Parent: [ad52b0] (diff)

Child: [482185] (diff)

Download this file

rcldb.cpp    1097 lines (1000 with data), 32.9 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
#ifndef lint
static char rcsid[] = "@(#$Id: rcldb.cpp,v 1.40 2005-11-24 07:16:15 dockes Exp $ (C) 2004 J.F.Dockes";
#endif
#include <stdio.h>
#include <sys/stat.h>
#include <iostream>
#include <string>
#include <vector>
#include <algorithm>
#ifndef NO_NAMESPACES
using namespace std;
#endif /* NO_NAMESPACES */
#include "rcldb.h"
#include "textsplit.h"
#include "transcode.h"
#include "unacpp.h"
#include "conftree.h"
#include "debuglog.h"
#include "pathut.h"
#include "smallut.h"
#include "pathhash.h"
#include "xapian.h"
#include <xapian/stem.h>
// Data for a xapian database. There could actually be 2 different
// ones for indexing or query as there is not much in common.
class Native {
public:
bool isopen;
bool iswritable;
string basedir;
// Indexing
Xapian::WritableDatabase wdb;
vector<bool> updated;
// Querying
Xapian::Database db;
Xapian::Query query; // query descriptor: terms and subqueries
// joined by operators (or/and etc...)
Xapian::Enquire *enquire;
Xapian::MSet mset;
Native() : isopen(false), iswritable(false), enquire(0) { }
~Native() {
delete enquire;
}
};
Rcl::Db::Db()
{
pdata = new Native;
}
Rcl::Db::~Db()
{
LOGDEB1(("Rcl::Db::~Db\n"));
if (pdata == 0)
return;
Native *ndb = (Native *)pdata;
LOGDEB(("Db::~Db: isopen %d iswritable %d\n", ndb->isopen,
ndb->iswritable));
if (ndb->isopen == false)
return;
string ermsg;
try {
LOGDEB(("Rcl::Db::~Db: closing native database\n"));
if (ndb->iswritable == true)
ndb->wdb.flush();
delete ndb;
return;
} catch (const Xapian::Error &e) {
ermsg = e.get_msg();
} catch (const string &s) {
ermsg = s;
} catch (const char *s) {
ermsg = s;
} catch (...) {
ermsg = "Caught unknown exception";
}
LOGERR(("Rcl::Db::~Db: got exception: %s\n", ermsg.c_str()));
}
bool Rcl::Db::open(const string& dir, OpenMode mode)
{
if (pdata == 0)
return false;
Native *ndb = (Native *)pdata;
LOGDEB(("Db::open: isopen %d iswritable %d\n", ndb->isopen,
ndb->iswritable));
if (ndb->isopen) {
LOGERR(("Rcl::Db::open: already open\n"));
return false;
}
string ermsg;
try {
switch (mode) {
case DbUpd:
ndb->wdb =
Xapian::WritableDatabase(dir, Xapian::DB_CREATE_OR_OPEN);
LOGDEB(("Rcl::Db::open: lastdocid: %d\n",
ndb->wdb.get_lastdocid()));
ndb->updated.resize(ndb->wdb.get_lastdocid() + 1);
for (unsigned int i = 0; i < ndb->updated.size(); i++)
ndb->updated[i] = false;
ndb->iswritable = true;
break;
case DbTrunc:
break;
case DbRO:
default:
ndb->iswritable = false;
ndb->db = Xapian::Database(dir);
break;
}
ndb->isopen = true;
ndb->basedir = dir;
return true;
} catch (const Xapian::Error &e) {
ermsg = e.get_msg();
} catch (const string &s) {
ermsg = s;
} catch (const char *s) {
ermsg = s;
} catch (...) {
ermsg = "Caught unknown exception";
}
LOGERR(("Rcl::Db::open: exception while opening '%s': %s\n",
dir.c_str(), ermsg.c_str()));
return false;
}
// Note: xapian has no close call, we delete and recreate the db
bool Rcl::Db::close()
{
if (pdata == 0)
return false;
Native *ndb = (Native *)pdata;
LOGDEB(("Db::close(): isopen %d iswritable %d\n", ndb->isopen,
ndb->iswritable));
if (ndb->isopen == false)
return true;
string ermsg;
try {
if (ndb->iswritable == true) {
ndb->wdb.flush();
LOGDEB(("Rcl:Db: Called xapian flush\n"));
}
delete ndb;
pdata = new Native;
if (pdata)
return true;
} catch (const Xapian::Error &e) {
ermsg = e.get_msg();
} catch (const string &s) {
ermsg = s;
} catch (const char *s) {
