Parent: [e7a669] (diff)

Child: [533068] (diff)

Download this file

searchdatatox.cpp    1041 lines (926 with data), 34.7 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
/* Copyright (C) 2006 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.
*/
// Handle translation from rcl's SearchData structures to Xapian Queries
#include "autoconfig.h"
#include <stdio.h>
#include <string>
#include <vector>
#include <algorithm>
#include <sstream>
#include <iostream>
using namespace std;
#include "xapian.h"
#include "cstr.h"
#include "rcldb.h"
#include "rcldb_p.h"
#include "searchdata.h"
#include "debuglog.h"
#include "smallut.h"
#include "textsplit.h"
#include "unacpp.h"
#include "utf8iter.h"
#include "stoplist.h"
#include "rclconfig.h"
#include "termproc.h"
#include "synfamily.h"
#include "stemdb.h"
#include "expansiondbs.h"
#include "base64.h"
#include "daterange.h"
namespace Rcl {
typedef vector<SearchDataClause *>::iterator qlist_it_t;
static const int original_term_wqf_booster = 10;
// Expand doc categories and mime type wild card expressions
//
// Categories are expanded against the configuration, mimetypes
// against the index.
bool SearchData::expandFileTypes(Db &db, vector<string>& tps)
{
const RclConfig *cfg = db.getConf();
if (!cfg) {
LOGFATAL(("Db::expandFileTypes: null configuration!!\n"));
return false;
}
vector<string> exptps;
for (vector<string>::iterator it = tps.begin(); it != tps.end(); it++) {
if (cfg->isMimeCategory(*it)) {
vector<string>tps;
cfg->getMimeCatTypes(*it, tps);
exptps.insert(exptps.end(), tps.begin(), tps.end());
} else {
TermMatchResult res;
string mt = stringtolower((const string&)*it);
// We set casesens|diacsens to get an equivalent of ixTermMatch()
db.termMatch(Db::ET_WILD|Db::ET_CASESENS|Db::ET_DIACSENS, string(),
mt, res, -1, "mtype");
if (res.entries.empty()) {
exptps.push_back(it->c_str());
} else {
for (vector<TermMatchEntry>::const_iterator rit =
res.entries.begin(); rit != res.entries.end(); rit++) {
exptps.push_back(strip_prefix(rit->term));
}
}
}
}
sort(exptps.begin(), exptps.end());
exptps.erase(unique(exptps.begin(), exptps.end()), exptps.end());
tps = exptps;
return true;
}
static const char *maxXapClauseMsg =
"Maximum Xapian query size exceeded. Increase maxXapianClauses "
"in the configuration. ";
static const char *maxXapClauseCaseDiacMsg =
"Or try to use case (C) or diacritics (D) sensitivity qualifiers, or less "
"wildcards ?"
;
// Walk the clauses list, translate each and add to top Xapian Query
bool SearchData::clausesToQuery(Rcl::Db &db, SClType tp,
vector<SearchDataClause*>& query,
string& reason, void *d)
{
Xapian::Query xq;
for (qlist_it_t it = query.begin(); it != query.end(); it++) {
Xapian::Query nq;
if (!(*it)->toNativeQuery(db, &nq)) {
LOGERR(("SearchData::clausesToQuery: toNativeQuery failed: %s\n",
(*it)->getReason().c_str()));
reason += (*it)->getReason() + " ";
return false;
}
if (nq.empty()) {
LOGDEB(("SearchData::clausesToQuery: skipping empty clause\n"));
continue;
}
// If this structure is an AND list, must use AND_NOT for excl clauses.
