From 3d68828201467d99b007065f5db68bbdef6c0d5c Mon Sep 17 00:00:00 2001 From: Sarah Hoffmann Date: Thu, 1 May 2014 17:53:13 +0200 Subject: [PATCH] experimental tweaking of ranks of searches - completely removes partial words with spaces from addresses - ensure that frequent partial search terms appear in NonAddress - higher penalty for partial words consisting of numbers --- lib/Geocode.php | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) diff --git a/lib/Geocode.php b/lib/Geocode.php index dd75b3a7..8616be13 100644 --- a/lib/Geocode.php +++ b/lib/Geocode.php @@ -960,7 +960,7 @@ { if (isset($aSearchTerm['word_id']) && $aSearchTerm['word_id']) { - if ((!$bStructuredPhrases || $iPhrase > 0) && sizeof($aCurrentSearch['aName']) && strlen($sToken) >= 4) + if ((!$bStructuredPhrases || $iPhrase > 0) && sizeof($aCurrentSearch['aName']) && strpos($sToken, ' ') === false) { $aSearch = $aCurrentSearch; $aSearch['iSearchRank'] += 1; @@ -969,8 +969,11 @@ $aSearch['aAddress'][$aSearchTerm['word_id']] = $aSearchTerm['word_id']; if ($aSearch['iSearchRank'] < $this->iMaxRank) $aNewWordsetSearches[] = $aSearch; } - elseif (isset($aValidTokens[' '.$sToken])) // revert to the token version? + elseif (isset($aValidTokens[' '.$sToken]) && strlen($sToken) >= 4) // revert to the token version? { + $aSearch['aAddressNonSearch'][$aSearchTerm['word_id']] = $aSearchTerm['word_id']; + $aSearch['iSearchRank'] += 1; + if ($aSearch['iSearchRank'] < $this->iMaxRank) $aNewWordsetSearches[] = $aSearch; foreach($aValidTokens[' '.$sToken] as $aSearchTermToken) { if (empty($aSearchTermToken['country_code']) @@ -987,6 +990,7 @@ else { $aSearch['aAddressNonSearch'][$aSearchTerm['word_id']] = $aSearchTerm['word_id']; + if (preg_match('#^[0-9]+$#', $sToken)) $aSearch['iSearchRank'] += 2; if ($aSearch['iSearchRank'] < $this->iMaxRank) $aNewWordsetSearches[] = $aSearch; } } @@ -994,7 +998,8 @@ if (!sizeof($aCurrentSearch['aName']) || $aCurrentSearch['iNamePhrase'] == $iPhrase) { $aSearch = $aCurrentSearch; - $aSearch['iSearchRank'] += 2; + $aSearch['iSearchRank'] += 1; + if (!sizeof($aCurrentSearch['aName'])) $aSearch['iSearchRank'] += 1; if (preg_match('#^[0-9]+$#', $sToken)) $aSearch['iSearchRank'] += 2; if ($aWordFrequencyScores[$aSearchTerm['word_id']] < CONST_Max_Word_Frequency) $aSearch['aName'][$aSearchTerm['word_id']] = $aSearchTerm['word_id']; -- 2.39.5