plugins/af_sort_bayes/lib/class.naivebayesianstorage.php

   1 <?php
   2         /*
   3          ***** BEGIN LICENSE BLOCK *****
   4          This file is part of PHP Naive Bayesian Filter.
   5
   6          The Initial Developer of the Original Code is
   7          Loic d'Anterroches [loic_at_xhtml.net].
   8          Portions created by the Initial Developer are Copyright (C) 2003
   9          the Initial Developer. All Rights Reserved.
  10
  11          Contributor(s):
  12
  13          PHP Naive Bayesian Filter is free software; you can redistribute it
  14          and/or modify it under the terms of the GNU General Public License as
  15          published by the Free Software Foundation; either version 2 of
  16          the License, or (at your option) any later version.
  17
  18          PHP Naive Bayesian Filter is distributed in the hope that it will
  19          be useful, but WITHOUT ANY WARRANTY; without even the implied
  20          warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
  21          See the GNU General Public License for more details.
  22
  23          You should have received a copy of the GNU General Public License
  24          along with Foobar; if not, write to the Free Software
  25          Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  26
  27          Alternatively, the contents of this file may be used under the terms of
  28          the GNU Lesser General Public License Version 2.1 or later (the "LGPL"),
  29          in which case the provisions of the LGPL are applicable instead
  30          of those above.
  31
  32          ***** END LICENSE BLOCK *****
  33          */
  34
  35         /** Access to the storage of the data for the filter.
  36
  37          To avoid dependency with respect to any database, this class handle all the
  38          access to the data storage. You can provide your own class as long as
  39          all the methods are available. The current one rely on a MySQL database.
  40
  41          methods:
  42          - array getCategories()
  43          - bool  wordExists(string $word)
  44          - array getWord(string $word, string $categoryid)
  45
  46          */
  47         class NaiveBayesianStorage {
  48                 var $con = null;
  49                 var $owner_uid = null;
  50
  51                 function NaiveBayesianStorage($owner_uid) {
  52                         $this->con = Db::get();
  53                         $this->owner_uid = $owner_uid;
  54
  55                         return true;
  56                 }
  57
  58                 /** get the list of categories with basic data.
  59
  60                  @return array key = category ids, values = array(keys = 'probability', 'word_count')
  61                  */
  62                 function getCategories() {
  63                         $categories = array();
  64                         $rs = $this->con->query('SELECT * FROM ttrss_plugin_af_sort_bayes_categories WHERE owner_uid = ' . $this->owner_uid);
  65
  66                         while ($line = $this->con->fetch_assoc($rs)) {
  67                                 $categories[$line['id']] = array('probability' => $line['probability'],
  68                                         'category' => $line['category'],
  69                                         'word_count' => $line['word_count']
  70                                 );
  71                         }
  72
  73                         return $categories;
  74                 }
  75
  76                 function getCategoryByName($category) {
  77                         $rs = $this->con->query("SELECT id FROM ttrss_plugin_af_sort_bayes_categories WHERE category = '" .
  78                                 $this->con->escape_string($category) . "' AND owner_uid = " . $this->owner_uid);
  79
  80                         if ($this->con->num_rows($rs) != 0) {
  81                                 return $this->con->fetch_result($rs, 0, "id");
  82                         }
  83
  84                         return false;
  85                 }
  86
  87                 /** see if the word is an already learnt word.
  88                  @return bool
  89                  @param string word
  90                  */
  91                 function wordExists($word) {
  92                         $rs = $this->con->query("SELECT * FROM ttrss_plugin_af_sort_bayes_wordfreqs WHERE word='" . $this->con->escape_string($word) . "' AND
  93                                 owner_uid = " . $this->owner_uid);
  94
  95                         return $this->con->num_rows($rs) != 0;
  96                 }
  97
  98                 /** get details of a word in a category.
  99                  @return array ('count' => count)
 100                  @param  string word
 101                  @param  string category id
 102                  */
 103                 function getWord($word, $category_id) {
 104                         $details = array();
 105
 106                         $rs = $this->con->query("SELECT * FROM ttrss_plugin_af_sort_bayes_wordfreqs WHERE word='" .
 107                                 $this->con->escape_string($word) . "' AND category_id=" . (int)$category_id);
 108
 109                         if ($this->con->num_rows($rs) == 0 ) {
 110                                 $details['count'] = 0;
 111                         } else {
 112                                 $details['count'] = $this->con->fetch_result($rs, 0, "count");
 113                         }
 114
 115                         return $details;
 116                 }
 117
 118                 /** update a word in a category.
 119                  If the word is new in this category it is added, else only the count is updated.
 120
 121                  @return bool success
 122                  @param string word
 123                  @param int    count
 124                  @paran string category id
 125                  */
 126                 function updateWord($word, $count, $category_id) {
 127                         $oldword = $this->getWord($word, $category_id);
 128
 129                         if (0 == $oldword['count']) {
 130                                 return $this->con->query("INSERT INTO ttrss_plugin_af_sort_bayes_wordfreqs (word, category_id, count, owner_uid)
 131                                         VALUES ('" . $this->con->escape_string($word) . "', '" .
 132                                         (int)$category_id . "', '" .
 133                                         (int)$count . "', '".
