size(); i++) {
if (whatDbIdx(candidates[i]) == (size_t)idxi) {
docids.push_back(candidates[i]);
}
}
LOGDEB0("Db::Native::subDocs: returning " << (docids.size()) << " ids\n" );
return true;
}
}
bool Db::Native::xdocToUdi(Xapian::Document& xdoc, string &udi)
{
Xapian::TermIterator xit;
XAPTRY(xit = xdoc.termlist_begin();
xit.skip_to(wrap_prefix(udi_prefix)),
xrdb, m_rcldb->m_reason);
if (!m_rcldb->m_reason.empty()) {
LOGERR("xdocToUdi: xapian error: " << (m_rcldb->m_reason) << "\n" );
return false;
}
if (xit != xdoc.termlist_end()) {
udi = *xit;
if (!udi.empty()) {
udi = udi.substr(wrap_prefix(udi_prefix).size());
return true;
}
}
return false;
}
// Clear term from document if its frequency is 0. This should
// probably be done by Xapian when the freq goes to 0 when removing a
// posting, but we have to do it ourselves
bool Db::Native::clearDocTermIfWdf0(Xapian::Document& xdoc, const string& term)
{
LOGDEB1("Db::clearDocTermIfWdf0: [" << (term) << "]\n" );
// Find the term
Xapian::TermIterator xit;
XAPTRY(xit = xdoc.termlist_begin(); xit.skip_to(term);,
xrdb, m_rcldb->m_reason);
if (!m_rcldb->m_reason.empty()) {
LOGERR("Db::clearDocTerm...: [" << (term) << "] skip failed: " << (m_rcldb->m_reason) << "\n" );
return false;
}
if (xit == xdoc.termlist_end() || term.compare(*xit)) {
LOGDEB0("Db::clearDocTermIFWdf0: term [" << (term) << "] not found. xit: [" << (xit == xdoc.termlist_end() ? "EOL":(*xit)) << "]\n" );
return false;
}
// Clear the term if its frequency is 0
if (xit.get_wdf() == 0) {
LOGDEB1("Db::clearDocTermIfWdf0: clearing [" << (term) << "]\n" );
XAPTRY(xdoc.remove_term(term), xwdb, m_rcldb->m_reason);
if (!m_rcldb->m_reason.empty()) {
LOGDEB0("Db::clearDocTermIfWdf0: failed [" << (term) << "]: " << (m_rcldb->m_reason) << "\n" );
}
}
return true;
}
// Holder for term + pos
struct DocPosting {
DocPosting(string t, Xapian::termpos ps)
: term(t), pos(ps) {}
string term;
Xapian::termpos pos;
};
// Clear all terms for given field for given document.
// The terms to be cleared are all those with the appropriate
// prefix. We also remove the postings for the unprefixed terms (that
// is, we undo what we did when indexing).
bool Db::Native::clearField(Xapian::Document& xdoc, const string& pfx,
Xapian::termcount wdfdec)
{
LOGDEB1("Db::clearField: clearing prefix [" << (pfx) << "] for docid " << (unsigned(xdoc.get_docid())) << "\n" );
vector<DocPosting> eraselist;
string wrapd = wrap_prefix(pfx);
m_rcldb->m_reason.clear();
for (int tries = 0; tries < 2; tries++) {
try {
Xapian::TermIterator xit;
xit = xdoc.termlist_begin();
xit.skip_to(wrapd);
while (xit != xdoc.termlist_end() &&
!(*xit).compare(0, wrapd.size(), wrapd)) {
LOGDEB1("Db::clearfield: erasing for [" << ((*xit)) << "]\n" );
Xapian::PositionIterator posit;
for (posit = xit.positionlist_begin();
posit != xit.positionlist_end(); posit++) {
eraselist.push_back(DocPosting(*xit, *posit));
eraselist.push_back(DocPosting(strip_prefix(*xit), *posit));
}
xit++;
}
} catch (const Xapian::DatabaseModifiedError &e) {
m_rcldb->m_reason = e.get_msg();
xrdb.reopen();
continue;
} XCATCHERROR(m_rcldb->m_reason);
break;
}
if (!m_rcldb->m_reason.empty()) {
LOGERR("Db::clearField: failed building erase list: " << (m_rcldb->m_reason) << "\n" );
return false;
}
// Now remove the found positions, and the terms if the wdf is 0
for (vector<DocPosting>::const_iterator it = eraselist.begin();
it != eraselist.end(); it++) {
LOGDEB1("Db::clearField: remove posting: [" << (it->term) << "] pos [" << (int(it->pos)) << "]\n" );
XAPTRY(xdoc.remove_posting(it->term, it->pos, wdfdec);,
xwdb,m_rcldb->m_reason);
if (!m_rcldb->m_reason.empty()) {
// Not that this normally fails for non-prefixed XXST and
// ND, don't make a fuss
LOGDEB1("Db::clearFiedl: remove_posting failed for [" << (it->term) << "]," << (int(it->pos)) << ": " << (m_rcldb->m_reason) << "\n" );
}
clearDocTermIfWdf0(xdoc, it->term);
}
return true;
}
// Check if doc given by udi is indexed by term
bool Db::Native::hasTerm(const string& udi, int idxi, const string& term)
{
LOGDEB2("Native::hasTerm: udi [" << (udi) << "] term [" << (term) << "]\n" );
Xapian::Document xdoc;
if (getDoc(udi, idxi, xdoc)) {
Xapian::TermIterator xit;
XAPTRY(xit = xdoc.termlist_begin();
xit.skip_to(term);,
xrdb, m_rcldb->m_reason);
if (!m_rcldb->m_reason.empty()) {
LOGERR("Rcl::Native::hasTerm: " << (m_rcldb->m_reason) << "\n" );
return false;
}
if (xit != xdoc.termlist_end() && !term.compare(*xit)) {
return true;
}
}
return false;
}
// Retrieve Xapian document, given udi. There may be several identical udis
// if we are using multiple indexes.
Xapian::docid Db::Native::getDoc(const string& udi, int idxi,
Xapian::Document& xdoc)
{
string uniterm = make_uniterm(udi);
for (int tries = 0; tries < 2; tries++) {
try {
Xapian::PostingIterator docid;
for (docid = xrdb.postlist_begin(uniterm);
docid != xrdb.postlist_end(uniterm); docid++) {
xdoc = xrdb.get_document(*docid);
if (whatDbIdx(*docid) == (size_t)idxi)
return *docid;
}
// Udi not in Db.
return 0;
} catch (const Xapian::DatabaseModifiedError &e) {
m_rcldb->m_reason = e.get_msg();
xrdb.reopen();
continue;
} XCATCHERROR(m_rcldb->m_reason);
break;
}
LOGERR("Db::Native::getDoc: Xapian error: " << (m_rcldb->m_reason) << "\n" );
return 0;
}
// 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" << (data) << "\n" );
ConfSimple parms(data);
if (!parms.ok())
return false;
doc.xdocid = docid;
doc.haspages = hasPages(docid);
// Compute what index this comes from, and check for path translations
string dbdir = m_rcldb->m_basedir;
doc.idxi = 0;
if (!m_rcldb->m_extraDbs.empty()) {
int idxi = int(whatDbIdx(docid));
// idxi is in [0, extraDbs.size()]. 0 is for the main index,
// idxi-1 indexes into the additional dbs array.
