remove direct chnuk search. only vector search

src/Knowledge/QueryCleaner.php

@@ -0,0 +1,39 @@
+<?php
+
+declare(strict_types=1);
+
+namespace App\Knowledge;
+
+final readonly class QueryCleaner
+{
+    public function __construct(
+        private StopWords $stopWords
+    ) {
+    }
+
+    public function clean(string $query): string
+    {
+        $query = mb_strtolower($query);
+        $query = preg_replace('/[^\p{L}\p{N}\s]/u', ' ', $query);
+        $query = preg_replace('/\s+/u', ' ', $query);
+        $query = trim($query);
+
+        if ($query === '') {
+            return '';
+        }
+
+        $tokens = explode(' ', $query);
+        $stopWords = $this->stopWords->getStopWords(); // <-- wichtig: nutzt deine Klasse
+
+        $filtered = array_filter(
+            $tokens,
+            function (string $word) use ($stopWords): bool {
+                return $word !== ''
+                    && mb_strlen($word) > 2
+                    && !in_array($word, $stopWords, true);
+            }
+        );
+
+        return implode(' ', $filtered);
+    }
+}

src/Knowledge/Retrieval/NdjsonHybridRetriever.php

@@ -4,6 +4,7 @@ declare(strict_types=1);
 
 namespace App\Knowledge\Retrieval;
 
+use App\Knowledge\QueryCleaner;
 use App\Vector\VectorSearchClient;
 
 final class NdjsonHybridRetriever implements RetrieverInterface
@@ -11,10 +12,11 @@ final class NdjsonHybridRetriever implements RetrieverInterface
     private const VECTOR_SCORE_THRESHOLD = 0.25;
 
     public function __construct(
-        private readonly NdjsonChunkLookup   $lookup,
+        private readonly NdjsonChunkLookup  $lookup,
-        private readonly VectorSearchClient  $vectorClient,
+        private readonly VectorSearchClient $vectorClient,
-        private readonly int                 $maxChunks = 10,
+        private readonly QueryCleaner       $queryCleaner,
-        private readonly int                 $vectorTopK = 10,
+        private readonly int                $maxChunks = 25,
+        private readonly int                $vectorTopK = 10,
     )
     {
     }
@@ -23,9 +25,10 @@ final class NdjsonHybridRetriever implements RetrieverInterface
     {
         $limit = $this->maxChunks;
         $keywordChunks = [];
+        $query = $this->queryCleaner->clean($prompt);
 
         // Vector / enrichment
-        $hits = $this->vectorClient->search($prompt, $this->vectorTopK);
+        $hits = $this->vectorClient->search($query, $this->vectorTopK);
         if ($hits === []) {
             return $this->diversifyByDevice($keywordChunks, $limit, 1);
         }