switch instruction to use wikimedia CSV file

[nominatim.git] / lib-php / tokenizer / legacy_tokenizer.php

diff --git a/lib-php/tokenizer/legacy_tokenizer.php b/lib-php/tokenizer/legacy_tokenizer.php
index 570b88289e7cd13a68dd13d8c5af64c519f33fe6..6f3d23043f47071c4995c76545cba98073b6de7b 100644 (file)
--- a/lib-php/tokenizer/legacy_tokenizer.php
+++ b/lib-php/tokenizer/legacy_tokenizer.php
@@ -1,13 +1,22 @@
 <?php
 <?php

+/**
+ * SPDX-License-Identifier: GPL-2.0-only
+ *
+ * This file is part of Nominatim. (https://nominatim.org)
+ *
+ * Copyright (C) 2022 by the Nominatim developer community.
+ * For a full list of authors see the git log.
+ */

 
 namespace Nominatim;
 
 
 namespace Nominatim;
 

+require_once(CONST_LibDir.'/SimpleWordList.php');
+

 class Tokenizer
 {
     private $oDB;
 
     private $oNormalizer = null;
 class Tokenizer
 {
     private $oDB;
 
     private $oNormalizer = null;

-    private $aCountryRestriction = null;

 
     public function __construct(&$oDB)
     {
 
     public function __construct(&$oDB)
     {


@@ -37,12 +46,6 @@ class Tokenizer
     }
 
 
     }
 
 

-    public function setCountryRestriction($aCountries)
-    {
-        $this->aCountryRestriction = $aCountries;
-    }
-
-

     public function normalizeString($sTerm)
     {
         if ($this->oNormalizer === null) {
     public function normalizeString($sTerm)
     {
         if ($this->oNormalizer === null) {


@@ -53,6 +56,14 @@ class Tokenizer
     }
 
 
     }
 
 

+    public function mostFrequentWords($iNum)
+    {
+        $sSQL = 'SELECT word FROM word WHERE word is not null ';
+        $sSQL .= 'ORDER BY search_name_count DESC LIMIT '.$iNum;
+        return $this->oDB->getCol($sSQL);
+    }
+
+

     public function tokensForSpecialTerm($sTerm)
     {
         $aResults = array();
     public function tokensForSpecialTerm($sTerm)
     {
         $aResults = array();


@@ -92,6 +103,23 @@ class Tokenizer
             $sNormQuery .= ','.$this->normalizeString($oPhrase->getPhrase());
             $sSQL .= 'make_standard_name(:' .$iPhrase.') as p'.$iPhrase.',';
             $aParams[':'.$iPhrase] = $oPhrase->getPhrase();
             $sNormQuery .= ','.$this->normalizeString($oPhrase->getPhrase());
             $sSQL .= 'make_standard_name(:' .$iPhrase.') as p'.$iPhrase.',';
             $aParams[':'.$iPhrase] = $oPhrase->getPhrase();

+
+            // Conflicts between US state abbreviations and various words
+            // for 'the' in different languages
+            switch (strtolower($oPhrase->getPhrase())) {
+                case 'il':
+                    $aParams[':'.$iPhrase] = 'illinois';
+                    break;
+                case 'al':
+                    $aParams[':'.$iPhrase] = 'alabama';
+                    break;
+                case 'la':
+                    $aParams[':'.$iPhrase] = 'louisiana';
+                    break;
+                default:
+                    $aParams[':'.$iPhrase] = $oPhrase->getPhrase();
+                    break;
+            }

         }
         $sSQL = substr($sSQL, 0, -1);
 
         }
         $sSQL = substr($sSQL, 0, -1);
 


@@ -106,13 +134,14 @@ class Tokenizer
         $aWordLists = array();
         $aTokens = array();
         foreach ($aNormPhrases as $sPhrase) {
         $aWordLists = array();
         $aTokens = array();
         foreach ($aNormPhrases as $sPhrase) {

-            if (strlen($sPhrase) > 0) {
-                $aWords = explode(' ', $sPhrase);
-                Tokenizer::addTokens($aTokens, $aWords);
-                $aWordLists[] = $aWords;
-            } else {
-                $aWordLists[] = array();
+            $oWordList = new SimpleWordList($sPhrase);
+
+            foreach ($oWordList->getTokens() as $sToken) {
+                $aTokens[' '.$sToken] = ' '.$sToken;
+                $aTokens[$sToken] = $sToken;

             }
             }

+
+            $aWordLists[] = $oWordList;

         }
 
         Debug::printVar('Tokens', $aTokens);
         }
 
         Debug::printVar('Tokens', $aTokens);


@@ -121,7 +150,7 @@ class Tokenizer
         $oValidTokens = $this->computeValidTokens($aTokens, $sNormQuery);
 
         foreach ($aPhrases as $iPhrase => $oPhrase) {
         $oValidTokens = $this->computeValidTokens($aTokens, $sNormQuery);
 
         foreach ($aPhrases as $iPhrase => $oPhrase) {

-            $oPhrase->computeWordSets($aWordLists[$iPhrase], $oValidTokens);
+            $oPhrase->setWordSets($aWordLists[$iPhrase]->getWordSets($oValidTokens));

         }
 
         return $oValidTokens;
         }
 
         return $oValidTokens;


@@ -206,12 +235,7 @@ class Tokenizer
                     );
                 }
             } elseif ($aWord['country_code']) {
                     );
                 }
             } elseif ($aWord['country_code']) {

-                // Filter country tokens that do not match restricted countries.
-                if (!$this->aCountryRestriction
-                    || in_array($aWord['country_code'], $this->aCountryRestriction)
-                ) {
-                    $oToken = new Token\Country($iId, $aWord['country_code']);
-                }
+                $oToken = new Token\Country($iId, $aWord['country_code']);

             } elseif ($aWord['word_token'][0] == ' ') {
                 $oToken = new Token\Word(
                     $iId,
             } elseif ($aWord['word_token'][0] == ' ') {
                 $oToken = new Token\Word(
                     $iId,


@@ -238,29 +262,4 @@ class Tokenizer
             }
         }
     }
             }
         }
     }

-
-
-    /**
-     * Add the tokens from this phrase to the given list of tokens.
-     *
-     * @param string[] $aTokens List of tokens to append.
-     *
-     * @return void
-     */
-    private static function addTokens(&$aTokens, $aWords)
-    {
-        $iNumWords = count($aWords);
-
-        for ($i = 0; $i < $iNumWords; $i++) {
-            $sPhrase = $aWords[$i];
-            $aTokens[' '.$sPhrase] = ' '.$sPhrase;
-            $aTokens[$sPhrase] = $sPhrase;
-
-            for ($j = $i + 1; $j < $iNumWords; $j++) {
-                $sPhrase .= ' '.$aWords[$j];
-                $aTokens[' '.$sPhrase] = ' '.$sPhrase;
-                $aTokens[$sPhrase] = $sPhrase;
-            }
-        }
-    }

 }
 }