if (idxi) {
dbdir = m_rcldb->m_extraDbs[idxi - 1];
doc.idxi = idxi;
}
}
parms.get(Doc::keyurl, doc.idxurl);
doc.url = doc.idxurl;
m_rcldb->m_config->urlrewrite(dbdir, doc.url);
if (!doc.url.compare(doc.idxurl))
doc.idxurl.clear();
// Special cases:
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(cstr_caption, doc.meta[Doc::keytt]);
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(cstr_syntAbs) == 0) {
doc.meta[Doc::keyabs] =
doc.meta[Doc::keyabs].substr(cstr_syntAbs.length());
doc.syntabs = true;
}
parms.get(Doc::keyipt, doc.ipath);
parms.get(Doc::keypcs, doc.pcbytes);
parms.get(Doc::keyfs, doc.fbytes);
parms.get(Doc::keyds, doc.dbytes);
parms.get(Doc::keysig, doc.sig);
// Normal key/value pairs:
vector<string> keys = parms.getNames(string());
for (vector<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::keyurl] = doc.url;
doc.meta[Doc::keymt] = doc.dmtime.empty() ? doc.fmtime : doc.dmtime;
return true;
}
bool Db::Native::hasPages(Xapian::docid docid)
{
string ermsg;
Xapian::PositionIterator pos;
XAPTRY(pos = xrdb.positionlist_begin(docid, page_break_term);
if (pos != xrdb.positionlist_end(docid, page_break_term)) {
return true;
},
xrdb, ermsg);
if (!ermsg.empty()) {
LOGERR("Db::Native::hasPages: xapian error: " << (ermsg) << "\n" );
}
return false;
}
// Return the positions list for the page break term
bool Db::Native::getPagePositions(Xapian::docid docid, vector<int>& vpos)
{
vpos.clear();
// Need to retrieve the document record to check for multiple page breaks
// that we store there for lack of better place
map<int, int> mbreaksmap;
try {
Xapian::Document xdoc = xrdb.get_document(docid);
string data = xdoc.get_data();
Doc doc;
string mbreaks;
if (dbDataToRclDoc(docid, data, doc) &&
doc.getmeta(cstr_mbreaks, &mbreaks)) {
vector<string> values;
stringToTokens(mbreaks, values, ",");
for (unsigned int i = 0; i < values.size() - 1; i += 2) {
int pos = atoi(values[i].c_str()) + baseTextPosition;
int incr = atoi(values[i+1].c_str());
mbreaksmap[pos] = incr;
}
}
} catch (...) {
}
string qterm = page_break_term;
Xapian::PositionIterator pos;
try {
for (pos = xrdb.positionlist_begin(docid, qterm);
pos != xrdb.positionlist_end(docid, qterm); pos++) {
int ipos = *pos;
if (ipos < int(baseTextPosition)) {
LOGDEB("getPagePositions: got page position " << (ipos) << " not in body\n" );
// Not in text body. Strange...
continue;
}
map<int, int>::iterator it = mbreaksmap.find(ipos);
if (it != mbreaksmap.end()) {
LOGDEB1("getPagePositions: found multibreak at " << (ipos) << " incr " << (it->second) << "\n" );
for (int i = 0 ; i < it->second; i++)
vpos.push_back(ipos);
}
vpos.push_back(ipos);
}
} catch (...) {
// Term does not occur. No problem.
}
return true;
}
int Db::Native::getPageNumberForPosition(const vector<int>& pbreaks, int pos)
{
if (pos < int(baseTextPosition)) // Not in text body
return -1;
vector<int>::const_iterator it =
upper_bound(pbreaks.begin(), pbreaks.end(), pos);
return int(it - pbreaks.begin() + 1);
}
// Note: we're passed a Xapian::Document* because Xapian
// reference-counting is not mt-safe. We take ownership and need
// to delete it before returning.
bool Db::Native::addOrUpdateWrite(const string& udi, const string& uniterm,
Xapian::Document *newdocument_ptr,
size_t textlen)
{
#ifdef IDX_THREADS
Chrono chron;
std::unique_lock<std::mutex> lock(m_mutex);
#endif
std::shared_ptr<Xapian::Document> doc_cleaner(newdocument_ptr);
// Check file system full every mbyte of indexed text. It's a bit wasteful
// to do this after having prepared the document, but it needs to be in
// the single-threaded section.
if (m_rcldb->m_maxFsOccupPc > 0 &&
(m_rcldb->m_occFirstCheck ||
(m_rcldb->m_curtxtsz - m_rcldb->m_occtxtsz) / MB >= 1)) {
LOGDEB("Db::add: checking file system usage\n" );
int pc;
m_rcldb->m_occFirstCheck = 0;
if (fsocc(m_rcldb->m_basedir, &pc) && pc >= m_rcldb->m_maxFsOccupPc) {
LOGERR("Db::add: stop indexing: file system " << pc << " %" <<
" full > max " << m_rcldb->m_maxFsOccupPc << " %" << "\n");
return false;
}
m_rcldb->m_occtxtsz = m_rcldb->m_curtxtsz;
}
const char *fnc = udi.c_str();
string ermsg;
// Add db entry or update existing entry:
try {
Xapian::docid did =
xwdb.replace_document(uniterm, *newdocument_ptr);
if (did < m_rcldb->updated.size()) {
// This is necessary because only the file-level docs are tested
// by needUpdate(), so the subdocs existence flags are only set
// here.
m_rcldb->updated[did] = true;
LOGINFO("Db::add: docid " << (did) << " updated [" << (fnc) << "]\n" );
} else {
LOGINFO("Db::add: docid " << (did) << " added [" << (fnc) << "]\n" );
}
} XCATCHERROR(ermsg);
if (!ermsg.empty()) {
LOGERR("Db::add: replace_document failed: " << (ermsg) << "\n" );
ermsg.erase();
// FIXME: is this ever actually needed?
try {
xwdb.add_document(*newdocument_ptr);
LOGDEB("Db::add: " << (fnc) << " added (failed re-seek for duplicate)\n" );
} XCATCHERROR(ermsg);
if (!ermsg.empty()) {
LOGERR("Db::add: add_document failed: " << (ermsg) << "\n" );
return false;
}
}
// Test if we're over the flush threshold (limit memory usage):
bool ret = m_rcldb->maybeflush(textlen);
#ifdef IDX_THREADS
m_totalworkns += chron.nanos();
#endif
return ret;
}
bool Db::Native::purgeFileWrite(bool orphansOnly, const string& udi,
const string& uniterm)
{
#if defined(IDX_THREADS)
// We need a mutex even if we have a write queue (so we can only
// be called by a single thread) to protect about multiple acces
// to xrdb from subDocs() which is also called from needupdate()
// (called from outside the write thread !