// Else this is an OR list, and there can't be excl clauses (checked by
// addClause())
Xapian::Query::op op;
if (tp == SCLT_AND) {
if ((*it)->getexclude()) {
op = Xapian::Query::OP_AND_NOT;
} else {
op = Xapian::Query::OP_AND;
}
} else {
op = Xapian::Query::OP_OR;
}
if (xq.empty()) {
if (op == Xapian::Query::OP_AND_NOT)
xq = Xapian::Query(op, Xapian::Query::MatchAll, nq);
else
xq = nq;
} else {
xq = Xapian::Query(op, xq, nq);
}
if (int(xq.get_length()) >= getMaxCl()) {
LOGERR(("%s\n", maxXapClauseMsg));
m_reason += maxXapClauseMsg;
if (!o_index_stripchars)
m_reason += maxXapClauseCaseDiacMsg;
return false;
}
}
LOGDEB0(("SearchData::clausesToQuery: got %d clauses\n", xq.get_length()));
if (xq.empty())
xq = Xapian::Query::MatchAll;
*((Xapian::Query *)d) = xq;
return true;
}
bool SearchData::toNativeQuery(Rcl::Db &db, void *d)
{
LOGDEB(("SearchData::toNativeQuery: stemlang [%s]\n", m_stemlang.c_str()));
m_reason.erase();
db.getConf()->getConfParam("maxTermExpand", &m_maxexp);
db.getConf()->getConfParam("maxXapianClauses", &m_maxcl);
// Walk the clause list translating each in turn and building the
// Xapian query tree
Xapian::Query xq;
if (!clausesToQuery(db, m_tp, m_query, m_reason, &xq)) {
LOGERR(("SearchData::toNativeQuery: clausesToQuery failed. reason: %s\n",
m_reason.c_str()));
return false;
}
if (m_haveDates) {
// If one of the extremities is unset, compute db extremas
if (m_dates.y1 == 0 || m_dates.y2 == 0) {
int minyear = 1970, maxyear = 2100;
if (!db.maxYearSpan(&minyear, &maxyear)) {
LOGERR(("Can't retrieve index min/max dates\n"));
//whatever, go on.
}
if (m_dates.y1 == 0) {
m_dates.y1 = minyear;
m_dates.m1 = 1;
m_dates.d1 = 1;
}
if (m_dates.y2 == 0) {
m_dates.y2 = maxyear;
m_dates.m2 = 12;
m_dates.d2 = 31;
}
}
LOGDEB(("Db::toNativeQuery: date interval: %d-%d-%d/%d-%d-%d\n",
m_dates.y1, m_dates.m1, m_dates.d1,
m_dates.y2, m_dates.m2, m_dates.d2));
Xapian::Query dq = date_range_filter(m_dates.y1, m_dates.m1, m_dates.d1,
m_dates.y2, m_dates.m2, m_dates.d2);
if (dq.empty()) {
LOGINFO(("Db::toNativeQuery: date filter is empty\n"));
}
// If no probabilistic query is provided then promote the daterange
// filter to be THE query instead of filtering an empty query.
if (xq.empty()) {
LOGINFO(("Db::toNativeQuery: proba query is empty\n"));
xq = dq;
} else {
xq = Xapian::Query(Xapian::Query::OP_FILTER, xq, dq);
}
}
if (m_minSize != size_t(-1) || m_maxSize != size_t(-1)) {
Xapian::Query sq;
char min[50], max[50];
sprintf(min, "%lld", (long long)m_minSize);
sprintf(max, "%lld", (long long)m_maxSize);
if (m_minSize == size_t(-1)) {
string value(max);
leftzeropad(value, 12);
sq = Xapian::Query(Xapian::Query::OP_VALUE_LE, VALUE_SIZE, value);
} else if (m_maxSize == size_t(-1)) {
string value(min);
leftzeropad(value, 12);
sq = Xapian::Query(Xapian::Query::OP_VALUE_GE, VALUE_SIZE, value);
} else {
string minvalue(min);
leftzeropad(minvalue, 12);
string maxvalue(max);
leftzeropad(maxvalue, 12);
sq = Xapian::Query(Xapian::Query::OP_VALUE_RANGE, VALUE_SIZE,
minvalue, maxvalue);
}
// If no probabilistic query is provided then promote the
// filter to be THE query instead of filtering an empty query.
if (xq.empty()) {
LOGINFO(("Db::toNativeQuery: proba query is empty\n"));
xq = sq;
} else {
xq = Xapian::Query(Xapian::Query::OP_FILTER, xq, sq);
}
}
// Add the autophrase if any
if (m_autophrase) {
Xapian::Query apq;
if (m_autophrase->toNativeQuery(db, &apq)) {
xq = xq.empty() ? apq :
Xapian::Query(Xapian::Query::OP_AND_MAYBE, xq, apq);
}
}
// Add the file type filtering clause if any
if (!m_filetypes.empty()) {
expandFileTypes(db, m_filetypes);
Xapian::Query tq;
for (vector<string>::iterator it = m_filetypes.begin();
it != m_filetypes.end(); it++) {
string term = wrap_prefix(mimetype_prefix) + *it;
LOGDEB0(("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);
}
// Add the neg file type filtering clause if any
if (!m_nfiletypes.empty()) {
expandFileTypes(db, m_nfiletypes);
Xapian::Query tq;
for (vector<string>::iterator it = m_nfiletypes.begin();
it != m_nfiletypes.end(); it++) {
string term = wrap_prefix(mimetype_prefix) + *it;
LOGDEB0(("Adding negative 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_AND_NOT, xq, tq);
}
*((Xapian::Query *)d) = xq;
return true;
}
// Splitter for breaking a user string into simple terms and
// phrases. This is for parts of the user entry which would appear as
// a single word because there is no white space inside, but are
// actually multiple terms to rcldb (ie term1,term2). Still, most of
// the time, the result of our splitting will be a single term.
