xapian-core/weight/ifb2weight.cc

   1 /** @file
   2  * @brief Xapian::IfB2Weight class - the IfB2 weighting scheme of the DFR framework.
   3  */
   4 /* Copyright (C) 2013,2014 Aarsh Shah
   5  *
   6  * This program is free software; you can redistribute it and/or
   7  * modify it under the terms of the GNU General Public License as
   8  * published by the Free Software Foundation; either version 2 of the
   9  * License, or (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License
  17  * along with this program; if not, write to the Free Software
  18  * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
  19  */
  20
  21 #include <config.h>
  22
  23 #include "xapian/weight.h"
  24
  25 #include "weightinternal.h"
  26
  27 #include "serialise-double.h"
  28
  29 #include "xapian/error.h"
  30
  31 #include <cmath>
  32
  33 using namespace std;
  34
  35 namespace Xapian {
  36
  37 IfB2Weight::IfB2Weight(double c)
  38     : param_c(c)
  39 {
  40     if (param_c <= 0)
  41         throw Xapian::InvalidArgumentError("Parameter c is invalid");
  42     need_stat(AVERAGE_LENGTH);
  43     need_stat(DOC_LENGTH);
  44     need_stat(DOC_LENGTH_MIN);
  45     need_stat(COLLECTION_SIZE);
  46     need_stat(COLLECTION_FREQ);
  47     need_stat(WDF);
  48     need_stat(WDF_MAX);
  49     need_stat(WQF);
  50     need_stat(TERMFREQ);
  51 }
  52
  53 IfB2Weight *
  54 IfB2Weight::clone() const
  55 {
  56     return new IfB2Weight(param_c);
  57 }
  58
  59 void
  60 IfB2Weight::init(double factor)
  61 {
  62     if (factor == 0.0) {
  63         // This object is for the term-independent contribution, and that's
  64         // always zero for this scheme.
  65         return;
  66     }
  67
  68     double wdfn_upper = get_wdf_upper_bound();
  69     if (wdfn_upper == 0) {
  70         upper_bound = 0.0;
  71         return;
  72     }
  73
  74     double F = get_collection_freq();
  75     double N = get_collection_size();
  76
  77     wdfn_upper *= log2(1 + (param_c * get_average_length()) /
  78                     get_doclength_lower_bound());
  79
  80     // This term is constant for all documents.
  81     double idf_max = log2((N + 1.0) / (F + 0.5));
  82
  83     /* Calculate constant values to be used in get_sumpart(). */
  84     wqf_product_idf = get_wqf() * idf_max * factor;
  85     c_product_avlen = param_c * get_average_length();
  86     B_constant = (F + 1.0) / get_termfreq();
  87
  88     // wdfn * B = wdfn * (F + 1.0) / (get_termfreq() * (wdfn + 1.0)).
  89     // By cancelling out wdfn, we get (F + 1.0) / (get_termfreq() * (1.0 + 1.0 / wdfn)).
  90     // In order to maximize the product, we need to minimize the denominator, and so we use wdfn_upper.
  91     double max_wdfn_product_B = wdfn_upper * B_constant / (wdfn_upper + 1.0);
  92
  93     upper_bound = wqf_product_idf * max_wdfn_product_B * factor;
  94 }
  95
  96 string
  97 IfB2Weight::name() const
  98 {
  99     return "Xapian::IfB2Weight";
 100 }
 101
 102 string
 103 IfB2Weight::short_name() const
 104 {
 105     return "ifb2";
 106 }
 107
 108 string
 109 IfB2Weight::serialise() const
 110 {
 111     return serialise_double(param_c);
 112 }
 113
 114 IfB2Weight *
 115 IfB2Weight::unserialise(const string & s) const
 116 {
 117     const char *ptr = s.data();
 118     const char *end = ptr + s.size();
 119     double c = unserialise_double(&ptr, end);
 120     if (rare(ptr != end))
 121         throw Xapian::SerialisationError("Extra data in IfB2Weight::unserialise()");
 122     return new IfB2Weight(c);
 123 }
 124
 125 double
 126 IfB2Weight::get_sumpart(Xapian::termcount wdf, Xapian::termcount len,
 127                         Xapian::termcount, Xapian::termcount) const
 128 {
 129     if (wdf == 0) return 0.0;
 130     double wdfn = wdf;
 131     wdfn *= log2(1 + c_product_avlen / len);
 132
 133     double wdfn_product_B = wdfn * B_constant / (wdfn + 1.0);
 134
 135     return (wqf_product_idf * wdfn_product_B);
 136 }
 137
 138 double
 139 IfB2Weight::get_maxpart() const
 140 {
 141     return upper_bound;
 142 }
 143
 144 static inline void
 145 parameter_error(const char* message)
 146 {
 147     Xapian::Weight::Internal::parameter_error(message, "ifb2");
 148 }
 149
 150 IfB2Weight *
 151 IfB2Weight::create_from_parameters(const char * p) const
 152 {
 153     if (*p == '\0')
 154         return new Xapian::IfB2Weight();
 155     double k = 1.0;
 156     if (!Xapian::Weight::Internal::double_param(&p, &k))
 157         parameter_error("Parameter is invalid");
 158     if (*p)
 159         parameter_error("Extra data after parameter");
 160     return new Xapian::IfB2Weight(k);
 161 }
 162
 163 }