xapian-core/api/omenquireinternal.h

   1 /** @file omenquireinternal.h
   2  * @brief Internals
   3  */
   4 /* Copyright 1999,2000,2001 BrightStation PLC
   5  * Copyright 2001,2002 Ananova Ltd
   6  * Copyright 2002,2003,2004,2005,2006,2007,2008,2009,2010,2011,2014,2015,2016 Olly Betts
   7  * Copyright 2009 Lemur Consulting Ltd
   8  * Copyright 2011 Action Without Borders
   9  *
  10  * This program is free software; you can redistribute it and/or
  11  * modify it under the terms of the GNU General Public License as
  12  * published by the Free Software Foundation; either version 2 of the
  13  * License, or (at your option) any later version.
  14  *
  15  * This program is distributed in the hope that it will be useful,
  16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18  * GNU General Public License for more details.
  19  *
  20  * You should have received a copy of the GNU General Public License
  21  * along with this program; if not, write to the Free Software
  22  * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301
  23  * USA
  24  */
  25
  26 #ifndef OM_HGUARD_OMENQUIREINTERNAL_H
  27 #define OM_HGUARD_OMENQUIREINTERNAL_H
  28
  29 #include "xapian/database.h"
  30 #include "xapian/document.h"
  31 #include "xapian/enquire.h"
  32 #include "xapian/query.h"
  33 #include "xapian/keymaker.h"
  34
  35 #include <algorithm>
  36 #include <cmath>
  37 #include <map>
  38 #include <set>
  39
  40 #include "weight/weightinternal.h"
  41
  42 using namespace std;
  43
  44 class OmExpand;
  45 class MultiMatch;
  46
  47 namespace Xapian {
  48
  49 class TermIterator;
  50
  51 namespace Internal {
  52
  53 /** An item resulting from a query.
  54  *  This item contains the document id, and the weight calculated for
  55  *  the document.
  56  */
  57 class MSetItem {
  58     public:
  59         MSetItem(double wt_, Xapian::docid did_)
  60                 : wt(wt_), did(did_), collapse_count(0) {}
  61
  62         MSetItem(double wt_, Xapian::docid did_, const string &key_)
  63                 : wt(wt_), did(did_), collapse_key(key_), collapse_count(0) {}
  64
  65         MSetItem(double wt_, Xapian::docid did_, const string &key_,
  66                  Xapian::doccount collapse_count_)
  67                 : wt(wt_), did(did_), collapse_key(key_),
  68                   collapse_count(collapse_count_) {}
  69
  70         void swap(MSetItem & o) {
  71             std::swap(wt, o.wt);
  72             std::swap(did, o.did);
  73             std::swap(collapse_key, o.collapse_key);
  74             std::swap(collapse_count, o.collapse_count);
  75             std::swap(sort_key, o.sort_key);
  76         }
  77
  78         /** Weight calculated. */
  79         double wt;
  80
  81         /** Document id. */
  82         Xapian::docid did;
  83
  84         /** Value which was used to collapse upon.
  85          *
  86          *  If the collapse option is not being used, this will always
  87          *  have a null value.
  88          *
  89          *  If the collapse option is in use, this will contain the collapse
  90          *  key's value for this particular item.  If the key is not present
  91          *  for this item, the value will be a null string.  Only one instance
  92          *  of each key value (apart from the null string) will be present in
  93          *  the items in the returned Xapian::MSet.
  94          */
  95         string collapse_key;
  96
  97         /** Count of collapses done on collapse_key so far
  98          *
  99          * This is normally 0, and goes up for each collapse done
 100          * It is not necessarily an indication of how many collapses
 101          * might be done if an exhaustive match was done
 102          */
 103         Xapian::doccount collapse_count;
 104
 105         /** Used when sorting by value. */
 106         string sort_key;
 107
 108         /// Return a string describing this object.
 109         string get_description() const;
 110 };
 111
 112 }
 113
 114 /** Internals of enquire system.
 115  *  This allows the implementation of Xapian::Enquire to be hidden and reference
 116  *  counted.
 117  */
 118 class Enquire::Internal : public Xapian::Internal::intrusive_base {
 119     friend class MSet::Internal;
 120     private:
 121         /// The database which this enquire object uses.
 122         const Xapian::Database db;
 123
 124         /// The user's query.
 125         Query query;
 126
 127         /// The query length.
 128         termcount qlen;
 129
 130         /// Copy not allowed
 131         Internal(const Internal &);
 132         /// Assignment not allowed
 133         void operator=(const Internal &);
 134
 135     public:
 136         typedef enum { REL, VAL, VAL_REL, REL_VAL } sort_setting;
 137
 138         Xapian::valueno collapse_key;
 139
 140         Xapian::doccount collapse_max;
 141
 142         Xapian::Enquire::docid_order order;
 143
 144         int percent_cutoff;
 145
 146         double weight_cutoff;
 147
 148         Xapian::valueno sort_key;
 149         sort_setting sort_by;
 150         bool sort_value_forward;
 151
 152         Xapian::Internal::opt_intrusive_ptr<KeyMaker> sorter;
 153
 154         double time_limit;
 155
 156         /** The weight to use for this query.
 157          *
 158          *  This is mutable so that the default BM25Weight object can be
 159          *  created lazily when first required.
 160          */
 161         mutable Weight * weight;
 162
 163         /// The weighting scheme to use for query expansion.
 164         std::string eweightname;
 165
 166         /// The parameter required for TradWeight query expansion.
 167         double expand_k;
 168
 169         vector<Xapian::Internal::opt_intrusive_ptr<MatchSpy>> spies;
 170
 171         explicit Internal(const Xapian::Database &databases);
 172         ~Internal();
 173
 174         /** Request a document from the database.