class TextSplitQ : public TextSplitP {
public:
TextSplitQ(Flags flags, TermProc *prc)
: TextSplitP(prc, flags), m_nostemexp(false) {
}
bool takeword(const std::string &term, int pos, int bs, int be) {
// Check if the first letter is a majuscule in which
// case we do not want to do stem expansion. Need to do this
// before unac of course...
m_nostemexp = unaciscapital(term);
return TextSplitP::takeword(term, pos, bs, be);
}
bool nostemexp() const {
return m_nostemexp;
}
private:
bool m_nostemexp;
};
class TermProcQ : public TermProc {
public:
TermProcQ() : TermProc(0), m_alltermcount(0), m_lastpos(0), m_ts(0) {}
// We need a ref to the splitter (only it knows about orig term
// capitalization for controlling stemming. The ref can't be set
// in the constructor because the splitter is not built yet when
// we are born (chicken and egg).
void setTSQ(const TextSplitQ *ts) {
m_ts = ts;
}
bool takeword(const std::string &term, int pos, int bs, int be) {
m_alltermcount++;
if (m_lastpos < pos)
m_lastpos = pos;
bool noexpand = be ? m_ts->nostemexp() : true;
LOGDEB1(("TermProcQ::takeword: pushing [%s] pos %d noexp %d\n",
term.c_str(), pos, noexpand));
if (m_terms[pos].size() < term.size()) {
m_terms[pos] = term;
m_nste[pos] = noexpand;
}
return true;
}
bool flush() {
for (map<int, string>::const_iterator it = m_terms.begin();
it != m_terms.end(); it++) {
m_vterms.push_back(it->second);
m_vnostemexps.push_back(m_nste[it->first]);
}
return true;
}
int alltermcount() const {
return m_alltermcount;
}
int lastpos() const {
return m_lastpos;
}
const vector<string>& terms() {
return m_vterms;
}
const vector<bool>& nostemexps() {
return m_vnostemexps;
}
private:
// Count of terms including stopwords: this is for adjusting
// phrase/near slack
int m_alltermcount;
int m_lastpos;
const TextSplitQ *m_ts;
vector<string> m_vterms;
vector<bool> m_vnostemexps;
map<int, string> m_terms;
map<int, bool> m_nste;
};
/** Expand term into term list, using appropriate mode: stem, wildcards,
* diacritics...
*
* @param mods stem expansion, case and diacritics sensitivity control.
* @param term input single word
* @param oexp output expansion list
* @param sterm output original input term if there were no wildcards
* @param prefix field prefix in index. We could recompute it, but the caller
* has it already. Used in the simple case where there is nothing to expand,
* and we just return the prefixed term (else Db::termMatch deals with it).
* @param multiwords it may happen that synonym processing results in multi-word
* expansions which should be processed as phrases.
*/
bool SearchDataClauseSimple::expandTerm(Rcl::Db &db,
string& ermsg, int mods,
const string& term,
vector<string>& oexp, string &sterm,
const string& prefix,
vector<string>* multiwords
)
{
LOGDEB0(("expandTerm: mods 0x%x fld [%s] trm [%s] lang [%s]\n",
mods, m_field.c_str(), term.c_str(), getStemLang().c_str()));
sterm.clear();
oexp.clear();
if (term.empty())
return true;
bool maxexpissoft = false;
int maxexpand = getSoftMaxExp();
if (maxexpand != -1) {
maxexpissoft = true;
} else {
maxexpand = getMaxExp();
}
bool haswild = term.find_first_of(cstr_minwilds) != string::npos;
// If there are no wildcards, add term to the list of user-entered terms
if (!haswild) {
m_hldata.uterms.insert(term);
sterm = term;
}
// No stem expansion if there are wildcards or if prevented by caller
bool nostemexp = (mods & SDCM_NOSTEMMING) != 0;
if (haswild || getStemLang().empty()) {
LOGDEB2(("expandTerm: found wildcards or stemlang empty: no exp\n"));
nostemexp = true;
}
bool diac_sensitive = (mods & SDCM_DIACSENS) != 0;
bool case_sensitive = (mods & SDCM_CASESENS) != 0;
bool synonyms = (mods & SDCM_NOSYNS) == 0;
// noexpansion can be modified further down by possible case/diac expansion
bool noexpansion = nostemexp && !haswild && !synonyms;
if (o_index_stripchars) {
diac_sensitive = case_sensitive = false;
} else {
// If we are working with a raw index, apply the rules for case and
// diacritics sensitivity.