std::unique_lock<std::mutex> lock(m_mutex);
#endif // IDX_THREADS
string ermsg;
try {
Xapian::PostingIterator docid = xwdb.postlist_begin(uniterm);
if (docid == xwdb.postlist_end(uniterm)) {
return true;
}
if (m_rcldb->m_flushMb > 0) {
Xapian::termcount trms = xwdb.get_doclength(*docid);
m_rcldb->maybeflush(trms * 5);
}
string sig;
if (orphansOnly) {
Xapian::Document doc = xwdb.get_document(*docid);
sig = doc.get_value(VALUE_SIG);
if (sig.empty()) {
LOGINFO("purgeFileWrite: got empty sig\n" );
return false;
}
} else {
LOGDEB("purgeFile: delete docid " << (*docid) << "\n" );
xwdb.delete_document(*docid);
}
vector<Xapian::docid> docids;
subDocs(udi, 0, docids);
LOGDEB("purgeFile: subdocs cnt " << (docids.size()) << "\n" );
for (vector<Xapian::docid>::iterator it = docids.begin();
it != docids.end(); it++) {
if (m_rcldb->m_flushMb > 0) {
Xapian::termcount trms = xwdb.get_doclength(*it);
m_rcldb->maybeflush(trms * 5);
}
string subdocsig;
if (orphansOnly) {
Xapian::Document doc = xwdb.get_document(*it);
subdocsig = doc.get_value(VALUE_SIG);
if (subdocsig.empty()) {
LOGINFO("purgeFileWrite: got empty sig for subdoc??\n" );
continue;
}
}
if (!orphansOnly || sig != subdocsig) {
LOGDEB("Db::purgeFile: delete subdoc " << (*it) << "\n" );
xwdb.delete_document(*it);
}
}
return true;
} XCATCHERROR(ermsg);
if (!ermsg.empty()) {
LOGERR("Db::purgeFileWrite: " << (ermsg) << "\n" );
}
return false;
}
/* Rcl::Db methods ///////////////////////////////// */
bool Db::o_inPlaceReset;
Db::Db(const RclConfig *cfp)
: m_ndb(0), m_mode(Db::DbRO), m_curtxtsz(0), m_flushtxtsz(0),
m_occtxtsz(0), m_occFirstCheck(1), m_idxMetaStoredLen(150),
m_idxAbsTruncLen(250), m_synthAbsLen(250), m_synthAbsWordCtxLen(4),
m_flushMb(-1), m_maxFsOccupPc(0)
{
m_config = new RclConfig(*cfp);
if (start_of_field_term.empty()) {
if (o_index_stripchars) {
start_of_field_term = "XXST";
end_of_field_term = "XXND";
} else {
start_of_field_term = "XXST/";
end_of_field_term = "XXND/";
}
}
m_ndb = new Native(this);
if (m_config) {
m_config->getConfParam("maxfsoccuppc", &m_maxFsOccupPc);
m_config->getConfParam("idxflushmb", &m_flushMb);
m_config->getConfParam("idxmetastoredlen", &m_idxMetaStoredLen);
}
}
Db::~Db()
{
LOGDEB2("Db::~Db\n" );
if (m_ndb == 0)
return;
LOGDEB("Db::~Db: isopen " << (m_ndb->m_isopen) << " m_iswritable " << (m_ndb->m_iswritable) << "\n" );
i_close(true);
delete m_config;
}
vector<string> Db::getStemmerNames()
{
vector<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 " << (m_ndb->m_isopen) << " m_iswritable " << (m_ndb->m_iswritable) << " mode " << (mode) << "\n" );
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(cstr_RCL_IDX_VERSION_KEY,
cstr_RCL_IDX_VERSION);
m_ndb->m_iswritable = true;
#ifdef IDX_THREADS
m_ndb->maybeStartThreads();
#endif
// We used to open a readonly object in addition to
// the r/w one because some operations were faster
// when performed through a Database: no forced
// flushes on allterms_begin(), used in
// subDocs(). This issue has been gone for a long time
// (now: Xapian 1.2) and the separate objects seem to
// trigger other Xapian issues, so the query db is now
// a clone of the update one.
m_ndb->xrdb = m_ndb->xwdb;
LOGDEB("Db::open: lastdocid: " << (m_ndb->xwdb.get_lastdocid()) << "\n" );
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 (vector<string>::iterator it = m_extraDbs.begin();
it != m_extraDbs.end(); it++) {
if (error)
*error = DbOpenExtraDb;
LOGDEB("Db::Open: adding query db [" << &(*it) << "]\n" );
// 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->xrdb.get_doccount() > 0) {
string version = m_ndb->xrdb.get_metadata(cstr_RCL_IDX_VERSION_KEY);
if (version.compare(cstr_RCL_IDX_VERSION)) {
m_ndb->m_noversionwrite = true;
LOGERR("Rcl::Db::open: file index [" << (version) << "], software [" << (cstr_RCL_IDX_VERSION) << "]\n" );
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 [" << (dir) << "]: " << (ermsg) << "\n" );
return false;
}
// Note: xapian has no close call, we delete and recreate the db
bool Db::close()
{
LOGDEB1("Db::close()\n" );
return i_close(false);
}
bool Db::i_close(bool final)
{
if (m_ndb == 0)
return false;
LOGDEB("Db::i_close(" << (final) << "): m_isopen " << (m_ndb->m_isopen) << " m_iswritable " << (m_ndb->m_iswritable) << "\n" );
if (m_ndb->m_isopen == false && !final)
return true;
string ermsg;
try {
bool w = m_ndb->m_iswritable;
if (w) {
#ifdef IDX_THREADS
waitUpdIdle();
#endif
if (!m_ndb->m_noversionwrite)
m_ndb->xwdb.set_metadata(cstr_RCL_IDX_VERSION_KEY,
cstr_RCL_IDX_VERSION);
LOGDEB("Rcl::Db:close: xapian will close. May take some time\n" );
}
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: " << (ermsg) << "\n" );
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->xrdb.get_doccount(), m_ndb->xrdb, m_reason);
if (!m_reason.empty()) {
LOGERR("Db::docCnt: got error: " << (m_reason) << "\n" );
return -1;
}
return res;
}
int Db::termDocCnt(const string& _term)
{
int res = -1;
if (!m_ndb || !m_ndb->m_isopen)
return -1;
string term = _term;
if (o_index_stripchars)
if (!unacmaybefold(_term, term, "UTF-8", UNACOP_UNACFOLD)) {
LOGINFO("Db::termDocCnt: unac failed for [" << (_term) << "]\n" );
return 0;
}
if (m_stops.isStop(term)) {
LOGDEB1("Db::termDocCnt [" << (term) << "] in stop list\n" );
return 0;
}
XAPTRY(res = m_ndb->xrdb.get_termfreq(term), m_ndb->xrdb, m_reason);
if (!m_reason.empty()) {
LOGERR("Db::termDocCnt: got error: " << (m_reason) << "\n" );
return -1;
}
return res;
}
bool Db::addQueryDb(const string &_dir)
{
string dir = _dir;
LOGDEB0("Db::addQueryDb: ndb " << (m_ndb) << " iswritable " << ((m_ndb)?m_ndb->m_iswritable:0) << " db [" << (dir) << "]\n" );
if (!m_ndb)
return false;
if (m_ndb->m_iswritable)
return false;
dir = path_canon(dir);
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 {
vector<string>::iterator it = find(m_extraDbs.begin(),
m_extraDbs.end(), dir);
if (it != m_extraDbs.end()) {
m_extraDbs.erase(it);
}
}
return adjustdbs();
}
// Determining what index a doc result comes from is based on the
// modulo of the docid against the db count. Ref:
// http://trac.xapian.org/wiki/FAQ/MultiDatabaseDocumentID
size_t Db::whatDbIdx(const Doc& doc)
{
return m_ndb->whatDbIdx(doc.xdocid);
}
size_t Db::Native::whatDbIdx(Xapian::docid id)
{
LOGDEB1("Db::whatDbIdx: xdocid " << ((unsigned long)id) << ", " <<
(m_rcldb->m_extraDbs.size()) << " extraDbs\n" );
if (id == 0)
return (size_t)-1;
if (m_rcldb->m_extraDbs.size() == 0)
return 0;
return (id - 1) % (m_rcldb->m_extraDbs.size() + 1);
}
bool Db::testDbDir(const string &dir, bool *stripped_p)
{
string aerr;
bool mstripped = true;
LOGDEB("Db::testDbDir: [" << (dir) << "]\n" );
try {
Xapian::Database db(dir);
// If we have terms with a leading ':' it's an
// unstripped index
Xapian::TermIterator term = db.allterms_begin(":");
if (term == db.allterms_end())
mstripped = true;
else
mstripped = false;
} XCATCHERROR(aerr);
if (!aerr.empty()) {
LOGERR("Db::Open: error while trying to open database from [" << (dir) << "]: " << (aerr) << "\n" );
return false;
}
if (stripped_p)
*stripped_p = mstripped;
return true;
}
bool Db::isopen()
{
if (m_ndb == 0)
return false;
return m_ndb->m_isopen;
}
// Try to translate field specification into field prefix.