ermsg = s;
} catch (...) {
ermsg = "Caught unknown exception";
}
LOGERR(("Rcl::Db:close: exception while deleting db: %s\n",
ermsg.c_str()));
return false;
}
bool Rcl::Db::isopen()
{
if (pdata == 0)
return false;
Native *ndb = (Native *)pdata;
return ndb->isopen;
}
// A small class to hold state while splitting text
class mySplitterCB : public TextSplitCB {
public:
Xapian::Document &doc;
Xapian::termpos basepos; // Base for document section
Xapian::termpos curpos; // Last position sent to callback
mySplitterCB(Xapian::Document &d) : doc(d), basepos(1), curpos(0)
{}
bool takeword(const std::string &term, int pos, int, int);
};
// Callback for the document to word splitting class during indexation
bool mySplitterCB::takeword(const std::string &term, int pos, int, int)
{
// cerr << "splitCb: term " << term << endl;
//string printable;
//transcode(term, printable, "UTF-8", "ISO-8859-1");
//cerr << "Adding " << printable << endl;
try {
// 1 is the value for wdfinc in index_text when called from omindex
// TOBEDONE: check what this is used for
curpos = pos;
doc.add_posting(term, basepos + curpos, 1);
} catch (...) {
LOGERR(("Rcl::Db: Error occurred during xapian add_posting\n"));
return false;
}
return true;
}
// Unaccent and lowercase data: use unac
// for accents, and do it by hand for upper / lower.
// TOBEDONE: lowercasing is done only for ascii letters, just A-Z -> a-z
// Removing crlfs is so that we can use the text in the document data fields.
bool Rcl::dumb_string(const string &in, string &out)
{
string inter;
out.erase();
if (in.empty())
return true;
if (!unac_cpp(in, inter)) {
LOGERR(("dumb_string: unac_cpp failed for %s\n", in.c_str()));
// Ok, no need to stop the whole show
inter = "";
}
out.reserve(inter.length());
for (unsigned int i = 0; i < inter.length(); i++) {
if (inter[i] >= 'A' && inter[i] <= 'Z') {
out += inter[i] + 'a' - 'A';
} else {
if (inter[i] == '\n' || inter[i] == '\r')
out += ' ';
else
out += inter[i];
}
}
return true;
}
/* omindex direct */
/* Truncate a string to a given maxlength, avoiding cutting off midword
* if reasonably possible. */
string
truncate_to_word(string & input, string::size_type maxlen)
{
string output;
if (input.length() <= maxlen) {
output = input;
} else {
output = input.substr(0, maxlen);
const char *SEPAR = " \t\n\r-:.;,/[]{}";
string::size_type space = output.find_last_of(SEPAR);
// Original version only truncated at space if space was found after
// maxlen/2. But we HAVE to truncate at space, else we'd need to do
// utf8 stuff to avoid truncating at multibyte char. In any case,
// not finding space means that the text probably has no value.
// Except probably for Asian languages, so we may want to fix this
// one day
if (space == string::npos) {
output.erase();
} else {
output.erase(space);
}
output += " ...";
}
// replace newlines with spaces
size_t i = 0;
while ((i = output.find('\n', i)) != string::npos) output[i] = ' ';
return output;
}
// Truncate longer path and uniquize with hash . The goad for this is
// to avoid xapian max term length limitations, not to gain space (we
// gain very little even with very short maxlens like 30)
#define HASHPATH
#define PATHHASHLEN 150
// Add document in internal form to the database: index the terms in
// the title abstract and body and add special terms for file name,
// date, mime type ... , create the document data record (more
// metadata), and update database
bool Rcl::Db::add(const string &fn, const Rcl::Doc &idoc)
{
LOGDEB1(("Rcl::Db::add: fn %s\n", fn.c_str()));
