xapian-core/weight/ifb2weight.cc

   1 /** @file ifb2weight.cc
   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 #include "common/log2.h"
  25 #include "weightinternal.h"
  26
  27 #include "serialise-double.h"
  28
  29 #include "xapian/error.h"
  30
  31 using namespace std;
  32
  33 namespace Xapian {
  34
  35 IfB2Weight::IfB2Weight(double c)
  36    : param_c(c)
  37 {
  38     if (param_c <= 0)
  39         throw Xapian::InvalidArgumentError("Parameter c is invalid.");
  40     need_stat(AVERAGE_LENGTH);
  41     need_stat(DOC_LENGTH);
  42     need_stat(DOC_LENGTH_MIN);
  43     need_stat(COLLECTION_SIZE);
  44     need_stat(COLLECTION_FREQ);
  45     need_stat(WDF);
  46     need_stat(WDF_MAX);
  47     need_stat(WQF);
  48     need_stat(TERMFREQ);
  49 }
  50
  51 IfB2Weight *
  52 IfB2Weight::clone() const
  53 {
  54     return new IfB2Weight(param_c);
  55 }
  56
  57 void
  58 IfB2Weight::init(double factor)
  59 {
  60     double wdfn_upper = get_wdf_upper_bound();
  61     if (wdfn_upper == 0) {
  62         upper_bound = 0.0;
  63         return;
  64     }
  65
  66     double F = get_collection_freq();
  67     double N = get_collection_size();
  68
  69     wdfn_upper *= log2(1 + (param_c * get_average_length()) /
  70                     get_doclength_lower_bound());
  71
  72     // This term is constant for all documents.
  73     double idf_max = log2((N + 1.0) / (F + 0.5));
  74
  75     /* Calculate constant values to be used in get_sumpart(). */
  76     wqf_product_idf = get_wqf() * idf_max * factor;
  77     c_product_avlen = param_c * get_average_length();
  78     B_constant = (F + 1.0) / get_termfreq();
  79
  80     // wdfn * B = wdfn * (F + 1.0) / (get_termfreq() * (wdfn + 1.0)).
  81     // By cancelling out wdfn, we get (F + 1.0) / (get_termfreq() * (1.0 + 1.0 / wdfn)).
  82     // In order to maximize the product, we need to minimize the denominator, and so we use wdfn_upper.
  83     double max_wdfn_product_B = wdfn_upper * B_constant / (wdfn_upper + 1.0);
  84
  85     upper_bound = wqf_product_idf * max_wdfn_product_B * factor;
  86 }
  87
  88 string
  89 IfB2Weight::name() const
  90 {
  91     return "Xapian::IfB2Weight";
  92 }
  93
  94 string
  95 IfB2Weight::short_name() const
  96 {
  97     return "ifb2";
  98 }
  99
 100 string
 101 IfB2Weight::serialise() const
 102 {
 103     return serialise_double(param_c);
 104 }
 105
 106 IfB2Weight *
 107 IfB2Weight::unserialise(const string & s) const
 108 {
 109     const char *ptr = s.data();
 110     const char *end = ptr + s.size();
 111     double c = unserialise_double(&ptr, end);
 112     if (rare(ptr != end))
 113         throw Xapian::SerialisationError("Extra data in IfB2Weight::unserialise()");
 114     return new IfB2Weight(c);
 115 }
 116
 117 double
 118 IfB2Weight::get_sumpart(Xapian::termcount wdf, Xapian::termcount len,
 119                         Xapian::termcount) const
 120 {
 121     if (wdf == 0) return 0.0;
 122     double wdfn = wdf;
 123     wdfn *= log2(1 + c_product_avlen / len);
 124
 125     double wdfn_product_B = wdfn * B_constant / (wdfn + 1.0);
 126
 127     return (wqf_product_idf * wdfn_product_B);
 128 }
 129
 130 double
 131 IfB2Weight::get_maxpart() const
 132 {
 133     return upper_bound;
 134 }
 135
 136 double
 137 IfB2Weight::get_sumextra(Xapian::termcount, Xapian::termcount) const
 138 {
 139     return 0;
 140 }
 141
 142 double
 143 IfB2Weight::get_maxextra() const
 144 {
 145     return 0;
 146 }
 147
 148 IfB2Weight *
 149 IfB2Weight::create_from_parameters(const char * p) const
 150 {
 151     if (*p == '\0')
 152         return new Xapian::IfB2Weight();
 153     double k = 1.0;
 154     if (!Xapian::Weight::Internal::double_param(&p, &k))
 155         Xapian::Weight::Internal::parameter_error("Parameter is invalid", "ifb2");
 156     if (*p)
 157         Xapian::Weight::Internal::parameter_error("Extra data after parameter", "ifb2");
 158     return new Xapian::IfB2Weight(k);
 159 }
 160
 161 }