bool Db::fieldToTraits(const string& fld, const FieldTraits **ftpp,
bool isquery)
{
if (m_config && m_config->getFieldTraits(fld, ftpp, isquery))
return true;
*ftpp = 0;
return false;
}
// The splitter breaks text into words and adds postings to the Xapian
// document. We use a single object to split all of the document
// fields and position jumps to separate fields
class TextSplitDb : public TextSplitP {
public:
Xapian::Document &doc; // Xapian document
// Base for document section. Gets large increment when we change
// sections, to avoid cross-section proximity matches.
Xapian::termpos basepos;
// Current relative position. This is the remembered value from
// the splitter callback. The term position is reset for each call
// to text_to_words(), so that the last value of curpos is the
// section size (last relative term position), and this is what
// gets added to basepos in addition to the inter-section increment
// to compute the first position of the next section.
Xapian::termpos curpos;
TextSplitDb(Xapian::Document &d, TermProc *prc)
: TextSplitP(prc), doc(d), basepos(1), curpos(0)
{}
// Reimplement text_to_words to insert the begin and end anchor terms.
virtual bool text_to_words(const string &in)
{
string ermsg;
try {
// Index the possibly prefixed start term.
doc.add_posting(ft.pfx + start_of_field_term, basepos, ft.wdfinc);
++basepos;
} XCATCHERROR(ermsg);
if (!ermsg.empty()) {
LOGERR("Db: xapian add_posting error " << (ermsg) << "\n" );
goto out;
}
if (!TextSplitP::text_to_words(in)) {
LOGDEB("TextSplitDb: TextSplit::text_to_words failed\n" );
goto out;
}
try {
// Index the possibly prefixed end term.
doc.add_posting(ft.pfx + end_of_field_term, basepos + curpos + 1,
ft.wdfinc);
++basepos;
} XCATCHERROR(ermsg);
if (!ermsg.empty()) {
LOGERR("Db: xapian add_posting error " << (ermsg) << "\n" );
goto out;
}
out:
basepos += curpos + 100;
return true;
}
void setTraits(const FieldTraits& ftp)
{
ft = ftp;
if (!ft.pfx.empty())
ft.pfx = wrap_prefix(ft.pfx);
}
friend class TermProcIdx;
private:
FieldTraits ft;
};
class TermProcIdx : public TermProc {
public:
TermProcIdx() : TermProc(0), m_ts(0), m_lastpagepos(0), m_pageincr(0) {}
void setTSD(TextSplitDb *ts) {m_ts = ts;}
bool takeword(const std::string &term, int pos, int, int)
{
// Compute absolute position (pos is relative to current segment),
// and remember relative.
m_ts->curpos = pos;
pos += m_ts->basepos;
// Don't try to add empty term Xapian doesnt like it... Safety check
// this should not happen.
if (term.empty())
return true;
string ermsg;
try {
// Index without prefix, using the field-specific weighting
LOGDEB1("Emitting term at " << pos << " : [" << term << "]\n" );
if (!m_ts->ft.pfxonly)
m_ts->doc.add_posting(term, pos, m_ts->ft.wdfinc);
#ifdef TESTING_XAPIAN_SPELL
if (Db::isSpellingCandidate(term)) {
m_ts->db.add_spelling(term);
}
#endif
// Index the prefixed term.
if (!m_ts->ft.pfx.empty()) {
m_ts->doc.add_posting(m_ts->ft.pfx + term, pos,
m_ts->ft.wdfinc);
}
return true;
} XCATCHERROR(ermsg);
LOGERR("Db: xapian add_posting error " << (ermsg) << "\n" );
return false;
}
void newpage(int pos)
{
pos += m_ts->basepos;
if (pos < int(baseTextPosition)) {
LOGDEB("newpage: not in body: " << (pos) << "\n" );
return;
}
m_ts->doc.add_posting(m_ts->ft.pfx + page_break_term, pos);
if (pos == m_lastpagepos) {
m_pageincr++;
LOGDEB2("newpage: same pos, pageincr " << (m_pageincr) << " lastpagepos " << (m_lastpagepos) << "\n" );
} else {
LOGDEB2("newpage: pos change, pageincr " << (m_pageincr) << " lastpagepos " << (m_lastpagepos) << "\n" );
if (m_pageincr > 0) {
// Remember the multiple page break at this position
unsigned int relpos = m_lastpagepos - baseTextPosition;
LOGDEB2("Remembering multiple page break. Relpos " << (relpos) << " cnt " << (m_pageincr) << "\n" );
m_pageincrvec.push_back(pair<int, int>(relpos, m_pageincr));
}
m_pageincr = 0;
}
m_lastpagepos = pos;
}
virtual bool flush()
{
if (m_pageincr > 0) {
unsigned int relpos = m_lastpagepos - baseTextPosition;
LOGDEB2("Remembering multiple page break. Position " << (relpos) << " cnt " << (m_pageincr) << "\n" );
m_pageincrvec.push_back(pair<int, int>(relpos, m_pageincr));
m_pageincr = 0;
}
return TermProc::flush();
}
TextSplitDb *m_ts;
// Auxiliary page breaks data for positions with multiple page breaks.
int m_lastpagepos;
// increment of page breaks at same pos. Normally 0, 1.. when several
// breaks at the same pos
int m_pageincr;
vector <pair<int, int> > m_pageincrvec;
};
#ifdef TESTING_XAPIAN_SPELL
string Db::getSpellingSuggestion(const string& word)
{
if (m_ndb == 0)
return string();
string term = word;
if (o_index_stripchars)
if (!unacmaybefold(word, term, "UTF-8", UNACOP_UNACFOLD)) {
LOGINFO("Db::getSpelling: unac failed for [" << (word) << "]\n" );
return string();
}
if (!isSpellingCandidate(term))
return string();
return m_ndb->xrdb.get_spelling_suggestion(term);
}
#endif
// Let our user set the parameters for abstract processing
void Db::setAbstractParams(int idxtrunc, int syntlen, int syntctxlen)
{
LOGDEB1("Db::setAbstractParams: trunc " << (idxtrunc) << " syntlen " << (syntlen) << " ctxlen " << (syntctxlen) << "\n" );
if (idxtrunc >= 0)
m_idxAbsTruncLen = idxtrunc;
if (syntlen > 0)
m_synthAbsLen = syntlen;
if (syntctxlen > 0)
m_synthAbsWordCtxLen = syntctxlen;
}
bool Db::setSynGroupsFile(const string& fn)
{
return m_syngroups.setfile(fn);
}
static const string cstr_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 etc. , create the document data record (more
// metadata), and update database
bool Db::addOrUpdate(const string &udi, const string &parent_udi, Doc &doc)
{
LOGDEB("Db::add: udi [" << (udi) << "] parent [" << (parent_udi) << "]\n" );
if (m_ndb == 0)
return false;
// This document is potentially going to be passed to the index
// update thread. The reference counters are not mt-safe, so we
// need to do this through a pointer. The reference is just there
// to avoid changing too much code (the previous version passed a copy).