if (pdata == 0)
return false;
Native *ndb = (Native *)pdata;
// Truncate abstract, title and keywords to reasonable lengths
Rcl::Doc doc = idoc;
if (doc.abstract.empty())
doc.abstract = truncate_to_word(doc.text, 100);
else
doc.abstract = truncate_to_word(doc.abstract, 100);
doc.title = truncate_to_word(doc.title, 100);
doc.keywords = truncate_to_word(doc.keywords, 300);
Xapian::Document newdocument;
mySplitterCB splitData(newdocument);
TextSplit splitter(&splitData);
///////// Split and index terms in document body and auxiliary fields
// Split title and index terms
string noacc;
if (!dumb_string(doc.title, noacc)) {
LOGERR(("Rcl::Db::add: unac failed\n"));
return false;
}
splitter.text_to_words(noacc);
// Split body and index terms
splitData.basepos += splitData.curpos + 100;
if (!dumb_string(doc.text, noacc)) {
LOGERR(("Rcl::Db::add: dumb_string failed\n"));
return false;
}
splitter.text_to_words(noacc);
// Split keywords and index terms
splitData.basepos += splitData.curpos + 100;
if (!dumb_string(doc.keywords, noacc)) {
LOGERR(("Rcl::Db::add: dumb_string failed\n"));
return false;
}
splitter.text_to_words(noacc);
// Split abstract and index terms
splitData.basepos += splitData.curpos + 100;
if (!dumb_string(doc.abstract, noacc)) {
LOGERR(("Rcl::Db::add: dumb_string failed\n"));
return false;
}
splitter.text_to_words(noacc);
////// Special terms for metadata
// Mime type
newdocument.add_term("T" + doc.mimetype);
// Path name
string hash;
#ifdef HASHPATH
pathHash(fn, hash, PATHHASHLEN);
#else
hash = fn;
#endif
LOGDEB2(("Rcl::Db::add: pathhash [%s]\n", hash.c_str()));
string pathterm = "P" + hash;
newdocument.add_term(pathterm);
// File path + internal path: document unique identifier for
// documents inside multidocument files.
string uniterm;
if (!doc.ipath.empty()) {
uniterm = "Q" + hash + "|" + doc.ipath;
newdocument.add_term(uniterm);
}
// Dates etc...
time_t mtime = atol(doc.dmtime.empty() ? doc.fmtime.c_str() :
doc.dmtime.c_str());
struct tm *tm = localtime(&mtime);
char buf[9];
sprintf(buf, "%04d%02d%02d",tm->tm_year+1900, tm->tm_mon + 1, tm->tm_mday);
newdocument.add_term("D" + string(buf)); // Date (YYYYMMDD)
buf[7] = '\0';
if (buf[6] == '3') buf[6] = '2';
newdocument.add_term("W" + string(buf)); // "Weak" - 10ish day interval
buf[6] = '\0';
newdocument.add_term("M" + string(buf)); // Month (YYYYMM)
buf[4] = '\0';
newdocument.add_term("Y" + string(buf)); // Year (YYYY)
// Document data record. omindex has the following nl separated fields:
// - url
// - sample
// - caption (title limited to 100 chars)
// - mime type
string record = "url=file://" + fn;
record += "\nmtype=" + doc.mimetype;
record += "\nfmtime=" + doc.fmtime;
if (!doc.dmtime.empty())
record += "\ndmtime=" + doc.dmtime;
record += "\norigcharset=" + doc.origcharset;
record += "\ncaption=" + doc.title;
record += "\nkeywords=" + doc.keywords;
record += "\nabstract=" + doc.abstract;
if (!doc.ipath.empty()) {
record += "\nipath=" + doc.ipath;
}
record += "\n";
LOGDEB1(("Newdocument data: %s\n", record.c_str()));
newdocument.set_data(record);
const char *fnc = fn.c_str();
// Add db entry or update existing entry:
try {
Xapian::docid did =
ndb->wdb.replace_document(uniterm.empty() ? pathterm : uniterm,
newdocument);
if (did < ndb->updated.size()) {
ndb->updated[did] = true;
LOGDEB(("Rcl::Db::add: docid %d updated [%s , %s]\n", did, fnc,
doc.ipath.c_str()));
} else {
LOGDEB(("Rcl::Db::add: docid %d added [%s , %s]\n", did, fnc,
doc.ipath.c_str()));
}
} catch (...) {
// FIXME: is this ever actually needed?
