Merge remote-tracking branch 'upstream/master'

[nominatim.git] / lib / Geocode.php

diff --git a/lib/Geocode.php b/lib/Geocode.php
index af3262f9e7e866780be904fa7b870aed52471b33..12f9da37550f856e47a128641a275813d9c4e9de 100644 (file)
--- a/lib/Geocode.php
+++ b/lib/Geocode.php
@@ -2,12 +2,12 @@
 
 namespace Nominatim;
 
 
 namespace Nominatim;
 

-require_once(CONST_BasePath.'/lib/PlaceLookup.php');
-require_once(CONST_BasePath.'/lib/Phrase.php');
-require_once(CONST_BasePath.'/lib/ReverseGeocode.php');
-require_once(CONST_BasePath.'/lib/SearchDescription.php');
-require_once(CONST_BasePath.'/lib/SearchContext.php');
-require_once(CONST_BasePath.'/lib/TokenList.php');
+require_once(CONST_LibDir.'/PlaceLookup.php');
+require_once(CONST_LibDir.'/Phrase.php');
+require_once(CONST_LibDir.'/ReverseGeocode.php');
+require_once(CONST_LibDir.'/SearchDescription.php');
+require_once(CONST_LibDir.'/SearchContext.php');
+require_once(CONST_LibDir.'/TokenList.php');

 
 class Geocode
 {
 
 class Geocode
 {


@@ -650,6 +650,8 @@ class Geocode
                     $this->oNormalizer
                 );
 
                     $this->oNormalizer
                 );
 

+                $oCtx->setFullNameWords($oValidTokens->getFullWordIDs());
+

                 // Try more interpretations for Tokens that could not be matched.
                 foreach ($aTokens as $sToken) {
                     if ($sToken[0] == ' ' && !$oValidTokens->contains($sToken)) {
                 // Try more interpretations for Tokens that could not be matched.
                 foreach ($aTokens as $sToken) {
                     if ($sToken[0] == ' ' && !$oValidTokens->contains($sToken)) {


@@ -770,7 +772,7 @@ class Geocode
                         }
                     }
 
                         }
                     }
 

-                    //if ($iQueryLoop > 20) break;
+                    if ($iQueryLoop > 20) break;

                 }
 
                 if (!empty($aResults)) {
                 }
 
                 if (!empty($aResults)) {


@@ -842,8 +844,8 @@ class Geocode
                 }
 
                 if (!empty($aResults)) break;
                 }
 
                 if (!empty($aResults)) break;

-                //if ($iGroupLoop > 4) break;
-                //if ($iQueryLoop > 30) break;
+                if ($iGroupLoop > 4) break;
+                if ($iQueryLoop > 30) break;

             }
         } else {
             // Just interpret as a reverse geocode
             }
         } else {
             // Just interpret as a reverse geocode


@@ -917,21 +919,14 @@ class Geocode
                     $aResult['lon'],
                     $aResult['lat']
                 );
                     $aResult['lon'],
                     $aResult['lat']
                 );

-                // Adjust importance for the number of exact string matches in the result
-                $iCountWords = 0;
-                $sAddress = $aResult['langaddress'];
-                foreach ($aRecheckWords as $i => $sWord) {
-                    if (stripos($sAddress, $sWord)!==false) {
-                        $iCountWords++;
-                        if (preg_match('/(^|,)\s*'.preg_quote($sWord, '/').'\s*(,|$)/', $sAddress)) $iCountWords += 0.1;
-                    }
-                }
-
-                $aResult['importance'] = $aResult['importance'] + ($iCountWords*0.1); // 0.1 is a completely arbitrary number but something in the range 0.1 to 0.5 would seem right

 
                 // secondary ordering (for results with same importance (the smaller the better):
                 // - approximate importance of address parts
 
                 // secondary ordering (for results with same importance (the smaller the better):
                 // - approximate importance of address parts

-                $aResult['foundorder'] = -$aResult['addressimportance']/10;
+                if (isset($aResult['addressimportance']) && $aResult['addressimportance']) {
+                    $aResult['foundorder'] = -$aResult['addressimportance']/10;
+                } else {
+                    $aResult['foundorder'] = -$aResult['importance'];
+                }

                 // - number of exact matches from the query
                 $aResult['foundorder'] -= $aResults[$aResult['place_id']]->iExactMatches;
                 // - importance of the class/type
                 // - number of exact matches from the query
                 $aResult['foundorder'] -= $aResults[$aResult['place_id']]->iExactMatches;
                 // - importance of the class/type


@@ -941,6 +936,21 @@ class Geocode
                 } else {
                     $aResult['foundorder'] += 0.01;
                 }
                 } else {
                     $aResult['foundorder'] += 0.01;
                 }

+                // - rank
+                $aResult['foundorder'] -= 0.00001 * (30 - $aResult['rank_search']);
+
+                // Adjust importance for the number of exact string matches in the result
+                $iCountWords = 0;
+                $sAddress = $aResult['langaddress'];
+                foreach ($aRecheckWords as $i => $sWord) {
+                    if (stripos($sAddress, $sWord)!==false) {
+                        $iCountWords++;
+                        if (preg_match('/(^|,)\s*'.preg_quote($sWord, '/').'\s*(,|$)/', $sAddress)) $iCountWords += 0.1;
+                    }
+                }
+
+                // 0.1 is a completely arbitrary number but something in the range 0.1 to 0.5 would seem right
+                $aResult['importance'] = $aResult['importance'] + ($iCountWords*0.1);

             }
             $aSearchResults[$iIdx] = $aResult;
         }
             }
             $aSearchResults[$iIdx] = $aResult;
         }