Merge remote-tracking branch 'upstream/master'

[nominatim.git] / lib / TokenList.php

diff --git a/lib/TokenList.php b/lib/TokenList.php
index 1dcaa7f5091df7d4fa9e3763c89ddfb2442635f2..a419da6a9ad3429bc9087070038d491392ba1785 100644 (file)
--- a/lib/TokenList.php
+++ b/lib/TokenList.php
@@ -2,36 +2,103 @@
 
 namespace Nominatim;
 
 
 namespace Nominatim;
 

-require_once(CONST_BasePath.'/lib/TokenCountry.php');
-require_once(CONST_BasePath.'/lib/TokenHousenumber.php');
-require_once(CONST_BasePath.'/lib/TokenPostcode.php');
-require_once(CONST_BasePath.'/lib/TokenSpecialTerm.php');
-require_once(CONST_BasePath.'/lib/TokenWord.php');
-require_once(CONST_BasePath.'/lib/SpecialSearchOperator.php');
+require_once(CONST_LibDir.'/TokenCountry.php');
+require_once(CONST_LibDir.'/TokenHousenumber.php');
+require_once(CONST_LibDir.'/TokenPostcode.php');
+require_once(CONST_LibDir.'/TokenSpecialTerm.php');
+require_once(CONST_LibDir.'/TokenWord.php');
+require_once(CONST_LibDir.'/SpecialSearchOperator.php');

 
 /**
  * Saves information about the tokens that appear in a search query.
  *
 
 /**
  * Saves information about the tokens that appear in a search query.
  *

+ * Tokens are sorted by their normalized form, the token word. There are different
+ * kinds of tokens, represented by different Token* classes. Note that
+ * tokens do not have a common base class. All tokens need to have a field
+ * with the word id that points to an entry in the `word` database table
+ * but otherwise the information saved about a token can be very different.
+ *
+ * There are two different kinds of token words: full words and partial terms.
+ *
+ * Full words start with a space. They represent a complete name of a place.
+ * All special tokens are normally full words.
+ *
+ * Partial terms have no space at the beginning. They may represent a part of
+ * a name of a place (e.g. in the name 'World Trade Center' a partial term
+ * would be 'Trade' or 'Trade Center'). They are only used in TokenWord.

  */
 class TokenList
 {
     // List of list of tokens indexed by their word_token.
     private $aTokens = array();
 
  */
 class TokenList
 {
     // List of list of tokens indexed by their word_token.
     private $aTokens = array();
 

+
+    /**
+     * Return total number of tokens.
+     *
+     * @return Integer
+     */
+    public function count()
+    {
+        return count($this->aTokens);
+    }
+
+    /**
+     * Check if there are tokens for the given token word.
+     *
+     * @param string $sWord Token word to look for.
+     *
+     * @return bool True if there is one or more token for the token word.
+     */

     public function contains($sWord)
     {
         return isset($this->aTokens[$sWord]);
     }
 
     public function contains($sWord)
     {
         return isset($this->aTokens[$sWord]);
     }
 

+    /**
+     * Check if there are partial or full tokens for the given word.
+     *
+     * @param string $sWord Token word to look for.
+     *
+     * @return bool True if there is one or more token for the token word.
+     */
+    public function containsAny($sWord)
+    {
+        return isset($this->aTokens[$sWord]) || isset($this->aTokens[' '.$sWord]);
+    }
+
+    /**
+     * Get the list of tokens for the given token word.
+     *
+     * @param string $sWord Token word to look for.
+     *
+     * @return object[] Array of tokens for the given token word or an
+     *                  empty array if no tokens could be found.
+     */

     public function get($sWord)
     {
         return isset($this->aTokens[$sWord]) ? $this->aTokens[$sWord] : array();
     }
 
     public function get($sWord)
     {
         return isset($this->aTokens[$sWord]) ? $this->aTokens[$sWord] : array();
     }
 

+    public function getFullWordIDs()
+    {
+        $ids = array();
+
+        foreach ($this->aTokens as $aTokenList) {
+            foreach ($aTokenList as $oToken) {
+                if (is_a($oToken, '\Nominatim\Token\Word') && !$oToken->bPartial) {
+                    $ids[$oToken->iId] = $oToken->iId;
+                }
+            }
+        }
+
+        return $ids;
+    }
+

     /**
      * Add token information from the word table in the database.
      *
     /**
      * Add token information from the word table in the database.
      *

-     * @param object   $oDB           Database connection.
+     * @param object   $oDB           Nominatim::DB instance.

      * @param string[] $aTokens       List of tokens to look up in the database.
      * @param string[] $aCountryCodes List of country restrictions.
      * @param string   $sNormQuery    Normalized query string.
      * @param string[] $aTokens       List of tokens to look up in the database.
      * @param string[] $aCountryCodes List of country restrictions.
      * @param string   $sNormQuery    Normalized query string.


@@ -45,11 +112,11 @@ class TokenList
         $sSQL = 'SELECT word_id, word_token, word, class, type, country_code,';
         $sSQL .= ' operator, coalesce(search_name_count, 0) as count';
         $sSQL .= ' FROM word WHERE word_token in (';
         $sSQL = 'SELECT word_id, word_token, word, class, type, country_code,';
         $sSQL .= ' operator, coalesce(search_name_count, 0) as count';
         $sSQL .= ' FROM word WHERE word_token in (';

-        $sSQL .= join(',', array_map('getDBQuoted', $aTokens)).')';
+        $sSQL .= join(',', $oDB->getDBQuotedList($aTokens)).')';

 
         Debug::printSQL($sSQL);
 
 
         Debug::printSQL($sSQL);
 

-        $aDBWords = chksql($oDB->getAll($sSQL), 'Could not get word tokens.');
+        $aDBWords = $oDB->getAll($sSQL, null, 'Could not get word tokens.');

 
         foreach ($aDBWords as $aWord) {
             $oToken = null;
 
         foreach ($aDBWords as $aWord) {
             $oToken = null;


@@ -85,7 +152,7 @@ class TokenList
                         $iId,
                         $aWord['class'],
                         $aWord['type'],
                         $iId,
                         $aWord['class'],
                         $aWord['type'],

-                        $aWord['operator'] ? Operator::NONE : Operator::NEAR
+                        $aWord['operator'] ? Operator::NEAR : Operator::NONE

                     );
                 }
             } elseif ($aWord['country_code']) {
                     );
                 }
             } elseif ($aWord['country_code']) {


@@ -98,8 +165,9 @@ class TokenList
             } else {
                 $oToken = new Token\Word(
                     $iId,
             } else {
                 $oToken = new Token\Word(
                     $iId,

-                    $aWord['word'][0] != ' ',
-                    (int) $aWord['count']
+                    $aWord['word_token'][0] != ' ',
+                    (int) $aWord['count'],
+                    substr_count($aWord['word_token'], ' ')

                 );
             }
 
                 );
             }