ndb->wdb.add_document(newdocument);
LOGDEB(("Rcl::Db::add: %s added (failed re-seek for duplicate)\n",
fnc));
}
return true;
}
bool Rcl::Db::needUpdate(const string &filename, const struct stat *stp)
{
if (pdata == 0)
return false;
Native *ndb = (Native *)pdata;
// If no document exist with this path, we do need update
string hash;
#ifdef HASHPATH
pathHash(filename, hash, PATHHASHLEN);
#else
hash = filename;
#endif
string pathterm = "P" + hash;
if (!ndb->wdb.term_exists(pathterm)) {
LOGDEB2(("Db::needUpdate: path inexistant: %s\n", filename.c_str()));
return true;
}
// Look for all documents with this path. We need to look at all
// to set their existence flag. We check the update time on the
// fmtime field which will be identical for all docs inside a
// multi-document file (we currently always reindex all if the
// file changed)
Xapian::PostingIterator doc;
try {
Xapian::PostingIterator docid0 = ndb->wdb.postlist_begin(pathterm);
for (Xapian::PostingIterator docid = docid0;
docid != ndb->wdb.postlist_end(pathterm); docid++) {
Xapian::Document doc = ndb->wdb.get_document(*docid);
// Check the date once. no need to look at the others if
// the db needs updating. Note that the fmtime used to be
// called mtime, and we're keeping compat
if (docid == docid0) {
string data = doc.get_data();
const char *cp = strstr(data.c_str(), "fmtime=");
if (cp) {
cp += 7;
} else {
cp = strstr(data.c_str(), "mtime=");
if (cp)
cp+= 6;
}
long mtime = cp ? atol(cp) : 0;
if (mtime < stp->st_mtime) {
LOGDEB2(("Db::needUpdate: yes: mtime: Db %ld file %ld\n",
(long)mtime, (long)stp->st_mtime));
// Db is not up to date. Let's index the file
return true;
}
}
// Db is up to date. Make a note that this document exists.
if (*docid < ndb->updated.size())
ndb->updated[*docid] = true;
}
} catch (...) {
return true;
}
return false;
}
/// Compute name of stem db for given base database and language
static string stemdbname(const string& basename, string lang)
{
string nm = basename;
path_cat(nm, string("stem_") + lang);
return nm;
}
// Is char non-lowercase ascii ?
inline static bool
p_notlowerorutf(unsigned int c)
{
if (c < 'a' || (c > 'z' && c < 128))
return true;
return false;
}
/**
* Create database of stem to parents associations for a given language.
* We walk the list of all terms, stem them, and create another Xapian db
* with documents indexed by a single term (the stem), and with the list of
* parent terms in the document data.
*/
bool Rcl::Db::createStemDb(const string& lang)
{
LOGDEB(("Rcl::Db::createStemDb(%s)\n", lang.c_str()));
if (pdata == 0)
return false;
Native *ndb = (Native *)pdata;
if (ndb->isopen == false || ndb->iswritable == false)
return false;
// First build the in-memory stem database:
// We walk the list of all terms, and stem each.
// If the stem is identical to the term, no need to create an entry
// Else, we add an entry to the multimap.
// At the end, we only save stem-terms associations with several terms, the
// others are not useful
multimap<string, string> assocs;
// Statistics
int nostem=0; // Dont even try: not-alphanum (incomplete for now)
int stemconst=0; // Stem == term
int stemdiff=0; // Count of all different stems
int stemmultiple = 0; // Count of stems with multiple derivatives
try {
Xapian::Stem stemmer(lang);
Xapian::TermIterator it;
for (it = ndb->wdb.allterms_begin();
it != ndb->wdb.allterms_end(); it++) {
// If it has any non-lowercase 7bit char, cant be stemmable
string::iterator sit = (*it).begin(), eit = sit + (*it).length();
if ((sit = find_if(sit, eit, p_notlowerorutf)) != eit) {
++nostem;
// LOGDEB(("stemskipped: '%s', because of 0x%x\n",
// (*it).c_str(), *sit));
continue;
}
string stem = stemmer.stem_word(*it);
//cerr << "word " << *it << " stem " << stem << endl;
if (stem == *it) {
++stemconst;
continue;
}
assocs.insert(pair<string,string>(stem, *it));
}
} catch (...) {
LOGERR(("Stem database build failed: no stemmer for %s ? \n",
lang.c_str()));
return false;
}
// Create xapian database for stem relations
string stemdbdir = stemdbname(ndb->basedir, lang);
string ermsg = "NOERROR";
