<?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;
require_once(CONST_LibDir.'/ReverseGeocode.php');
require_once(CONST_LibDir.'/SearchDescription.php');
require_once(CONST_LibDir.'/SearchContext.php');
+require_once(CONST_LibDir.'/SearchPosition.php');
require_once(CONST_LibDir.'/TokenList.php');
require_once(CONST_TokenizerDir.'/tokenizer.php');
}
$this->iFinalLimit = $iLimit;
- $this->iLimit = $iLimit + min($iLimit, 10);
+ $this->iLimit = $iLimit + max($iLimit, 10);
}
public function setFeatureType($sFeatureType)
$this->bFallback = $oParams->getBool('fallback', $this->bFallback);
- // List of excluded Place IDs - used for more acurate pageing
+ // List of excluded Place IDs - used for more accurate pageing
$sExcluded = $oParams->getStringList('exclude_place_ids');
if ($sExcluded) {
foreach ($sExcluded as $iExcludedPlaceID) {
*/
foreach ($aPhrases as $iPhrase => $oPhrase) {
$aNewPhraseSearches = array();
- $sPhraseType = $oPhrase->getPhraseType();
+ $oPosition = new SearchPosition(
+ $oPhrase->getPhraseType(),
+ $iPhrase,
+ count($aPhrases)
+ );
foreach ($oPhrase->getWordSets() as $aWordset) {
$aWordsetSearches = $aSearches;
// Add all words from this wordset
foreach ($aWordset as $iToken => $sToken) {
$aNewWordsetSearches = array();
+ $oPosition->setTokenPosition($iToken, count($aWordset));
foreach ($aWordsetSearches as $oCurrentSearch) {
- // Tokens with full name matches.
- foreach ($oValidTokens->get(' '.$sToken) as $oSearchTerm) {
- $aNewSearches = $oCurrentSearch->extendWithFullTerm(
- $oSearchTerm,
- $sPhraseType,
- $iToken == 0 && $iPhrase == 0,
- $iPhrase == 0,
- $iToken + 1 == count($aWordset)
- && $iPhrase + 1 == count($aPhrases)
- );
-
- foreach ($aNewSearches as $oSearch) {
- if ($oSearch->getRank() < $this->iMaxRank) {
- $aNewWordsetSearches[] = $oSearch;
- }
- }
- }
- // Look for partial matches.
- // Note that there is no point in adding country terms here
- // because country is omitted in the address.
- if ($sPhraseType != 'country') {
- // Allow searching for a word - but at extra cost
- foreach ($oValidTokens->get($sToken) as $oSearchTerm) {
- $aNewSearches = $oCurrentSearch->extendWithPartialTerm(
- $sToken,
- $oSearchTerm,
- (bool) $sPhraseType,
- $iPhrase,
- $oValidTokens->get(' '.$sToken)
+ foreach ($oValidTokens->get($sToken) as $oSearchTerm) {
+ if ($oSearchTerm->isExtendable($oCurrentSearch, $oPosition)) {
+ $aNewSearches = $oSearchTerm->extendSearch(
+ $oCurrentSearch,
+ $oPosition
);
foreach ($aNewSearches as $oSearch) {
if ($this->aCountryCodes) {
$oCtx->setCountryList($this->aCountryCodes);
}
- $this->oTokenizer->setCountryRestriction($this->aCountryCodes);
Debug::newSection('Query Preprocessing');
userError('Query string is not UTF-8 encoded.');
}
- // Conflicts between US state abreviations and various words for 'the' in different languages
- if (isset($this->aLangPrefOrder['name:en'])) {
- $sQuery = preg_replace('/(^|,)\s*il\s*(,|$)/i', '\1illinois\2', $sQuery);
- $sQuery = preg_replace('/(^|,)\s*al\s*(,|$)/i', '\1alabama\2', $sQuery);
- $sQuery = preg_replace('/(^|,)\s*la\s*(,|$)/i', '\1louisiana\2', $sQuery);
- }
-
// Do we have anything that looks like a lat/lon pair?
$sQuery = $oCtx->setNearPointFromQuery($sQuery);
if (!empty($aTokens)) {
$aNewSearches = array();
+ $oPosition = new SearchPosition('', 0, 1);
+ $oPosition->setTokenPosition(0, 1);
+
foreach ($aSearches as $oSearch) {
foreach ($aTokens as $oToken) {
- $oNewSearch = clone $oSearch;
- $oNewSearch->setPoiSearch(
- $oToken->iOperator,
- $oToken->sClass,
- $oToken->sType
+ $aNewSearches = array_merge(
+ $aNewSearches,
+ $oToken->extendSearch($oSearch, $oPosition)
);
- $aNewSearches[] = $oNewSearch;
}
}
$aSearches = $aNewSearches;
}
$aReverseGroupedSearches = $this->getGroupedSearches($aSearches, $aPhrases, $oValidTokens);
- foreach ($aGroupedSearches as $aSearches) {
+ foreach ($aReverseGroupedSearches as $aSearches) {
foreach ($aSearches as $aSearch) {
- if (!isset($aReverseGroupedSearches[$aSearch->getRank()])) {
- $aReverseGroupedSearches[$aSearch->getRank()] = array();
+ if (!isset($aGroupedSearches[$aSearch->getRank()])) {
+ $aGroupedSearches[$aSearch->getRank()] = array();
}
- $aReverseGroupedSearches[$aSearch->getRank()][] = $aSearch;
+ $aGroupedSearches[$aSearch->getRank()][] = $aSearch;
}
}
- $aGroupedSearches = $aReverseGroupedSearches;
ksort($aGroupedSearches);
}
} else {
$aResult['importance'] = 0.001;
$aResult['foundorder'] = $aResult['addressimportance'];
} else {
- $aResult['importance'] = max(0.001, $aResult['importance']);
+ if ($aResult['importance'] == 0) {
+ $aResult['importance'] = 0.0001;
+ }
$aResult['importance'] *= $this->viewboxImportanceFactor(
$aResult['lon'],
$aResult['lat']
$iCountWords = 0;
$sAddress = $aResult['langaddress'];
foreach ($aRecheckWords as $i => $sWord) {
- if (stripos($sAddress, $sWord)!==false) {
+ if (grapheme_stripos($sAddress, $sWord)!==false) {
$iCountWords++;
if (preg_match('/(^|,)\s*'.preg_quote($sWord, '/').'\s*(,|$)/', $sAddress)) {
$iCountWords += 0.1;