Xapian::Document *newdocument_ptr = new Xapian::Document;
Xapian::Document &newdocument(*newdocument_ptr);
// The term processing pipeline:
TermProcIdx tpidx;
TermProc *nxt = &tpidx;
TermProcStop tpstop(nxt, m_stops);nxt = &tpstop;
//TermProcCommongrams tpcommon(nxt, m_stops); nxt = &tpcommon;
TermProcPrep tpprep(nxt);
if (o_index_stripchars)
nxt = &tpprep;
TextSplitDb splitter(newdocument, nxt);
tpidx.setTSD(&splitter);
// Udi unique term: this is used for file existence/uptodate
// checks, and unique id for the replace_document() call.
string uniterm = make_uniterm(udi);
if (doc.onlyxattr) {
// Only updating an existing doc with new extended attributes
// data. Need to read the old doc and its data record
// first. This is so different from the normal processing that
// it uses a fully separate code path (with some duplication
// unfortunately)
if (!m_ndb->docToXdocXattrOnly(&splitter, udi, doc, newdocument)) {
delete newdocument_ptr;
return false;
}
} else {
// 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);
}
}
// Split and index the path from the url for path-based filtering
{
string path = url_gpathS(doc.url);
#ifdef _WIN32
// Windows file names are case-insensitive, so we
// translate to UTF-8 and lowercase
string upath = compute_utf8fn(m_config, path, false);
unacmaybefold(upath, path, "UTF-8", UNACOP_FOLD);
#endif
vector<string> vpath;
stringToTokens(path, vpath, "/");
// If vpath is not /, the last elt is the file/dir name, not a
// part of the path.
if (vpath.size())
vpath.resize(vpath.size()-1);
splitter.curpos = 0;
newdocument.add_posting(wrap_prefix(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(wrap_prefix(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;
for (meta_it = doc.meta.begin(); meta_it != doc.meta.end(); meta_it++) {
if (!meta_it->second.empty()) {
const FieldTraits *ftp;
// We don't test for an empty prefix here. Some fields are part
// of the internal conf with an empty prefix (ie: abstract).
if (!fieldToTraits(meta_it->first, &ftp)) {
LOGDEB0("Db::add: no prefix for field [" << (meta_it->first) << "], no indexing\n" );
continue;
}
LOGDEB0("Db::add: field [" << (meta_it->first) << "] pfx [" << (ftp->pfx) << "] inc " << (ftp->wdfinc) << ": [" << (meta_it->second) << "]\n" );
splitter.setTraits(*ftp);
if (!splitter.text_to_words(meta_it->second))
LOGDEB("Db::addOrUpdate: split failed for " << (meta_it->first) << "\n" );
}
}
// Reset to no prefix and default params
splitter.setTraits(FieldTraits());
if (splitter.curpos < baseTextPosition)
splitter.basepos = baseTextPosition;
// Split and index body text
LOGDEB2("Db::add: split body: [" << (doc.text) << "]\n" );
#ifdef TEXTSPLIT_STATS
splitter.resetStats();
#endif
if (!splitter.text_to_words(doc.text))
LOGDEB("Db::addOrUpdate: split failed for main text\n" );
#ifdef TEXTSPLIT_STATS
// Reject bad data. unrecognized base64 text is characterized by
// high avg word length and high variation (because there are
// word-splitters like +/ inside the data).
TextSplit::Stats::Values v = splitter.getStats();
// v.avglen > 15 && v.sigma > 12
if (v.count > 200 && (v.avglen > 10 && v.sigma / v.avglen > 0.8)) {
LOGINFO("RclDb::addOrUpdate: rejecting doc for bad stats count " << (v.count) << " avglen " << (v.avglen) << " sigma " << (v.sigma) << " url [" << (doc.url) << "] ipath [" << (doc.ipath) << "] text " << (doc.text) << "\n" );
delete newdocument_ptr;
return true;
}
#endif
////// Special terms for other metadata. No positions for these.
// Mime type
newdocument.add_boolean_term(wrap_prefix(mimetype_prefix) + doc.mimetype);
// Simple file name indexed unsplit for specific "file name"
// searches. This is not the same as a filename: clause inside the
// query language.
// We also add a term for the filename extension if any.
string utf8fn;
if (doc.getmeta(Doc::keyfn, &utf8fn) && !utf8fn.empty()) {
string fn;
if (unacmaybefold(utf8fn, fn, "UTF-8", UNACOP_UNACFOLD)) {
// 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_boolean_term(wrap_prefix(fileext_prefix) +
fn.substr(pos + 1));
}
newdocument.add_term(wrap_prefix(unsplitfilename_prefix) + fn, 0);
}
}
newdocument.add_boolean_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_boolean_term(make_parentterm(parent_udi));
}
// Dates etc.
time_t mtime = atoll(doc.dmtime.empty() ? doc.fmtime.c_str() :
doc.dmtime.c_str());
struct tm tmb;
localtime_r(&mtime, &tmb);
char buf[9];
snprintf(buf, 9, "%04d%02d%02d",
tmb.tm_year+1900, tmb.tm_mon + 1, tmb.tm_mday);
// Date (YYYYMMDD)
newdocument.add_boolean_term(wrap_prefix(xapday_prefix) + string(buf));
// Month (YYYYMM)
buf[6] = '\0';
newdocument.add_boolean_term(wrap_prefix(xapmonth_prefix) + string(buf));
// Year (YYYY)
buf[4] = '\0';
newdocument.add_boolean_term(wrap_prefix(xapyear_prefix) + string(buf));
//////////////////////////////////////////////////////////////////
// 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())
doc.fbytes = doc.pcbytes;
if (!doc.fbytes.empty()) {
RECORD_APPEND(record, Doc::keyfs, doc.fbytes);
leftzeropad(doc.fbytes, 12);
newdocument.add_value(VALUE_SIZE, doc.fbytes);
}
if (doc.haschildren) {
newdocument.add_boolean_term(has_children_term);
}
if (!doc.pcbytes.empty())
RECORD_APPEND(record, Doc::keypcs, doc.pcbytes);
char sizebuf[30];
sprintf(sizebuf, "%u", (unsigned int)doc.text.length());
RECORD_APPEND(record, Doc::keyds, sizebuf);
// 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);
}
if (!doc.ipath.empty())
RECORD_APPEND(record, Doc::keyipt, doc.ipath);
// Fields from the Meta array. Handle title specially because it has a
// different name inside the data record (history...)
string& ttref = doc.meta[Doc::keytt];
ttref = neutchars(truncate_to_word(ttref, m_idxMetaStoredLen), cstr_nc);
if (!ttref.empty()) {
RECORD_APPEND(record, cstr_caption, ttref);
ttref.clear();
}
// 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.