Xapian::WritableDatabase sdb;
try {
sdb = Xapian::WritableDatabase(stemdbdir,
Xapian::DB_CREATE_OR_OVERWRITE);
} catch (const Xapian::Error &e) {
ermsg = e.get_msg();
} catch (const string &s) {
ermsg = s;
} catch (const char *s) {
ermsg = s;
} catch (...) {
ermsg = "Caught unknown exception";
}
if (ermsg != "NOERROR") {
LOGERR(("Rcl::Db::createstemdb: exception while opening '%s': %s\n",
stemdbdir.c_str(), ermsg.c_str()));
return false;
}
// Enter pseud-docs in db. Walk the multimap, only enter
// associations where there are several parent terms
string stem;
list<string> derivs;
for (multimap<string,string>::const_iterator it = assocs.begin();
it != assocs.end(); it++) {
if (stem == it->first) {
// Staying with same stem
derivs.push_back(it->second);
// cerr << " " << it->second << endl;
} else {
// Changing stems
++stemdiff;
if (derivs.size() > 1) {
// Previous stem has multiple derivatives. Enter in db
++stemmultiple;
Xapian::Document newdocument;
newdocument.add_term(stem);
// The doc data is just parents=blank-separated-list
string record = "parents=";
for (list<string>::const_iterator it = derivs.begin();
it != derivs.end(); it++) {
record += *it + " ";
}
record += "\n";
LOGDEB1(("stemdocument data: %s\n", record.c_str()));
newdocument.set_data(record);
try {
sdb.replace_document(stem, newdocument);
} catch (...) {
LOGERR(("Rcl::Db::createstemdb: replace failed\n"));
return false;
}
}
derivs.clear();
stem = it->first;
derivs.push_back(it->second);
// cerr << "\n" << stem << " " << it->second;
}
}
LOGDEB(("Stem map size: %d stems %d mult %d no %d const %d\n",
assocs.size(), stemdiff, stemmultiple, nostem, stemconst));
return true;
}
/**
* This is called at the end of an indexing session, to delete the
* documents for files that are no longer there. We also build the
* stem database while we are at it.
*/
bool Rcl::Db::purge()
{
LOGDEB(("Rcl::Db::purge\n"));
if (pdata == 0)
return false;
Native *ndb = (Native *)pdata;
LOGDEB(("Rcl::Db::purge: isopen %d iswritable %d\n", ndb->isopen,
ndb->iswritable));
if (ndb->isopen == false || ndb->iswritable == false)
return false;
// There seems to be problems with the document delete code, when
// we do this, the database is not actually updated. Especially,
// if we delete a bunch of docs, so that there is a hole in the
// docids at the beginning, we can't add anything (appears to work
// and does nothing). Maybe related to the exceptions below when
// trying to delete an unexistant document ?
// Flushing before trying the deletes seeems to work around the problem
ndb->wdb.flush();
for (Xapian::docid docid = 1; docid < ndb->updated.size(); ++docid) {
if (!ndb->updated[docid]) {
try {
ndb->wdb.delete_document(docid);
LOGDEB(("Rcl::Db::purge: deleted document #%d\n", docid));
} catch (const Xapian::DocNotFoundError &) {
LOGDEB(("Rcl::Db::purge: document #%d not found\n", docid));
}
}
}
ndb->wdb.flush();
return true;
}
/**
* Expand term to list of all terms which stem to the same term.
*/
static list<string> stemexpand(Native *ndb, string term, const string& lang)
{
list<string> explist;
try {
Xapian::Stem stemmer(lang);
string stem = stemmer.stem_word(term);
LOGDEB(("stemexpand: '%s' stem-> '%s'\n", term.c_str(), stem.c_str()));
// Try to fetch the doc from the stem db
string stemdbdir = stemdbname(ndb->basedir, lang);
Xapian::Database sdb(stemdbdir);
LOGDEB1(("stemexpand: %s lastdocid: %d\n",
stemdbdir.c_str(), sdb.get_lastdocid()));
if (!sdb.term_exists(stem)) {
LOGDEB1(("Rcl::Db::stemexpand: no term for %s\n", stem.c_str()));
explist.push_back(term);
return explist;
}
Xapian::PostingIterator did = sdb.postlist_begin(stem);
if (did == sdb.postlist_end(stem)) {
LOGDEB1(("stemexpand: no term(1) for %s\n",stem.c_str()));
explist.push_back(term);
return explist;
}
Xapian::Document doc = sdb.get_document(*did);
string data = doc.get_data();
// No need for a conftree, but we need to massage the data a little
string::size_type pos = data.find_first_of("=");
++pos;
string::size_type pos1 = data.find_last_of("\n");
if (pos == string::npos || pos1 == string::npos ||pos1 <= pos) { // ??
