optimize texts retrieval

src/Intent/CatalogIntentLite.php

@@ -4,6 +4,7 @@ declare(strict_types=1);
 
 namespace App\Intent;
 
+use App\Knowledge\Retrieval\QueryCleaner;
 use App\Tag\TagVectorSearchClient;
 use App\Tag\TagTypes;
 
@@ -29,7 +30,7 @@ final class CatalogIntentLite
      * Minimaler Similarity-Score.
      * Verhindert Rauschen.
      */
-    private const MIN_SCORE = 0.75;
+    private const MIN_SCORE = 0.72;
 
     /**
      * Differenz zwischen Top1 und Top2,
@@ -39,6 +40,7 @@ final class CatalogIntentLite
 
     public function __construct(
         private readonly TagVectorSearchClient $tagVectorClient,
+        private readonly QueryCleaner   $queryCleaner,
     ) {}
 
     /**
@@ -52,8 +54,10 @@ final class CatalogIntentLite
             return null;
         }
 
-        // 1) Vector-Suche
-        $hits = $this->tagVectorClient->search($prompt, 3);
+        $promptTag = $this->queryCleaner->clean($prompt);
+
+        // 1) Tag-Vector-Suche
+        $hits = $this->tagVectorClient->search($promptTag, 3);
 
         if ($hits === []) {
             return null;
@@ -62,7 +66,7 @@ final class CatalogIntentLite
         $best = $hits[0];
         $bestScore = (float)($best['score'] ?? 0.0);
 
-        // 2) Score-Gate
+        // 2) Score-Tags
         if ($bestScore < self::MIN_SCORE) {
             return null;
         }

src/Service/Admin/ModelGenerationConfigAdminService.php

@@ -74,7 +74,7 @@ final class ModelGenerationConfigAdminService
             return [];
         }
 
-        return $this->retriever->retrieveInternal($prompt, $config);
+        return $this->retriever->retrieveDebug($prompt);
     }
 
     private function requireString(mixed $value, string $field): string

templates/admin/model_config/test_retrieval.html.twig

@@ -77,13 +77,37 @@
                 </h5>
 
                 <div style="max-height: 500px; overflow-y: auto;">
-
                     {% for chunk in results %}
                         <div class="border border-secondary p-3 mb-3 small">
-                            {{ chunk|nl2br }}
+
+                            {# ================= META-ZEILE ================= #}
+                            <div class="mb-2 text-warning" style="font-size: 11px; line-height: 1.4;">
+                                <span class="text-info"><strong>rank:</strong> {{ chunk.rank }}</span> |
+                                <span class="text-info"><strong>chunk_id:</strong> {{ chunk.chunk_id }}</span> |
+                                <span class="text-info"><strong>document_id:</strong> <a
+                                            class="text-info"
+                                            href="{{ path('admin_document_show', { id: chunk.document_id }) }}">
+                                {{ chunk.document_id }}
+                                </a></span> |
+                                <span><strong>raw_score:</strong> {{ chunk.raw_score|number_format(6, '.', '') }}</span>
+                                |
+                                <span><strong>rrf_score:</strong> {{ chunk.rrf_score|number_format(6, '.', '') }}</span>
+                                |
+                                <span><strong>threshold:</strong> {{ chunk.threshold }}</span> |
+                                <span><strong>intent:</strong> {{ chunk.intent }}</span> |
+                                <span>
+                    <strong>is_list_query:</strong>
+                    {{ chunk.is_list_query ? 'true' : 'false' }}
+                </span>
+                            </div>
+
+                            {# ================= CHUNK TEXT ================= #}
+                            <div>
+                                {{ chunk.text }}
+                            </div>
+
                         </div>
                     {% endfor %}
-
                 </div>
 
             </div>