safety-screen/Nominatim
1
0
Fork 0
forked from hans/Nominatim
Code Pull Requests Activity

move SearchDescription building into tokens

Browse Source 
Moving the logic for extending the SearchDescription into the
token classes splits up the code and makes it more readable.
More importantly: it allows tokenizer to define custom token
classes in the future.
This commit is contained in:
Sarah Hoffmann
2021-07-17 20:24:33 +02:00
parent 3cd85eaaf1
commit a48ebd9b47
9 changed files with 464 additions and 270 deletions
Download Patch File Download Diff File Expand all files Collapse all files

82
lib-php/TokenPartial.php
View File

@@ -8,19 +8,86 @@ namespace Nominatim\Token;
class Partial
{
/// Database word id, if applicable.
public $iId;
private $iId;
/// Number of appearances in the database.
public $iSearchNameCount;
/// Normalised version of the partial word.
public $sToken;
private $iSearchNameCount;
/// True, if the token consists exclusively of digits and spaces.
private $bNumberToken;
public function __construct($iId, $sToken, $iSearchNameCount)
{
$this->iId = $iId;
$this->sToken = $sToken;
$this->bNumberToken = (bool) preg_match('#^[0-9 ]+$#', $sToken);
$this->iSearchNameCount = $iSearchNameCount;
}
public function getId()
{
return $this->iId;
}
/**
* Derive new searches by adding this token to an existing search.
*
* @param object $oSearch Partial search description derived so far.
* @param object $oPosition Description of the token position within
the query.
*
* @return SearchDescription[] List of derived search descriptions.
*/
public function extendSearch($oSearch, $oPosition)
{
if ($oPosition->isPhrase('country')) {
return array();
}
$aNewSearches = array();
// Partial token in Address.
if (($oPosition->isPhrase('') || !$oPosition->isFirstPhrase())
&& $oSearch->hasName()
) {
$iSearchCost = $this->bNumberToken ? 2 : 1;
if ($this->iSearchNameCount >= CONST_Max_Word_Frequency) {
$iSearchCost += 1;
}
$oNewSearch = $oSearch->clone($iSearchCost);
$oNewSearch->addAddressToken(
$this->iId,
$this->iSearchNameCount < CONST_Max_Word_Frequency
);
$aNewSearches[] = $oNewSearch;
}
// Partial token in Name.
if ((!$oSearch->hasPostcode() && !$oSearch->hasAddress())
&& (!$oSearch->hasName(true)
|| $oSearch->getNamePhrase() == $oPosition->getPhrase())
) {
$iSearchCost = 1;
if (!$oSearch->hasName(true)) {
$iSearchCost += 1;
}
if ($this->bNumberToken) {
$iSearchCost += 1;
}
$oNewSearch = $oSearch->clone($iSearchCost);
$oNewSearch->addPartialNameToken(
$this->iId,
$this->iSearchNameCount < CONST_Max_Word_Frequency,
$oPosition->getPhrase()
);
$aNewSearches[] = $oNewSearch;
}
return $aNewSearches;
}
public function debugInfo()
{
return array(
@@ -31,4 +98,9 @@ class Partial
)
);
}
public function debugCode()
{
return 'w';
}
}