explist.push_back(term);
return explist;
}
ConfTree::stringToStrings(data.substr(pos, pos1-pos), explist);
if (find(explist.begin(), explist.end(), term) == explist.end()) {
explist.push_back(term);
}
LOGDEB(("stemexpand: %s -> %s\n", stem.c_str(),
stringlistdisp(explist).c_str()));
} catch (...) {
LOGERR(("stemexpand: error accessing stem db\n"));
explist.push_back(term);
return explist;
}
return explist;
}
class wsQData : public TextSplitCB {
public:
vector<string> terms;
string catterms() {
string s;
for (unsigned int i=0;i<terms.size();i++) {
s += "[" + terms[i] + "] ";
}
return s;
}
bool takeword(const std::string &term, int , int, int) {
LOGDEB1(("wsQData::takeword: %s\n", term.c_str()));
terms.push_back(term);
return true;
}
void dumball() {
for (vector<string>::iterator it=terms.begin(); it !=terms.end();it++){
string dumb;
Rcl::dumb_string(*it, dumb);
*it = dumb;
}
}
};
//
// Turn string into list of xapian queries. There is little
// interpretation done on the string (no +term -term or filename:term
// stuff). We just separate words and phrases, and interpret
// capitalized terms as wanting no stem expansion. Elements of the
// list corresponding to a stem-expanded part are an OR query of the
// expanded elements
static void stringToXapianQueries(const string &iq,
const string& stemlang,
Native *ndb,
list<Xapian::Query> &pqueries,
Rcl::Db::QueryOpts opts = Rcl::Db::QO_NONE)
{
string qstring = iq;
#if 0
// Unaccent and lowerterm. Note that lowerterming here may not be
// such a good idea because it forbids using capitalized words to
// indicate that a term should not use stem expansion, for
// example.
if (!Rcl::dumb_string(iqstring, qstring))
return false;
#endif
// Split into (possibly single word) phrases ("this is a phrase"):
list<string> phrases;
ConfTree::stringToStrings(qstring, phrases);
// Then process each phrase: split into terms and transform into
// appropriate Xapian Query
for (list<string>::iterator it=phrases.begin(); it !=phrases.end(); it++) {
LOGDEB(("strToXapianQ: phrase or word: [%s]\n", it->c_str()));
wsQData splitData;
TextSplit splitter(&splitData, true);
splitter.text_to_words(*it);
LOGDEB(("strToXapianQ: splitter term count: %d\n",
splitData.terms.size()));
switch(splitData.terms.size()) {
case 0: continue;// ??
case 1: // Not a real phrase: one term
{
string term = splitData.terms.front();
bool nostemexp = false;
// Yes this doesnt work with accented or non-european
// majuscules. TOBEDONE: something :)
if (term.length() > 0 && term[0] >= 'A' && term[0] <= 'Z')
nostemexp = true;
LOGDEB(("Term: %s\n", term.c_str()));
// Possibly perform stem compression/expansion
list<string> exp;
string term1;
Rcl::dumb_string(term, term1);
if (!nostemexp && (opts & Rcl::Db::QO_STEM)) {
exp = stemexpand(ndb, term1, stemlang);
} else {
exp.push_back(term1);
}
// Push either term or OR of stem-expanded set
pqueries.push_back(Xapian::Query(Xapian::Query::OP_OR,
exp.begin(), exp.end()));
}
break;
default:
// Phrase: no stem expansion
splitData.dumball();
LOGDEB(("Pushing phrase: [%s]\n", splitData.catterms().c_str()));
pqueries.push_back(Xapian::Query(Xapian::Query::OP_PHRASE,
splitData.terms.begin(),
splitData.terms.end()));
}
}
}
// Prepare query out of simple query string
bool Rcl::Db::setQuery(const std::string &iqstring, QueryOpts opts,
const string& stemlang)
{
LOGDEB(("Rcl::Db::setQuery: q: '%s', opts 0x%x, stemlang %s\n",
iqstring.c_str(), (unsigned int)opts, stemlang.c_str()));
Native *ndb = (Native *)pdata;
if (!ndb)
return false;
asdata.erase();
dbindices.clear();
list<Xapian::Query> pqueries;
stringToXapianQueries(iqstring, stemlang, ndb, pqueries, opts);
ndb->query = Xapian::Query(Xapian::Query::OP_OR, pqueries.begin(),
pqueries.end());
delete ndb->enquire;
ndb->enquire = new Xapian::Enquire(ndb->db);
ndb->enquire->set_query(ndb->query);
ndb->mset = Xapian::MSet();
return true;
}
// Prepare query out of "advanced search" data