// Note that the map accesses by operator[] create empty entries if they
// don't exist yet.
if (m_idxAbsTruncLen > 0) {
string& absref = doc.meta[Doc::keyabs];
trimstring(absref, " \t\r\n");
if (absref.empty()) {
if (!doc.text.empty())
absref = cstr_syntAbs +
neutchars(truncate_to_word(doc.text, m_idxAbsTruncLen),
cstr_nc);
} else {
absref = neutchars(truncate_to_word(absref, m_idxAbsTruncLen),
cstr_nc);
}
// Do the append here to avoid the different truncation done
// in the regular "stored" loop
if (!absref.empty()) {
RECORD_APPEND(record, Doc::keyabs, absref);
absref.clear();
}
}
// Append all regular "stored" meta fields
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[nm].empty()) {
string value =
neutchars(truncate_to_word(doc.meta[nm],
m_idxMetaStoredLen), cstr_nc);
RECORD_APPEND(record, nm, value);
}
}
// At this point, if the document "filename" field was empty,
// try to store the "container file name" value. This is done
// after indexing because we don't want search matches on
// this, but the filename is often useful for display
// purposes.
const string *fnp = 0;
if (!doc.peekmeta(Rcl::Doc::keyfn, &fnp) || fnp->empty()) {
if (doc.peekmeta(Rcl::Doc::keytcfn, &fnp) && !fnp->empty()) {
string value =
neutchars(truncate_to_word(*fnp,
m_idxMetaStoredLen), cstr_nc);
RECORD_APPEND(record, Rcl::Doc::keyfn, value);
}
}
// If empty pages (multiple break at same pos) were recorded, save
// them (this is because we have no way to record them in the
// Xapian list
if (!tpidx.m_pageincrvec.empty()) {
ostringstream multibreaks;
for (unsigned int i = 0; i < tpidx.m_pageincrvec.size(); i++) {
if (i != 0)
multibreaks << ",";
multibreaks << tpidx.m_pageincrvec[i].first << "," <<
tpidx.m_pageincrvec[i].second;
}
RECORD_APPEND(record, string(cstr_mbreaks), multibreaks.str());
}
// If the file's md5 was computed, add value and term.
// The value is optionally used for query result duplicate elimination,
// and the term to find the duplicates.
// We don't do this for empty docs.
const string *md5;
if (doc.peekmeta(Doc::keymd5, &md5) && !md5->empty() &&
md5->compare(cstr_md5empty)) {
string digest;
MD5HexScan(*md5, digest);
newdocument.add_value(VALUE_MD5, digest);
newdocument.add_boolean_term(wrap_prefix("XM") + *md5);
}
LOGDEB0("Rcl::Db::add: new doc record:\n" << (record) << "\n" );
newdocument.set_data(record);
}
#ifdef IDX_THREADS
if (m_ndb->m_havewriteq) {
DbUpdTask *tp = new DbUpdTask(DbUpdTask::AddOrUpdate, udi, uniterm,
newdocument_ptr, doc.text.length());
if (!m_ndb->m_wqueue.put(tp)) {
LOGERR("Db::addOrUpdate:Cant queue task\n" );
delete newdocument_ptr;
return false;
} else {
return true;
}
}
#endif
return m_ndb->addOrUpdateWrite(udi, uniterm, newdocument_ptr,
doc.text.length());
}
bool Db::Native::docToXdocXattrOnly(TextSplitDb *splitter, const string &udi,
Doc &doc, Xapian::Document& xdoc)
{
LOGDEB0("Db::docToXdocXattrOnly\n" );
#ifdef IDX_THREADS
std::unique_lock<std::mutex> lock(m_mutex);
#endif
// Read existing document and its data record
if (getDoc(udi, 0, xdoc) == 0) {
LOGERR("docToXdocXattrOnly: existing doc not found\n" );
return false;
}
string data;
XAPTRY(data = xdoc.get_data(), xrdb, m_rcldb->m_reason);
if (!m_rcldb->m_reason.empty()) {
LOGERR("Db::xattrOnly: got error: " << (m_rcldb->m_reason) << "\n" );
return false;
}
// Clear the term lists for the incoming fields and index the new values
map<string, string>::iterator meta_it;
for (meta_it = doc.meta.begin(); meta_it != doc.meta.end(); meta_it++) {
const FieldTraits *ftp;
if (!m_rcldb->fieldToTraits(meta_it->first, &ftp) || ftp->pfx.empty()) {
LOGDEB0("Db::xattrOnly: no prefix for field [" << (meta_it->first) << "], skipped\n" );
continue;
}
// Clear the previous terms for the field
clearField(xdoc, ftp->pfx, ftp->wdfinc);
LOGDEB0("Db::xattrOnly: field [" << (meta_it->first) << "] pfx [" << (ftp->pfx) << "] inc " << (ftp->wdfinc) << ": [" << (meta_it->second) << "]\n" );
splitter->setTraits(*ftp);
if (!splitter->text_to_words(meta_it->second))
LOGDEB("Db::xattrOnly: split failed for " << (meta_it->first) << "\n" );
}
xdoc.add_value(VALUE_SIG, doc.sig);
// Parse current data record into a dict for ease of processing
ConfSimple datadic(data);
if (!datadic.ok()) {
LOGERR("db::docToXdocXattrOnly: failed turning data rec to dict\n" );
return false;
}
// For each "stored" field, check if set in doc metadata and
// update the value if it is
const set<string>& stored = m_rcldb->m_config->getStoredFields();
for (set<string>::const_iterator it = stored.begin();
it != stored.end(); it++) {
string nm = m_rcldb->m_config->fieldCanon(*it);
if (doc.getmeta(nm, 0)) {
string value = neutchars(
truncate_to_word(doc.meta[nm], m_rcldb->m_idxMetaStoredLen),
cstr_nc);
datadic.set(nm, value, "");
}
}
// Recreate the record. We want to do this with the local RECORD_APPEND
// method for consistency in format, instead of using ConfSimple print
vector<string> names = datadic.getNames("");
data.clear();
for (vector<string>::const_iterator it = names.begin();
it != names.end(); it++) {
string value;
datadic.get(*it, value, "");
RECORD_APPEND(data, *it, value);
}
RECORD_APPEND(data, Doc::keysig, doc.sig);
xdoc.set_data(data);
return true;
}
#ifdef IDX_THREADS
void Db::waitUpdIdle()
{
if (m_ndb->m_iswritable && m_ndb->m_havewriteq) {
Chrono chron;
m_ndb->m_wqueue.waitIdle();
// We flush here just for correct measurement of the thread work time
string ermsg;
try {
m_ndb->xwdb.commit();
} XCATCHERROR(ermsg);
if (!ermsg.empty()) {
LOGERR("Db::waitUpdIdle: flush() failed: " << (ermsg) << "\n" );
}
m_ndb->m_totalworkns += chron.nanos();
LOGINFO("Db::waitUpdIdle: total xapian work " << (lltodecstr(m_ndb->m_totalworkns/1000000)) << " mS\n" );
}
}
#endif
// Flush when idxflushmbs is reached
bool Db::maybeflush(off_t moretext)
{
if (m_flushMb > 0) {
m_curtxtsz += moretext;
if ((m_curtxtsz - m_flushtxtsz) / MB >= m_flushMb) {
LOGDEB("Db::add/delete: txt size >= " << (m_flushMb) << " Mb, flushing\n" );
return doFlush();
}
}
return true;
}
bool Db::doFlush()
{
if (!m_ndb) {
LOGERR("Db::doFLush: no ndb??\n" );
return false;
}
string ermsg;
try {
m_ndb->xwdb.commit();
} XCATCHERROR(ermsg);
if (!ermsg.empty()) {
LOGERR("Db::doFlush: flush() failed: " << (ermsg) << "\n" );
return false;
}
m_flushtxtsz = m_curtxtsz;
return true;
}
void Db::setExistingFlags(const string& udi, unsigned int docid)
{
if (m_mode == DbRO)
return;
if (docid == (unsigned int)-1) {
LOGERR("Db::setExistingFlags: called with bogus docid !!\n" );
return;
}
#ifdef IDX_THREADS
std::unique_lock<std::mutex> lock(m_ndb->m_mutex);
#endif
i_setExistingFlags(udi, docid);
}
void Db::i_setExistingFlags(const string& udi, unsigned int docid)
{
// Set the up to date flag for the document and its subdocs
if (docid >= updated.size()) {
LOGERR("needUpdate: existing docid beyond updated.size(). Udi [" << (udi) << "], docid " << (unsigned(docid)) << ", updated.size() " << ((unsigned)updated.size()) << "\n" );
return;
} else {
updated[docid] = true;
}
// Set the existence flag for all the subdocs (if any)
vector<Xapian::docid> docids;
if (!m_ndb->subDocs(udi, 0, docids)) {
LOGERR("Rcl::Db::needUpdate: can't get subdocs\n" );
return;
}
for (vector<Xapian::docid>::iterator it = docids.begin();
it != docids.end(); it++) {
if (*it < updated.size()) {
LOGDEB2("Db::needUpdate: docid " << (*it) << " set\n" );
updated[*it] = true;
}
}
}
// Test if doc given by udi has changed since last indexed (test sigs)
bool Db::needUpdate(const string &udi, const string& sig,
unsigned int *docidp, string *osigp)
{
if (m_ndb == 0)
return false;
if (osigp)
osigp->clear();
if (docidp)
*docidp = 0;
// If we are doing an in place or full reset, no need to test.
