X-Git-Url: https://git.openstreetmap.org./nominatim.git/blobdiff_plain/6ad35aca4aadbf436e9dce776134f9ce4d2e69c6..0fb8eade136ea03e7853aca0795ca69833c33661:/lib-php/tokenizer/legacy_icu_tokenizer.php

diff --git a/lib-php/tokenizer/legacy_icu_tokenizer.php b/lib-php/tokenizer/legacy_icu_tokenizer.php
index 796635ee..690ef136 100644
--- a/lib-php/tokenizer/legacy_icu_tokenizer.php
+++ b/lib-php/tokenizer/legacy_icu_tokenizer.php
@@ -8,7 +8,6 @@ class Tokenizer
 
     private $oNormalizer;
     private $oTransliterator;
-    private $aCountryRestriction;
 
     public function __construct(&$oDB)
     {
@@ -19,23 +18,17 @@ class Tokenizer
 
     public function checkStatus()
     {
-        $sSQL = "SELECT word_id FROM word WHERE word_token == 'a'";
+        $sSQL = 'SELECT word_id FROM word WHERE word_id is not null limit 1';
         $iWordID = $this->oDB->getOne($sSQL);
         if ($iWordID === false) {
-            throw new Exception('Query failed', 703);
+            throw new \Exception('Query failed', 703);
         }
         if (!$iWordID) {
-            throw new Exception('No value', 704);
+            throw new \Exception('No value', 704);
         }
     }
 
 
-    public function setCountryRestriction($aCountries)
-    {
-        $this->aCountryRestriction = $aCountries;
-    }
-
-
     public function normalizeString($sTerm)
     {
         if ($this->oNormalizer === null) {
@@ -145,10 +138,9 @@ class Tokenizer
     private function addTokensFromDB(&$oValidTokens, $aTokens, $sNormQuery)
     {
         // Check which tokens we have, get the ID numbers
-        $sSQL = 'SELECT word_id, word_token, type';
-        $sSQL .= "      info->>'cc' as country, info->>'postcode' as postcode,";
+        $sSQL = 'SELECT word_id, word_token, type, word,';
         $sSQL .= "      info->>'op' as operator,";
-        $sSQL .= "      info->>'class' as class, info->>'type' as type,";
+        $sSQL .= "      info->>'class' as class, info->>'type' as ctype,";
         $sSQL .= "      info->>'count' as count";
         $sSQL .= ' FROM word WHERE word_token in (';
         $sSQL .= join(',', $this->oDB->getDBQuotedList($aTokens)).')';
@@ -159,66 +151,63 @@ class Tokenizer
 
         foreach ($aDBWords as $aWord) {
             $iId = (int) $aWord['word_id'];
+            $sTok = $aWord['word_token'];
 
             switch ($aWord['type']) {
-                'C':  // country name tokens
-                    if ($aWord['country'] === null
-                        || ($this->aCountryRestriction
-                            && !in_array($aWord['country'], $this->aCountryRestriction))
-                    ) {
-                        continue;
+                case 'C':  // country name tokens
+                    if ($aWord['word'] !== null) {
+                        $oValidTokens->addToken(
+                            $sTok,
+                            new Token\Country($iId, $aWord['word'])
+                        );
                     }
-                    $oToken = new Token\Country($iId, $aWord['country'])
                     break;
-                'H':  // house number tokens
-                    $oToken = new Token\HouseNumber($iId, $aWord['word_token']);
+                case 'H':  // house number tokens
+                    $oValidTokens->addToken($sTok, new Token\HouseNumber($iId, $aWord['word_token']));
                     break;
-                'P':  // postcode tokens
+                case 'P':  // postcode tokens
                     // Postcodes are not normalized, so they may have content
                     // that makes SQL injection possible. Reject postcodes
                     // that would need special escaping.
-                    if ($aWord['postcode'] === null
-                        || pg_escape_string($aWord['postcode']) == $aWord['postcode']
+                    if ($aWord['word'] !== null
+                        && pg_escape_string($aWord['word']) == $aWord['word']
                     ) {
-                       continue;
-                    }
-                    $sNormPostcode = $this->normalizeString($aWord['postcode']);
-                    if (strpos($sNormQuery, $sNormPostcode) === false) {
-                        continue;
+                        $sNormPostcode = $this->normalizeString($aWord['word']);
+                        if (strpos($sNormQuery, $sNormPostcode) !== false) {
+                            $oValidTokens->addToken(
+                                $sTok,
+                                new Token\Postcode($iId, $aWord['word'], null)
+                            );
+                        }
                     }
-                    $oToken = new Token\Postcode($iId, $aWord['postcode'], null);
                     break;
-                'S':  // tokens for classification terms (special phrases)
-                    if ($aWord['class'] === null || $aWord['type'] === null
-                    ) {
-                        continue;
+                case 'S':  // tokens for classification terms (special phrases)
+                    if ($aWord['class'] !== null && $aWord['ctype'] !== null) {
+                        $oValidTokens->addToken($sTok, new Token\SpecialTerm(
+                            $iId,
+                            $aWord['class'],
+                            $aWord['ctype'],
+                            (isset($aWord['operator'])) ? Operator::NEAR : Operator::NONE
+                        ));
                     }
-                    $oToken = new Token\SpecialTerm(
-                        $iId,
-                        $aWord['class'],
-                        $aWord['type'],
-                        $aWord['op'] ? Operator::NEAR : Operator::NONE
-                    );
                     break;
-                'W': // full-word tokens
-                    $oToken = new Token\Word(
+                case 'W': // full-word tokens
+                    $oValidTokens->addToken($sTok, new Token\Word(
                         $iId,
                         (int) $aWord['count'],
                         substr_count($aWord['word_token'], ' ')
-                    );
+                    ));
                     break;
-                'w':  // partial word terms
-                    $oToken = new Token\Partial(
+                case 'w':  // partial word terms
+                    $oValidTokens->addToken($sTok, new Token\Partial(
                         $iId,
                         $aWord['word_token'],
                         (int) $aWord['count']
-                    );
+                    ));
                     break;
                 default:
-                    continue;
+                    break;
             }
-
-            $oValidTokens->addToken($aWord['word_token'], $oToken);
         }
     }