bool Rcl::Db::setQuery(AdvSearchData &sdata, QueryOpts opts,
const string& stemlang)
{
LOGDEB(("Rcl::Db::setQuery: adv:\n"));
LOGDEB((" allwords: %s\n", sdata.allwords.c_str()));
LOGDEB((" phrase: %s\n", sdata.phrase.c_str()));
LOGDEB((" orwords: %s\n", sdata.orwords.c_str()));
LOGDEB((" nowords: %s\n", sdata.nowords.c_str()));
string ft;
for (list<string>::iterator it = sdata.filetypes.begin();
it != sdata.filetypes.end(); it++) {ft += *it + " ";}
if (!ft.empty())
LOGDEB((" searched file types: %s\n", ft.c_str()));
if (!sdata.topdir.empty())
LOGDEB((" restricted to: %s\n", sdata.topdir.c_str()));
asdata = sdata;
dbindices.clear();
Native *ndb = (Native *)pdata;
if (!ndb)
return false;
list<Xapian::Query> pqueries;
Xapian::Query xq;
if (!sdata.allwords.empty()) {
stringToXapianQueries(sdata.allwords, stemlang, ndb, pqueries, opts);
if (!pqueries.empty()) {
xq = Xapian::Query(Xapian::Query::OP_AND, pqueries.begin(),
pqueries.end());
pqueries.clear();
}
}
if (!sdata.orwords.empty()) {
stringToXapianQueries(sdata.orwords, stemlang, ndb, pqueries, opts);
if (!pqueries.empty()) {
Xapian::Query nq;
nq = Xapian::Query(Xapian::Query::OP_OR, pqueries.begin(),
pqueries.end());
xq = xq.empty() ? nq :
Xapian::Query(Xapian::Query::OP_AND_MAYBE, xq, nq);
pqueries.clear();
}
}
// We do no stem expansion on 'No' words. Should we ?
if (!sdata.nowords.empty()) {
stringToXapianQueries(sdata.nowords, stemlang, ndb, pqueries);
if (!pqueries.empty()) {
Xapian::Query nq;
nq = Xapian::Query(Xapian::Query::OP_OR, pqueries.begin(),
pqueries.end());
xq = xq.empty() ? nq :
Xapian::Query(Xapian::Query::OP_AND_NOT, xq, nq);
pqueries.clear();
}
}
if (!sdata.phrase.empty()) {
Xapian::Query nq;
string s = string("\"") + sdata.phrase + string("\"");
stringToXapianQueries(s, stemlang, ndb, pqueries);
if (!pqueries.empty()) {
// There should be a single list element phrase query.
xq = xq.empty() ? *pqueries.begin() :
Xapian::Query(Xapian::Query::OP_AND, xq, *pqueries.begin());
pqueries.clear();
}
}
if (!sdata.filetypes.empty()) {
Xapian::Query tq;
for (list<string>::iterator it = sdata.filetypes.begin();
it != sdata.filetypes.end(); it++) {
string term = "T" + *it;
LOGDEB(("Adding file type term: [%s]\n", term.c_str()));
tq = tq.empty() ? Xapian::Query(term) :
Xapian::Query(Xapian::Query::OP_OR, tq, Xapian::Query(term));
}
xq = xq.empty() ? tq : Xapian::Query(Xapian::Query::OP_FILTER, xq, tq);
}
ndb->query = xq;
delete ndb->enquire;
ndb->enquire = new Xapian::Enquire(ndb->db);
ndb->enquire->set_query(ndb->query);
ndb->mset = Xapian::MSet();
return true;
}
bool Rcl::Db::getQueryTerms(list<string>& terms)
{
Native *ndb = (Native *)pdata;
if (!ndb)
return false;
terms.clear();
Xapian::TermIterator it;
for (it = ndb->query.get_terms_begin(); it != ndb->query.get_terms_end();
it++) {
terms.push_back(*it);
}
return true;
}
int Rcl::Db::getResCnt()
{
Native *ndb = (Native *)pdata;
if (!ndb || !ndb->enquire) {
LOGERR(("Rcl::Db::getResCnt: no query opened\n"));
return -1;
}
if (ndb->mset.size() <= 0)
return -1;
return ndb->mset.get_matches_lower_bound();
}
// This class (friend to RclDb) exists so that we can have functions that
// access private RclDb data and have Xapian-specific parameters (so that we
// don't want them to appear in the public rcldb.h).
class Rcl::DbPops {
public:
static bool filterMatch(Rcl::Db *rdb, Xapian::Document &xdoc) {
// Parse xapian document's data and populate doc fields
string data = xdoc.get_data();
ConfSimple parms(&data);
// The only filtering for now is on file path (subtree)
string url;
parms.get(string("url"), url);
url = url.substr(7);
if (url.find(rdb->asdata.topdir) == 0)
return true;
return false;
}
};
// Get document at rank i in query (i is the index in the whole result
// set, as in the enquire class. We check if the current mset has the
// doc, else ask for an other one. We use msets of 10 documents. Don't
// know if the whole thing makes sense at all but it seems to work.