 134                                         $this->owner_uid . "')");
 135                         }
 136                         else {
 137                                 return $this->con->query("UPDATE ttrss_plugin_af_sort_bayes_wordfreqs SET count = count + " . (int) $count . " WHERE category_id = '" . $this->con->escape_string($category_id) . "' AND word = '" . $this->con->escape_string($word) . "'");
 138                         }
 139                 }
 140
 141                 /** remove a word from a category.
 142
 143                  @return bool success
 144                  @param string word
 145                  @param int  count
 146                  @param string category id
 147                  */
 148                 function removeWord($word, $count, $category_id) {
 149                         $oldword = $this->getWord($word, $category_id);
 150
 151                         if (0 != $oldword['count'] && 0 >= ($oldword['count'] - $count)) {
 152                                 return $this->con->query("DELETE FROM ttrss_plugin_af_sort_bayes_wordfreqs WHERE word='" .
 153                                         $this->con->escape_string($word) . "' AND category_id='" .
 154                                         $this->con->escape_string($category_id) . "'");
 155                         }
 156                         else {
 157                                 return $this->con->query("UPDATE ttrss_plugin_af_sort_bayes_wordfreqs SET count = count - " .
 158                                         (int) $count . " WHERE category_id = '" . $this->con->escape_string($category_id) . "'
 159                                         AND word = '" . $this->con->escape_string($word) . "'");
 160                         }
 161                 }
 162
 163                 /** update the probabilities of the categories and word count.
 164                  This function must be run after a set of training
 165
 166                  @return bool sucess
 167                  */
 168                 function updateProbabilities() {
 169                         // first update the word count of each category
 170                         $rs = $this->con->query("SELECT SUM(count) AS total FROM ttrss_plugin_af_sort_bayes_wordfreqs WHERE owner_uid = ".$this->owner_uid);
 171
 172                         $total_words = $this->con->fetch_result($rs, 0, "total");
 173
 174                         if ($total_words == 0) {
 175                                 $this->con->query("UPDATE ttrss_plugin_af_sort_bayes_categories SET word_count=0, probability=0 WHERE owner_uid = " . $this->owner_uid);
 176                                 return true;
 177                         }
 178
 179                         $rs = $this->con->query("SELECT tc.id AS category_id, SUM(count) AS total FROM ttrss_plugin_af_sort_bayes_categories AS tc
 180                                 LEFT JOIN ttrss_plugin_af_sort_bayes_wordfreqs AS tw ON (tc.id = tw.category_id) WHERE tc.owner_uid = ".$this->owner_uid." GROUP BY tc.id");
 181
 182                         while ($line = $this->con->fetch_assoc($rs)) {
 183
 184                                 $proba = (int)$line['total'] / $total_words;
 185                                 $this->con->query("UPDATE ttrss_plugin_af_sort_bayes_categories SET word_count=" . (int) $line['total'] .
 186                                         ", probability=" . $proba . " WHERE id = '" . $line['category_id'] . "'");
 187                         }
 188
 189                         return true;
 190                 }
 191
 192                 /** save a reference in the database.
 193
 194                  @return bool success
 195                  @param  string reference if, must be unique
 196                  @param  string category id
 197                  @param  string content of the reference
 198                  */
 199                 function saveReference($doc_id, $category_id, $content) {
 200                         return $this->con->query("INSERT INTO ttrss_plugin_af_sort_bayes_references (document_id, category_id, owner_uid) VALUES
 201                                 ('" . $this->con->escape_string($doc_id) . "', '" .
 202                                         (int)$category_id . "', " .
 203                                         (int)$this->owner_uid . ")");
 204                 }
 205
 206                 /** get a reference from the database.
 207
 208                  @return array  reference( category_id => ...., content => ....)
 209                  @param  string id
 210                  */
 211                 function getReference($doc_id, $include_content = true)
 212                 {
 213
 214                         $ref = array();
 215                         $rs = $this->con->query("SELECT * FROM ttrss_plugin_af_sort_bayes_references WHERE document_id='" .
 216                                 $this->con->escape_string($doc_id) . "' AND owner_uid = " . $this->owner_uid);
 217
 218                         if ($this->con->num_rows($rs) == 0) {
 219                                 return $ref;
 220                         }
 221
 222                         $ref['category_id'] = $this->con->fetch_result($rs, 0, 'category_id');
 223                         $ref['id'] = $this->con->fetch_result($rs, 0, 'id');
 224                         $ref['document_id'] = $this->con->fetch_result($rs, 0, 'document_id');
 225
 226                         if ($include_content) {
 227                                 $rs = $this->con->query("SELECT content, title FROM ttrss_entries WHERE guid = '" .
 228                                         $this->con->escape_string($ref['document_id']) . "'");
 229
 230                                 if ($this->con->num_rows($rs) != 0) {
 231                                         $ref['content'] = mb_strtolower($this->con->fetch_result($rs, 0, 'title') . ' ' . strip_tags($this->con->fetch_result($rs, 0, 'content')));
 232                                 }
 233                         }
 234
 235                         return $ref;
 236                 }
 237
 238                 /** remove a reference from the database
 239
 240                  @return bool sucess
 241                  @param  string reference id
 242                  */
 243                 function removeReference($doc_id) {
 244
 245                         return $this->con->query("DELETE FROM ttrss_plugin_af_sort_bayes_references WHERE document_id='" . $this->con->escape_string($doc_id) . "' AND owner_uid = " . $this->owner_uid);
 246                 }
 247
 248         }