p33b

src/Config/NdjsonHybridRetrieverConfig.php

@@ -196,6 +196,11 @@ final class NdjsonHybridRetrieverConfig
         return $this->requiredStringList('exact_detail_tokens');
     }
 
+    public function genericExactSelectionCleanupProfile(): string
+    {
+        return $this->requiredString('generic_exact_selection_cleanup_profile');
+    }
+
     /** @return string[] */
     public function genericExactSelectionTokens(): array
     {
@@ -316,6 +321,7 @@ final class NdjsonHybridRetrieverConfig
             'exact_selection_indicator_table_required_primary_terms' => $this->exactSelectionIndicatorTableRequiredPrimaryTerms(),
             'exact_selection_indicator_table_required_context_terms' => $this->exactSelectionIndicatorTableRequiredContextTerms(),
             'exact_detail_tokens' => $this->exactDetailTokens(),
+            'generic_exact_selection_cleanup_profile' => $this->genericExactSelectionCleanupProfile(),
             'generic_exact_selection_tokens' => $this->genericExactSelectionTokens(),
             'generic_product_tokens' => $this->genericProductTokens(),
             'important_short_model_tokens' => $this->importantShortModelTokens(),
@@ -369,6 +375,22 @@ final class NdjsonHybridRetrieverConfig
         return $value;
     }
 
+    private function requiredString(string $key): string
+    {
+        $value = $this->requiredValue($key);
+
+        if (!is_scalar($value)) {
+            throw $this->invalid($key, 'must be a non-empty string');
+        }
+
+        $value = trim((string) $value);
+        if ($value === '') {
+            throw $this->invalid($key, 'must be a non-empty string');
+        }
+
+        return $value;
+    }
+
     /** @return string[] */
     private function requiredStringList(string $key): array
     {

src/Config/RetriexEffectiveConfigProvider.php

@@ -1059,6 +1059,13 @@ final readonly class RetriexEffectiveConfigProvider
             $errors[] = 'retrieval.hard_max_chunks must be greater than 0.';
         }
 
+        $cleanupProfile = $retrieval['generic_exact_selection_cleanup_profile'] ?? null;
+        if (!is_string($cleanupProfile) || trim($cleanupProfile) === '') {
+            $errors[] = 'retrieval.generic_exact_selection_cleanup_profile must be a non-empty string.';
+        } elseif (!in_array(trim($cleanupProfile), $this->languageCleanupConfig->getCleanupProfileNames(), true)) {
+            $errors[] = 'retrieval.generic_exact_selection_cleanup_profile references unknown language cleanup profile: ' . trim($cleanupProfile) . '.';
+        }
+
         $this->validateStringListMap($retrieval['vocabulary'] ?? [], 'retrieval.vocabulary', $errors, $warnings);
 
         $inventory = $retrieval['inventory_parameter'] ?? [];

src/Knowledge/Retrieval/NdjsonHybridRetriever.php

@@ -5,6 +5,7 @@ declare(strict_types=1);
 namespace App\Knowledge\Retrieval;
 
 use App\Catalog\EntityCatalogService;
+use App\Config\LanguageCleanupConfig;
 use App\Config\NdjsonHybridRetrieverConfig;
 use App\Entity\ModelGenerationConfig;
 use App\Intent\CatalogIntentLite;
@@ -46,6 +47,7 @@ final readonly class NdjsonHybridRetriever implements RetrieverInterface
         private EntityCatalogService            $entityCatalogService,
         private QueryEnricher                   $queryEnricher,
         private NdjsonHybridRetrieverConfig     $retrieverConfig,
+        private LanguageCleanupConfig           $languageCleanupConfig,
     )
     {
     }
@@ -942,7 +944,19 @@ final readonly class NdjsonHybridRetriever implements RetrieverInterface
 
     private function isGenericExactSelectionToken(string $token): bool
     {
-        return in_array($token, $this->retrieverConfig->genericExactSelectionTokens(), true);
+        return in_array($token, $this->genericExactSelectionCleanupTokens(), true);
+    }
+
+    /** @return string[] */
+    private function genericExactSelectionCleanupTokens(): array
+    {
+        $profileName = $this->retrieverConfig->genericExactSelectionCleanupProfile();
+
+        return array_values(array_unique(array_merge(
+            $this->languageCleanupConfig->getStopWordsForProfile($profileName),
+            $this->languageCleanupConfig->getMetaTermsForProfile($profileName),
+            $this->retrieverConfig->genericExactSelectionTokens()
+        )));
     }
 
     /**