// If any character has a diacritic, we become
// diacritic-sensitive. Note that the way that the test is
// performed (conversion+comparison) will automatically ignore
// accented characters which are actually a separate letter
if (getAutoDiac() && unachasaccents(term)) {
LOGDEB0(("expandTerm: term has accents -> diac-sensitive\n"));
diac_sensitive = true;
}
// If any character apart the first is uppercase, we become
// case-sensitive. The first character is reserved for
// turning off stemming. You need to use a query language
// modifier to search for Floor in a case-sensitive way.
Utf8Iter it(term);
it++;
if (getAutoCase() && unachasuppercase(term.substr(it.getBpos()))) {
LOGDEB0(("expandTerm: term has uppercase -> case-sensitive\n"));
case_sensitive = true;
}
// If we are sensitive to case or diacritics turn stemming off
if (diac_sensitive || case_sensitive) {
LOGDEB0(("expandTerm: diac or case sens set -> "
"stemexpand and synonyms off\n"));
nostemexp = true;
synonyms = false;
}
if (!case_sensitive || !diac_sensitive)
noexpansion = false;
}
if (noexpansion) {
oexp.push_back(prefix + term);
m_hldata.terms[term] = term;
LOGDEB(("ExpandTerm: noexpansion: final: %s\n",
stringsToString(oexp).c_str()));
return true;
}
int termmatchsens = 0;
if (case_sensitive)
termmatchsens |= Db::ET_CASESENS;
if (diac_sensitive)
termmatchsens |= Db::ET_DIACSENS;
if (synonyms)
termmatchsens |= Db::ET_SYNEXP;
Db::MatchType mtyp = haswild ? Db::ET_WILD :
nostemexp ? Db::ET_NONE : Db::ET_STEM;
TermMatchResult res;
if (!db.termMatch(mtyp | termmatchsens, getStemLang(),
term, res, maxexpand, m_field, multiwords)) {
// Let it go through
}
// Term match entries to vector of terms
if (int(res.entries.size()) >= maxexpand && !maxexpissoft) {
ermsg = "Maximum term expansion size exceeded."
" Maybe use case/diacritics sensitivity or increase maxTermExpand.";
return false;
}
for (vector<TermMatchEntry>::const_iterator it = res.entries.begin();
it != res.entries.end(); it++) {
oexp.push_back(it->term);
}
// If the term does not exist at all in the db, the return from
// termMatch() is going to be empty, which is not what we want (we
// would then compute an empty Xapian query)
if (oexp.empty())
oexp.push_back(prefix + term);
// Remember the uterm-to-expansion links
for (vector<string>::const_iterator it = oexp.begin();
it != oexp.end(); it++) {
m_hldata.terms[strip_prefix(*it)] = term;
}
LOGDEB(("ExpandTerm: final: %s\n", stringsToString(oexp).c_str()));
return true;
}
// Do distribution of string vectors: a,b c,d -> a,c a,d b,c b,d
void multiply_groups(vector<vector<string> >::const_iterator vvit,
vector<vector<string> >::const_iterator vvend,
vector<string>& comb,
vector<vector<string> >&allcombs)
{
// Remember my string vector and compute next, for recursive calls.
