fix parameters for TokenWord creation

[nominatim.git] / 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 8cff6f322410366d2e0ca2ceaf143d2b2035ce64..eac964e43cd314ffeaaf32b906b585a358b2e758 100644 (file)
--- a/lib-php/tokenizer/legacy_icu_tokenizer.php
+++ b/lib-php/tokenizer/legacy_icu_tokenizer.php
@@ -120,14 +120,14 @@ class Tokenizer
 
             // Try more interpretations for Tokens that could not be matched.
             foreach ($aTokens as $sToken) {
-                if ($sToken[0] == ' ' && !$oValidTokens->contains($sToken)) {
-                    if (preg_match('/^ ([0-9]{5}) [0-9]{4}$/', $sToken, $aData)) {
+                if ($sToken[0] != ' ' && !$oValidTokens->contains($sToken)) {
+                    if (preg_match('/^([0-9]{5}) [0-9]{4}$/', $sToken, $aData)) {
                         // US ZIP+4 codes - merge in the 5-digit ZIP code
                         $oValidTokens->addToken(
                             $sToken,
                             new Token\Postcode(null, $aData[1], 'us')
                         );
-                    } elseif (preg_match('/^ [0-9]+$/', $sToken)) {
+                    } elseif (preg_match('/^[0-9]+$/', $sToken)) {
                         // Unknown single word token with a number.
                         // Assume it is a house number.
                         $oValidTokens->addToken(
@@ -198,13 +198,13 @@ class Tokenizer
             } elseif ($aWord['word_token'][0] == ' ') {
                  $oToken = new Token\Word(
                      $iId,
-                     $aWord['word_token'][0] != ' ',
                      (int) $aWord['count'],
                      substr_count($aWord['word_token'], ' ')
                  );
             } else {
                 $oToken = new Token\Partial(
                     $iId,
+                    $aWord['word_token'],
                     (int) $aWord['count']
                 );
             }