//
// If there is a postquery filter (ie: file names), we have to
// maintain a correspondance from the sequential external index
// sequence to the internal Xapian hole-y one (the holes being the documents
// that dont match the filter).
bool Rcl::Db::getDoc(int exti, Doc &doc, int *percent)
{
const int qquantum = 30;
LOGDEB1(("Rcl::Db::getDoc: exti %d\n", exti));
Native *ndb = (Native *)pdata;
if (!ndb || !ndb->enquire) {
LOGERR(("Rcl::Db::getDoc: no query opened\n"));
return false;
}
// For now the only post-query filter is on dir subtree
bool postqfilter = !asdata.topdir.empty();
LOGDEB1(("Topdir %s postqflt %d\n", asdata.topdir.c_str(), postqfilter));
int xapi;
if (postqfilter) {
// There is a postquery filter, does this fall in already known area ?
if (exti >= (int)dbindices.size()) {
// Have to fetch xapian docs and filter until we get
// enough or fail
dbindices.reserve(exti+1);
// First xapian doc we fetch is the one after last stored
int first = dbindices.size() > 0 ? dbindices.back() + 1 : 0;
// Loop until we get enough docs
while (exti >= (int)dbindices.size()) {
LOGDEB(("Rcl::Db::getDoc: fetching %d starting at %d\n",
qquantum, first));
try {
ndb->mset = ndb->enquire->get_mset(first, qquantum);
} catch (const Xapian::DatabaseModifiedError &error) {
ndb->db.reopen();
ndb->mset = ndb->enquire->get_mset(first, qquantum);
} catch (const Xapian::Error & error) {
LOGERR(("enquire->get_mset: exception: %s\n",
error.get_msg().c_str()));
abort();
}
if (ndb->mset.empty()) {
LOGDEB(("Rcl::Db::getDoc: got empty mset\n"));
return false;
}
first = ndb->mset.get_firstitem();
for (unsigned int i = 0; i < ndb->mset.size() ; i++) {
LOGDEB(("Rcl::Db::getDoc: [%d]\n", i));
Xapian::Document xdoc = ndb->mset[i].get_document();
if (Rcl::DbPops::filterMatch(this, xdoc)) {
dbindices.push_back(first + i);
}
}
first = first + ndb->mset.size();
}
}
xapi = dbindices[exti];
} else {
xapi = exti;
}
// From there on, we work with a xapian enquire item number. Fetch it
int first = ndb->mset.get_firstitem();
int last = first + ndb->mset.size() -1;
if (!(xapi >= first && xapi <= last)) {
LOGDEB(("Fetching for first %d, count %d\n", xapi, qquantum));
try {
ndb->mset = ndb->enquire->get_mset(xapi, qquantum);
} catch (const Xapian::DatabaseModifiedError &error) {
ndb->db.reopen();
ndb->mset = ndb->enquire->get_mset(xapi, qquantum);
} catch (const Xapian::Error & error) {
LOGERR(("enquire->get_mset: exception: %s\n",
error.get_msg().c_str()));
abort();
}
if (ndb->mset.empty())
return false;
first = ndb->mset.get_firstitem();
last = first + ndb->mset.size() -1;
}
LOGDEB1(("Rcl::Db::getDoc: Qry '%s' win [%d-%d] Estimated results: %d",
ndb->query.get_description().c_str(),
first, last,
ndb->mset.get_matches_lower_bound()));
Xapian::Document xdoc = ndb->mset[xapi-first].get_document();
if (percent)
*percent = ndb->mset.convert_to_percent(ndb->mset[xapi-first]);
// Parse xapian document's data and populate doc fields
string data = xdoc.get_data();
LOGDEB1(("Rcl::Db::getDoc: data: %s\n", data.c_str()));
ConfSimple parms(&data);
parms.get(string("url"), doc.url);
parms.get(string("mtype"), doc.mimetype);
parms.get(string("fmtime"), doc.fmtime);
parms.get(string("dmtime"), doc.dmtime);
parms.get(string("origcharset"), doc.origcharset);
parms.get(string("caption"), doc.title);
parms.get(string("keywords"), doc.keywords);
parms.get(string("abstract"), doc.abstract);
parms.get(string("ipath"), doc.ipath);
return true;
}