vector<vector<string> >::const_iterator myvit = vvit++;
// Walk the string vector I'm called upon and, for each string,
// add it to current result, an call myself recursively on the
// next string vector. The last call (last element of the vector of
// vectors), adds the elementary result to the output
// Walk my string vector
for (vector<string>::const_iterator strit = (*myvit).begin();
strit != (*myvit).end(); strit++) {
// Add my current value to the string vector we're building
comb.push_back(*strit);
if (vvit == vvend) {
// Last call: store current result
allcombs.push_back(comb);
} else {
// Call recursively on next string vector
multiply_groups(vvit, vvend, comb, allcombs);
}
// Pop the value I just added (make room for the next element in my
// vector)
comb.pop_back();
}
}
static void prefix_vector(vector<string>& v, const string& prefix)
{
for (vector<string>::iterator it = v.begin(); it != v.end(); it++) {
*it = prefix + *it;
}
}
void SearchDataClauseSimple::
processSimpleSpan(Rcl::Db &db, string& ermsg,
const string& span,
int mods, void * pq)
{
vector<Xapian::Query>& pqueries(*(vector<Xapian::Query>*)pq);
LOGDEB0(("StringToXapianQ::processSimpleSpan: [%s] mods 0x%x\n",
span.c_str(), (unsigned int)mods));
vector<string> exp;
string sterm; // dumb version of user term
string prefix;
const FieldTraits *ftp;
if (!m_field.empty() && db.fieldToTraits(m_field, &ftp, true)) {
if (ftp->noterms)
addModifier(SDCM_NOTERMS); // Don't add terms to highlight data
prefix = wrap_prefix(ftp->pfx);
}
vector<string> multiwords;
if (!expandTerm(db, ermsg, mods, span, exp, sterm, prefix, &multiwords))
return;
// Set up the highlight data. No prefix should go in there
for (vector<string>::const_iterator it = exp.begin();
it != exp.end(); it++) {
m_hldata.groups.push_back(vector<string>(1, it->substr(prefix.size())));
m_hldata.slacks.push_back(0);
m_hldata.grpsugidx.push_back(m_hldata.ugroups.size() - 1);
}
// Push either term or OR of stem-expanded set
Xapian::Query xq(Xapian::Query::OP_OR, exp.begin(), exp.end());
m_curcl += exp.size();
// If sterm (simplified original user term) is not null, give it a
// relevance boost. We do this even if no expansion occurred (else
// the non-expanded terms in a term list would end-up with even
// less wqf). This does not happen if there are wildcards anywhere
// in the search.
// We normally boost the original term in the stem expansion list. Don't
// do it if there are wildcards anywhere, this would skew the results.
bool doBoostUserTerm =
(m_parentSearch && !m_parentSearch->haveWildCards()) ||
(m_parentSearch == 0 && !m_haveWildCards);
if (doBoostUserTerm && !sterm.empty()) {
xq = Xapian::Query(Xapian::Query::OP_OR, xq,
Xapian::Query(prefix+sterm,
original_term_wqf_booster));
}
// Push phrases for the multi-word expansions
for (vector<string>::const_iterator mwp = multiwords.begin();
mwp != multiwords.end(); mwp++) {
vector<string> phr;
// We just do a basic split to keep things a bit simpler here
// (no textsplit). This means though that no punctuation is
// allowed in multi-word synonyms.
stringToTokens(*mwp, phr);
if (!prefix.empty())
prefix_vector(phr, prefix);
xq = Xapian::Query(Xapian::Query::OP_OR, xq,
Xapian::Query(Xapian::Query::OP_PHRASE,
phr.begin(), phr.end()));
m_curcl++;
}
pqueries.push_back(xq);
}
// User entry element had several terms: transform into a PHRASE or
// NEAR xapian query, the elements of which can themselves be OR
// queries if the terms get expanded by stemming or wildcards (we
// don't do stemming for PHRASE though)
void SearchDataClauseSimple::processPhraseOrNear(Rcl::Db &db, string& ermsg,
TermProcQ *splitData,
int mods, void *pq,
bool useNear, int slack)
{
vector<Xapian::Query> &pqueries(*(vector<Xapian::Query>*)pq);
Xapian::Query::op op = useNear ? Xapian::Query::OP_NEAR :
Xapian::Query::OP_PHRASE;
vector<Xapian::Query> orqueries;
#ifdef XAPIAN_NEAR_EXPAND_SINGLE_BUF
bool hadmultiple = false;
#endif
vector<vector<string> >groups;
string prefix;
const FieldTraits *ftp;
if (!m_field.empty() && db.fieldToTraits(m_field, &ftp, true)) {
prefix = wrap_prefix(ftp->pfx);
}
if (mods & Rcl::SearchDataClause::SDCM_ANCHORSTART) {
orqueries.push_back(Xapian::Query(prefix + start_of_field_term));
slack++;
}
// Go through the list and perform stem/wildcard expansion for each element
vector<bool>::const_iterator nxit = splitData->nostemexps().begin();
for (vector<string>::const_iterator it = splitData->terms().begin();
it != splitData->terms().end(); it++, nxit++) {
LOGDEB0(("ProcessPhrase: processing [%s]\n", it->c_str()));
// Adjust when we do stem expansion. Not if disabled by
// caller, not inside phrases, and some versions of xapian
// will accept only one OR clause inside NEAR.