if (o_inPlaceReset || m_mode == DbTrunc) {
// For in place reset, pretend the doc existed, to enable
// subdoc purge. The value is only used as a boolean in this case.
if (docidp && o_inPlaceReset) {
*docidp = -1;
}
return true;
}
string uniterm = make_uniterm(udi);
string ermsg;
#ifdef IDX_THREADS
// Need to protect against interaction with the doc update/insert
// thread which also updates the existence map, and even multiple
// accesses to the readonly Xapian::Database are not allowed
// anyway
std::unique_lock<std::mutex> lock(m_ndb->m_mutex);
#endif
// Try to find the document indexed by the uniterm.
Xapian::PostingIterator docid;
XAPTRY(docid = m_ndb->xrdb.postlist_begin(uniterm), m_ndb->xrdb, m_reason);
if (!m_reason.empty()) {
LOGERR("Db::needUpdate: xapian::postlist_begin failed: " << (m_reason) << "\n" );
return false;
}
if (docid == m_ndb->xrdb.postlist_end(uniterm)) {
// No document exists with this path: we do need update
LOGDEB("Db::needUpdate:yes (new): [" << (uniterm) << "]\n" );
return true;
}
Xapian::Document xdoc;
XAPTRY(xdoc = m_ndb->xrdb.get_document(*docid), m_ndb->xrdb, m_reason);
if (!m_reason.empty()) {
LOGERR("Db::needUpdate: get_document error: " << (m_reason) << "\n" );
return true;
}
if (docidp) {
*docidp = *docid;
}
// Retrieve old file/doc signature from value
string osig;
XAPTRY(osig = xdoc.get_value(VALUE_SIG), m_ndb->xrdb, m_reason);
if (!m_reason.empty()) {
LOGERR("Db::needUpdate: get_value error: " << (m_reason) << "\n" );
return true;
}
LOGDEB2("Db::needUpdate: oldsig [" << (osig) << "] new [" << (sig) << "]\n" );
if (osigp) {
*osigp = osig;
}
// Compare new/old sig
if (sig != osig) {
LOGDEB("Db::needUpdate:yes: olsig [" << (osig) << "] new [" << (sig) << "] [" << (uniterm) << "]\n" );
// Db is not up to date. Let's index the file
return true;
}
// Up to date. Set the existance flags in the map for the doc and
// its subdocs.
LOGDEB("Db::needUpdate:no: [" << (uniterm) << "]\n" );
i_setExistingFlags(udi, *docid);
return false;
}
// Return existing stem db languages
vector<string> Db::getStemLangs()
{
LOGDEB("Db::getStemLang\n" );
vector<string> langs;
if (m_ndb == 0 || m_ndb->m_isopen == false)
return langs;
StemDb db(m_ndb->xrdb);
db.getMembers(langs);
return langs;
}
/**
* Delete stem db for given language
*/
bool Db::deleteStemDb(const string& lang)
{
LOGDEB("Db::deleteStemDb(" << (lang) << ")\n" );
if (m_ndb == 0 || m_ndb->m_isopen == false || !m_ndb->m_iswritable)
return false;
XapWritableSynFamily db(m_ndb->xwdb, synFamStem);
return db.deleteMember(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::createStemDbs(const vector<string>& langs)
{
LOGDEB("Db::createStemDbs\n" );
if (m_ndb == 0 || m_ndb->m_isopen == false || !m_ndb->m_iswritable) {
LOGERR("createStemDb: db not open or not writable\n" );
return false;
}
return createExpansionDbs(m_ndb->xwdb, langs);
}
/**
* 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 " << (m_ndb->m_isopen) << " m_iswritable " << (m_ndb->m_iswritable) << "\n" );
if (m_ndb->m_isopen == false || m_ndb->m_iswritable == false)
return false;
#ifdef IDX_THREADS
// If we manage our own write queue, make sure it's drained and closed
if (m_ndb->m_havewriteq)
m_ndb->m_wqueue.setTerminateAndWait();
// else we need to lock out other top level threads. This is just
// a precaution as they should have been waited for by the top
// level actor at this point
std::unique_lock<std::mutex> lock(m_ndb->m_mutex);
#endif // IDX_THREADS
// 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.commit();
} 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 {
if (m_flushMb > 0) {
// We use an average term length of 5 for
// estimating the doc sizes which is probably not
// accurate but gives rough consistency with what
// we do for add/update. I should fetch the doc
// size from the data record, but this would be
// bad for performance.
Xapian::termcount trms = m_ndb->xwdb.get_doclength(docid);
maybeflush(trms * 5);
}
m_ndb->xwdb.delete_document(docid);
LOGDEB("Db::purge: deleted document #" << (docid) << "\n" );
} catch (const Xapian::DocNotFoundError &) {
LOGDEB0("Db::purge: document #" << (docid) << " not found\n" );
} catch (const Xapian::Error &e) {
LOGERR("Db::purge: document #" << (docid) << ": " << (e.get_msg()) << "\n" );
} catch (...) {
LOGERR("Db::purge: document #" << (docid) << ": unknown error\n" );
}
purgecount++;
}
}
try {
m_ndb->xwdb.commit();
} catch (...) {
LOGERR("Db::purge: 2nd flush failed\n" );
}
return true;
}
// Test for doc existence.
bool Db::docExists(const string& uniterm)
{
#ifdef IDX_THREADS
// Need to protect read db against multiaccess.