 175          */
 176         void request_doc(const Xapian::Internal::MSetItem &item) const;
 177
 178         /** Read a previously requested document from the database.
 179          */
 180         Xapian::Document read_doc(const Xapian::Internal::MSetItem &item) const;
 181
 182         Xapian::Document get_document(const Xapian::Internal::MSetItem &item) const;
 183
 184         void set_query(const Query & query_, termcount qlen_);
 185         const Query & get_query() const;
 186         MSet get_mset(Xapian::doccount first, Xapian::doccount maxitems,
 187                       Xapian::doccount check_at_least,
 188                       const RSet *omrset,
 189                       const MatchDecider *mdecider) const;
 190
 191         ESet get_eset(Xapian::termcount maxitems, const RSet & omrset, int flags,
 192                       const ExpandDecider *edecider, double min_wt) const;
 193
 194         TermIterator get_matching_terms(Xapian::docid did) const;
 195         TermIterator get_matching_terms(const Xapian::MSetIterator &it) const;
 196
 197         Xapian::doccount get_termfreq(const string &tname) const;
 198
 199         string get_description() const;
 200 };
 201
 202 class MSet::Internal : public Xapian::Internal::intrusive_base {
 203     public:
 204         /// Factor to multiply weights by to convert them to percentages.
 205         double percent_factor;
 206
 207     private:
 208         /** The set of documents which have been requested but not yet
 209          *  collected.
 210          */
 211         mutable set<Xapian::doccount> requested_docs;
 212
 213         /// Cache of documents, indexed by MSet index.
 214         mutable map<Xapian::doccount, Xapian::Document> indexeddocs;
 215
 216         /// Read and cache the documents so far requested.
 217         void read_docs() const;
 218
 219         /// Copy not allowed
 220         Internal(const Internal &);
 221         /// Assignment not allowed
 222         void operator=(const Internal &);
 223
 224     public:
 225         /// Xapian::Enquire reference, for getting documents.
 226         Xapian::Internal::intrusive_ptr<const Enquire::Internal> enquire;
 227
 228         /** Provides the term frequency and weight for each term in the query. */
 229         Xapian::Weight::Internal * stats;
 230
 231         /// A list of items comprising the (selected part of the) MSet.
 232         vector<Xapian::Internal::MSetItem> items;
 233
 234         /// Rank of first item in MSet.
 235         Xapian::doccount firstitem;
 236
 237         Xapian::doccount matches_lower_bound;
 238
 239         Xapian::doccount matches_estimated;
 240
 241         Xapian::doccount matches_upper_bound;
 242
 243         Xapian::doccount uncollapsed_lower_bound;
 244
 245         Xapian::doccount uncollapsed_estimated;
 246
 247         Xapian::doccount uncollapsed_upper_bound;
 248
 249         double max_possible;
 250
 251         double max_attained;
 252
 253         Internal()
 254                 : percent_factor(0),
 255                   stats(NULL),
 256                   firstitem(0),
 257                   matches_lower_bound(0),
 258                   matches_estimated(0),
 259                   matches_upper_bound(0),
 260                   uncollapsed_lower_bound(0),
 261                   uncollapsed_estimated(0),
 262                   uncollapsed_upper_bound(0),
 263                   max_possible(0),
 264                   max_attained(0) {}
 265
 266         /// Note: destroys parameter items.
 267         Internal(Xapian::doccount firstitem_,
 268              Xapian::doccount matches_upper_bound_,
 269              Xapian::doccount matches_lower_bound_,
 270              Xapian::doccount matches_estimated_,
 271              Xapian::doccount uncollapsed_upper_bound_,
 272              Xapian::doccount uncollapsed_lower_bound_,
 273              Xapian::doccount uncollapsed_estimated_,
 274              double max_possible_,
 275              double max_attained_,
 276              vector<Xapian::Internal::MSetItem> &items_,
 277              double percent_factor_)
 278                 : percent_factor(percent_factor_),
 279                   stats(NULL),
 280                   firstitem(firstitem_),
 281                   matches_lower_bound(matches_lower_bound_),
 282                   matches_estimated(matches_estimated_),
 283                   matches_upper_bound(matches_upper_bound_),
 284                   uncollapsed_lower_bound(uncollapsed_lower_bound_),
 285                   uncollapsed_estimated(uncollapsed_estimated_),
 286                   uncollapsed_upper_bound(uncollapsed_upper_bound_),
 287                   max_possible(max_possible_),
 288                   max_attained(max_attained_) {
 289             std::swap(items, items_);
 290         }
 291
 292         ~Internal() { delete stats; }
 293
 294         /// get a document by index in MSet, via the cache.
 295         Xapian::Document get_doc_by_index(Xapian::doccount index) const;
 296
 297         /// Converts a weight to a percentage weight
 298         int convert_to_percent_internal(double wt) const;
 299
 300         std::string snippet(const std::string & text, size_t length,
 301                             const Xapian::Stem & stemmer,
 302                             unsigned flags,
 303                             const std::string & hi_start,
 304                             const std::string & hi_end,
 305                             const std::string & omit) const;
 306
 307         /// Return a string describing this object.
 308         string get_description() const;
 309
 310         /** Fetch items specified into the document cache.
 311          */
 312         void fetch_items(Xapian::doccount first, Xapian::doccount last) const;
 313 };
 314
 315 class RSet::Internal : public Xapian::Internal::intrusive_base {
 316     friend class Xapian::RSet;
 317
 318     private:
 319         /// Items in the relevance set.
 320         set<Xapian::docid> items;
 321
 322     public:
 323         const set<Xapian::docid> & get_items() const { return items; }
 324
 325         /// Return a string describing this object.
 326         string get_description() const;
 327 };
 328
 329 }
 330
 331 #endif // OM_HGUARD_OMENQUIREINTERNAL_H