bool nostemexp = *nxit || (op == Xapian::Query::OP_PHRASE)
#ifdef XAPIAN_NEAR_EXPAND_SINGLE_BUF
|| hadmultiple
#endif // single OR inside NEAR
;
int lmods = mods;
if (nostemexp)
lmods |= SearchDataClause::SDCM_NOSTEMMING;
string sterm;
vector<string> exp;
if (!expandTerm(db, ermsg, lmods, *it, exp, sterm, prefix))
return;
LOGDEB0(("ProcessPhraseOrNear: exp size %d, exp: %s\n", exp.size(),
stringsToString(exp).c_str()));
// groups is used for highlighting, we don't want prefixes in there.
vector<string> noprefs;
for (vector<string>::const_iterator it = exp.begin();
it != exp.end(); it++) {
noprefs.push_back(it->substr(prefix.size()));
}
groups.push_back(noprefs);
orqueries.push_back(Xapian::Query(Xapian::Query::OP_OR,
exp.begin(), exp.end()));
m_curcl += exp.size();
if (m_curcl >= getMaxCl())
return;
#ifdef XAPIAN_NEAR_EXPAND_SINGLE_BUF
if (exp.size() > 1)
hadmultiple = true;
#endif
}
if (mods & Rcl::SearchDataClause::SDCM_ANCHOREND) {
orqueries.push_back(Xapian::Query(prefix + end_of_field_term));
slack++;
}
// Generate an appropriate PHRASE/NEAR query with adjusted slack
// For phrases, give a relevance boost like we do for original terms
LOGDEB2(("PHRASE/NEAR: alltermcount %d lastpos %d\n",
splitData->alltermcount(), splitData->lastpos()));
Xapian::Query xq(op, orqueries.begin(), orqueries.end(),
splitData->lastpos() + 1 + slack);
if (op == Xapian::Query::OP_PHRASE)
xq = Xapian::Query(Xapian::Query::OP_SCALE_WEIGHT, xq,
original_term_wqf_booster);
pqueries.push_back(xq);
// Add all combinations of NEAR/PHRASE groups to the highlighting data.
vector<vector<string> > allcombs;
vector<string> comb;
multiply_groups(groups.begin(), groups.end(), comb, allcombs);
// Insert the search groups and slacks in the highlight data, with
// a reference to the user entry that generated them:
m_hldata.groups.insert(m_hldata.groups.end(),
allcombs.begin(), allcombs.end());
m_hldata.slacks.insert(m_hldata.slacks.end(), allcombs.size(), slack);
m_hldata.grpsugidx.insert(m_hldata.grpsugidx.end(), allcombs.size(),
m_hldata.ugroups.size() - 1);
}
// Trim string beginning with ^ or ending with $ and convert to flags
static int stringToMods(string& s)
{
int mods = 0;
// Check for an anchored search
trimstring(s);
if (s.length() > 0 && s[0] == '^') {
mods |= Rcl::SearchDataClause::SDCM_ANCHORSTART;
s.erase(0, 1);
}
if (s.length() > 0 && s[s.length()-1] == '$') {
mods |= Rcl::SearchDataClause::SDCM_ANCHOREND;
s.erase(s.length()-1);
}
return mods;
}
/**
* Turn user entry string (NOT query language) into a list of xapian queries.
* We just separate words and phrases, and do wildcard and stem expansion,
*
* This is used to process data entered into an OR/AND/NEAR/PHRASE field of
* the GUI (in the case of NEAR/PHRASE, clausedist adds dquotes to the user
* entry).
*
* This appears awful, and it would seem that the split into
* terms/phrases should be performed in the upper layer so that we
* only receive pure term or near/phrase pure elements here, but in
* fact there are things that would appear like terms to naive code,
* and which will actually may be turned into phrases (ie: tom:jerry),
* in a manner which intimately depends on the index implementation,
* so that it makes sense to process this here.
*
* The final list contains one query for each term or phrase
* - Elements corresponding to a stem-expanded part are an OP_OR
* composition of the stem-expanded terms (or a single term query).