std::unique_lock<std::mutex> lock(m_ndb->m_mutex);
#endif
string ermsg;
try {
Xapian::PostingIterator docid = m_ndb->xrdb.postlist_begin(uniterm);
if (docid == m_ndb->xrdb.postlist_end(uniterm)) {
return false;
} else {
return true;
}
} XCATCHERROR(ermsg);
if (!ermsg.empty()) {
LOGERR("Db::docExists(" << (uniterm) << ") " << (ermsg) << "\n" );
}
return false;
}
/* Delete document(s) for given unique identifier (doc and descendents) */
bool Db::purgeFile(const string &udi, bool *existed)
{
LOGDEB("Db:purgeFile: [" << (udi) << "]\n" );
if (m_ndb == 0 || !m_ndb->m_iswritable)
return false;
string uniterm = make_uniterm(udi);
bool exists = docExists(uniterm);
if (existed)
*existed = exists;
if (!exists)
return true;
#ifdef IDX_THREADS
if (m_ndb->m_havewriteq) {
DbUpdTask *tp = new DbUpdTask(DbUpdTask::Delete, udi, uniterm,
0, (size_t)-1);
if (!m_ndb->m_wqueue.put(tp)) {
LOGERR("Db::purgeFile:Cant queue task\n" );
return false;
} else {
return true;
}
}
#endif
/* We get there is IDX_THREADS is not defined or there is no queue */
return m_ndb->purgeFileWrite(false, udi, uniterm);
}
/* Delete subdocs with an out of date sig. We do this to purge
obsolete subdocs during a partial update where no general purge
will be done */
bool Db::purgeOrphans(const string &udi)
{
LOGDEB("Db:purgeOrphans: [" << (udi) << "]\n" );
if (m_ndb == 0 || !m_ndb->m_iswritable)
return false;
string uniterm = make_uniterm(udi);
#ifdef IDX_THREADS
if (m_ndb->m_havewriteq) {
DbUpdTask *tp = new DbUpdTask(DbUpdTask::PurgeOrphans, udi, uniterm,
0, (size_t)-1);
if (!m_ndb->m_wqueue.put(tp)) {
LOGERR("Db::purgeFile:Cant queue task\n" );
return false;
} else {
return true;
}
}
#endif
/* We get there is IDX_THREADS is not defined or there is no queue */
return m_ndb->purgeFileWrite(true, udi, uniterm);
}
bool Db::dbStats(DbStats& res)
{
if (!m_ndb || !m_ndb->m_isopen)
return false;
Xapian::Database xdb = m_ndb->xrdb;
XAPTRY(res.dbdoccount = xdb.get_doccount();
res.dbavgdoclen = xdb.get_avlength();
res.mindoclen = xdb.get_doclength_lower_bound();
res.maxdoclen = xdb.get_doclength_upper_bound();
, xdb, m_reason);
if (!m_reason.empty())
return false;
return true;
}
// Retrieve document defined by Unique doc identifier. This is used
// by the GUI history feature and by open parent/getenclosing
// ! The return value is always true except for fatal errors. Document
// existence should be tested by looking at doc.pc
bool Db::getDoc(const string &udi, const Doc& idxdoc, Doc &doc)
{
LOGDEB("Db:getDoc: [" << (udi) << "]\n" );
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;
Xapian::Document xdoc;
Xapian::docid docid;
int idxi = idxdoc.idxi;
if ((docid = m_ndb->getDoc(udi, idxi, xdoc))) {
string data = xdoc.get_data();
doc.meta[Rcl::Doc::keyudi] = udi;
return m_ndb->dbDataToRclDoc(docid, data, doc);
} else {
// 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: [" << (udi) << "]\n" );
return true;
}
}
bool Db::hasSubDocs(const Doc &idoc)
{
if (m_ndb == 0)
return false;
string inudi;
if (!idoc.getmeta(Doc::keyudi, &inudi) || inudi.empty()) {
LOGERR("Db::hasSubDocs: no input udi or empty\n" );
return false;
}
LOGDEB1("Db::hasSubDocs: idxi " << (idoc.idxi) << " inudi [" << (inudi) << "]\n" );
// Not sure why we perform both the subDocs() call and the test on
// has_children. The former will return docs if the input is a
// file-level document, but the latter should be true both in this
// case and if the input is already a subdoc, so the first test
// should be redundant. Does not hurt much in any case, to be
// checked one day.
vector<Xapian::docid> docids;
if (!m_ndb->subDocs(inudi, idoc.idxi, docids)) {
LOGDEB("Db::hasSubDocs: lower level subdocs failed\n" );
return false;
}
if (!docids.empty())
return true;
// Check if doc has an "has_children" term
if (m_ndb->hasTerm(inudi, idoc.idxi, has_children_term))
return true;
return false;
}
// Retrieve all subdocuments of a given one, which may not be a file-level
// one (in which case, we have to retrieve this first, then filter the ipaths)
bool Db::getSubDocs(const Doc &idoc, vector<Doc>& subdocs)
{
if (m_ndb == 0)
return false;
string inudi;
if (!idoc.getmeta(Doc::keyudi, &inudi) || inudi.empty()) {
LOGERR("Db::getSubDocs: no input udi or empty\n" );
return false;
}
string rootudi;
string ipath = idoc.ipath;
LOGDEB0("Db::getSubDocs: idxi " << (idoc.idxi) << " inudi [" << (inudi) << "] ipath [" << (ipath) << "]\n" );
if (ipath.empty()) {
// File-level doc. Use it as root
rootudi = inudi;
} else {
// See if we have a parent term
Xapian::Document xdoc;
if (!m_ndb->getDoc(inudi, idoc.idxi, xdoc)) {
LOGERR("Db::getSubDocs: can't get Xapian document\n" );
return false;
}
Xapian::TermIterator xit;
XAPTRY(xit = xdoc.termlist_begin();
xit.skip_to(wrap_prefix(parent_prefix)),
m_ndb->xrdb, m_reason);
if (!m_reason.empty()) {
LOGERR("Db::getSubDocs: xapian error: " << (m_reason) << "\n" );
return false;
}
if (xit == xdoc.termlist_end()) {
LOGERR("Db::getSubDocs: parent term not found\n" );
return false;
}
rootudi = strip_prefix(*xit);
}
LOGDEB("Db::getSubDocs: root: [" << (rootudi) << "]\n" );
// Retrieve all subdoc xapian ids for the root
vector<Xapian::docid> docids;
if (!m_ndb->subDocs(rootudi, idoc.idxi, docids)) {
LOGDEB("Db::getSubDocs: lower level subdocs failed\n" );
return false;
}
// Retrieve doc, filter, and build output list
for (int tries = 0; tries < 2; tries++) {
try {
for (vector<Xapian::docid>::const_iterator it = docids.begin();
it != docids.end(); it++) {
Xapian::Document xdoc = m_ndb->xrdb.get_document(*it);
string data = xdoc.get_data();
string docudi;
m_ndb->xdocToUdi(xdoc, docudi);
Doc doc;
doc.meta[Doc::keyudi] = docudi;
doc.meta[Doc::keyrr] = "100%";
doc.pc = 100;
if (!m_ndb->dbDataToRclDoc(*it, data, doc)) {
LOGERR("Db::getSubDocs: doc conversion error\n" );
return false;
}
if (ipath.empty() ||
FileInterner::ipathContains(ipath, doc.ipath)) {
subdocs.push_back(doc);
}
}
return true;
} catch (const Xapian::DatabaseModifiedError &e) {
m_reason = e.get_msg();
m_ndb->xrdb.reopen();
continue;
} XCATCHERROR(m_reason);
break;
}
LOGERR("Db::getSubDocs: Xapian error: " << (m_reason) << "\n" );
return false;
}
} // End namespace Rcl