move intent an config value into config files
This commit is contained in:
40
src/Config/StopWordsConfig.php
Normal file
40
src/Config/StopWordsConfig.php
Normal file
@@ -0,0 +1,40 @@
|
||||
<?php
|
||||
|
||||
declare(strict_types=1);
|
||||
|
||||
namespace App\Config;
|
||||
|
||||
final class StopWordsConfig
|
||||
{
|
||||
/**
|
||||
* Retrieval-optimized stop-word list.
|
||||
*
|
||||
* Important:
|
||||
* - keep negations
|
||||
* - keep question words
|
||||
* - keep domain terms
|
||||
* - remove only structural filler words
|
||||
*
|
||||
* @return string[]
|
||||
*/
|
||||
public function getStopWords(): array
|
||||
{
|
||||
return [
|
||||
'mit',
|
||||
'der', 'die', 'das',
|
||||
'ein', 'eine', 'einer', 'eines',
|
||||
'den', 'dem', 'des',
|
||||
'und', 'oder', 'aber', 'sowie',
|
||||
'ich', 'du', 'er', 'sie', 'es',
|
||||
'wir', 'ihr',
|
||||
'halt', 'eben', 'auch', 'schon',
|
||||
'noch', 'mal', 'bitte', 'danke',
|
||||
'also', 'nun', 'tja',
|
||||
'dann', 'danach', 'davor',
|
||||
'hier', 'dort',
|
||||
'heute', 'gestern', 'morgen',
|
||||
'könnte', 'kannst', 'kann',
|
||||
'würde', 'würdest', 'würden',
|
||||
];
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user