* - Elements corresponding to phrase/near are an OP_PHRASE/NEAR
* composition of the phrase terms (no stem expansion in this case)
* @return the subquery count (either or'd stem-expanded terms or phrase word
* count)
*/
bool SearchDataClauseSimple::processUserString(Rcl::Db &db, const string &iq,
string &ermsg, void *pq,
int slack, bool useNear)
{
vector<Xapian::Query> &pqueries(*(vector<Xapian::Query>*)pq);
int mods = m_modifiers;
LOGDEB(("StringToXapianQ:pUS:: qstr [%s] fld [%s] mods 0x%x "
"slack %d near %d\n",
iq.c_str(), m_field.c_str(), mods, slack, useNear));
ermsg.erase();
m_curcl = 0;
const StopList stops = db.getStopList();
// Simple whitespace-split input into user-level words and
// double-quoted phrases: word1 word2 "this is a phrase".
//
// The text splitter may further still decide that the resulting
// "words" are really phrases, this depends on separators:
// [paul@dom.net] would still be a word (span), but [about:me]
// will probably be handled as a phrase.
vector<string> phrases;
TextSplit::stringToStrings(iq, phrases);
// Process each element: textsplit into terms, handle stem/wildcard
// expansion and transform into an appropriate Xapian::Query
try {
for (vector<string>::iterator it = phrases.begin();
it != phrases.end(); it++) {
LOGDEB0(("strToXapianQ: phrase/word: [%s]\n", it->c_str()));
// Anchoring modifiers
int amods = stringToMods(*it);
int terminc = amods != 0 ? 1 : 0;
mods |= amods;
// If there are multiple spans in this element, including
// at least one composite, we have to increase the slack
// else a phrase query including a span would fail.
// Ex: "term0@term1 term2" is onlyspans-split as:
// 0 term0@term1 0 12
// 2 term2 13 18
// The position of term2 is 2, not 1, so a phrase search
// would fail.
// We used to do word split, searching for
// "term0 term1 term2" instead, which may have worse
// performance, but will succeed.
// We now adjust the phrase/near slack by comparing the term count
// and the last position
// The term processing pipeline:
// split -> [unac/case ->] stops -> store terms
TermProcQ tpq;
TermProc *nxt = &tpq;
TermProcStop tpstop(nxt, stops); nxt = &tpstop;
//TermProcCommongrams tpcommon(nxt, stops); nxt = &tpcommon;
//tpcommon.onlygrams(true);
TermProcPrep tpprep(nxt);
if (o_index_stripchars)
nxt = &tpprep;
TextSplitQ splitter(TextSplit::Flags(TextSplit::TXTS_ONLYSPANS |
TextSplit::TXTS_KEEPWILD),
nxt);
tpq.setTSQ(&splitter);
splitter.text_to_words(*it);
slack += tpq.lastpos() - tpq.terms().size() + 1;
LOGDEB0(("strToXapianQ: termcount: %d\n", tpq.terms().size()));
switch (tpq.terms().size() + terminc) {
case 0:
continue;// ??
case 1: {
int lmods = mods;
if (tpq.nostemexps().front())
lmods |= SearchDataClause::SDCM_NOSTEMMING;
m_hldata.ugroups.push_back(tpq.terms());
processSimpleSpan(db, ermsg, tpq.terms().front(),
lmods, &pqueries);
}
break;
default:
m_hldata.ugroups.push_back(tpq.terms());
processPhraseOrNear(db, ermsg, &tpq, mods, &pqueries,
useNear, slack);
}
if (m_curcl >= getMaxCl()) {
ermsg = maxXapClauseMsg;
if (!o_index_stripchars)
ermsg += maxXapClauseCaseDiacMsg;
break;
}
}
} 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.empty()) {
LOGERR(("stringToXapianQueries: %s\n", ermsg.c_str()));
return false;
}
return true;
}
// Translate a simple OR or AND search clause.
bool SearchDataClauseSimple::toNativeQuery(Rcl::Db &db, void *p)
{
LOGDEB(("SearchDataClauseSimple::toNativeQuery: fld [%s] val [%s] "
"stemlang [%s]\n", m_field.c_str(), m_text.c_str(),
getStemLang().c_str()));
Xapian::Query *qp = (Xapian::Query *)p;
*qp = Xapian::Query();
Xapian::Query::op op;
switch (m_tp) {
case SCLT_AND: op = Xapian::Query::OP_AND; break;
case SCLT_OR: op = Xapian::Query::OP_OR; break;
default:
LOGERR(("SearchDataClauseSimple: bad m_tp %d\n", m_tp));
m_reason = "Internal error";
return false;
}
vector<Xapian::Query> pqueries;
if (!processUserString(db, m_text, m_reason, &pqueries))
return false;
if (pqueries.empty()) {
LOGERR(("SearchDataClauseSimple: resolved to null query\n"));
m_reason = string("Resolved to null query. Term too long ? : [" +
m_text + string("]"));
return false;
}
*qp = Xapian::Query(op, pqueries.begin(), pqueries.end());
if (m_weight != 1.0) {
*qp = Xapian::Query(Xapian::Query::OP_SCALE_WEIGHT, *qp, m_weight);
}
return true;
}
// Translate a FILENAME search clause. This always comes
// from a "filename" search from the gui or recollq. A query language
// "filename:"-prefixed field will not go through here, but through
// the generic field-processing code.
//
// We do not split the entry any more (used to do some crazy thing
// about expanding multiple fragments in the past). We just take the
// value blanks and all and expand this against the indexed unsplit
// file names
bool SearchDataClauseFilename::toNativeQuery(Rcl::Db &db, void *p)
{
Xapian::Query *qp = (Xapian::Query *)p;
*qp = Xapian::Query();
int maxexp = getSoftMaxExp();
if (maxexp == -1)
maxexp = getMaxExp();
vector<string> names;
db.filenameWildExp(m_text, names, maxexp);
*qp = Xapian::Query(Xapian::Query::OP_OR, names.begin(), names.end());
if (m_weight != 1.0) {
*qp = Xapian::Query(Xapian::Query::OP_SCALE_WEIGHT, *qp, m_weight);
}
return true;
}
// Translate a dir: path filtering clause. See comments in .h
bool SearchDataClausePath::toNativeQuery(Rcl::Db &db, void *p)
{
LOGDEB(("SearchDataClausePath::toNativeQuery: [%s]\n", m_text.c_str()));
Xapian::Query *qp = (Xapian::Query *)p;
*qp = Xapian::Query();
if (m_text.empty()) {
LOGERR(("SearchDataClausePath: empty path??\n"));
m_reason = "Empty path ?";
return false;
}
vector<Xapian::Query> orqueries;
if (m_text[0] == '/')
orqueries.push_back(Xapian::Query(wrap_prefix(pathelt_prefix)));
else
m_text = path_tildexpand(m_text);
vector<string> vpath;
stringToTokens(m_text, vpath, "/");
for (vector<string>::const_iterator pit = vpath.begin();
pit != vpath.end(); pit++){
string sterm;
vector<string> exp;
if (!expandTerm(db, m_reason,
SDCM_NOSTEMMING|SDCM_CASESENS|SDCM_DIACSENS,
*pit, exp, sterm, wrap_prefix(pathelt_prefix))) {
return false;
}
LOGDEB0(("SDataPath::toNative: exp size %d. Exp: %s\n", exp.size(),
stringsToString(exp).c_str()));
if (exp.size() == 1)
orqueries.push_back(Xapian::Query(exp[0]));
else
orqueries.push_back(Xapian::Query(Xapian::Query::OP_OR,
exp.begin(), exp.end()));
m_curcl += exp.size();
if (m_curcl >= getMaxCl())
return false;
}
*qp = Xapian::Query(Xapian::Query::OP_PHRASE,
orqueries.begin(), orqueries.end());
if (m_weight != 1.0) {
*qp = Xapian::Query(Xapian::Query::OP_SCALE_WEIGHT, *qp, m_weight);
}
return true;
}
// Translate NEAR or PHRASE clause.
bool SearchDataClauseDist::toNativeQuery(Rcl::Db &db, void *p)
{
LOGDEB(("SearchDataClauseDist::toNativeQuery\n"));
Xapian::Query *qp = (Xapian::Query *)p;
*qp = Xapian::Query();
vector<Xapian::Query> pqueries;
// We produce a single phrase out of the user entry then use
// stringToXapianQueries() to lowercase and simplify the phrase
// terms etc. This will result into a single (complex)
// Xapian::Query.
if (m_text.find('\"') != string::npos) {
m_text = neutchars(m_text, "\"");
}
string s = cstr_dquote + m_text + cstr_dquote;
bool useNear = (m_tp == SCLT_NEAR);
if (!processUserString(db, s, m_reason, &pqueries, m_slack, useNear))
return false;
if (pqueries.empty()) {
LOGERR(("SearchDataClauseDist: resolved to null query\n"));
m_reason = string("Resolved to null query. Term too long ? : [" +
m_text + string("]"));
return false;
}
*qp = *pqueries.begin();
if (m_weight != 1.0) {
*qp = Xapian::Query(Xapian::Query::OP_SCALE_WEIGHT, *qp, m_weight);
}
return true;
}
} // Namespace Rcl