second step

src/Config/StopWordsConfig.php

@@ -4,39 +4,18 @@ declare(strict_types=1);
 
 namespace App\Config;
 
+/**
+ * YAML-backed stop-word configuration.
+ *
+ * This class intentionally has no PHP fallback list. The complete list lives in
+ * config/retriex/language.yaml.
+ */
 final class StopWordsConfig
 {
-    /**
-     * Retrieval-optimized stop-word list.
-     *
-     * Important:
-     * - keep negations
-     * - keep question words
-     * - keep domain terms
-     * - remove only structural filler words
-     */
-    private const DEFAULT_STOP_WORDS = [
-        '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',
-    ];
-
     /**
      * @param array<string, mixed> $config
      */
-    public function __construct(private array $config = [])
+    public function __construct(private readonly array $config)
     {
     }
 
@@ -45,19 +24,22 @@ final class StopWordsConfig
      */
     public function getStopWords(): array
     {
-        return $this->stringList('words', self::DEFAULT_STOP_WORDS);
+        return $this->requiredStringList('words');
     }
 
     /**
-     * @param string[] $default
      * @return string[]
      */
-    private function stringList(string $key, array $default): array
+    private function requiredStringList(string $key): array
     {
-        $value = $this->config[$key] ?? $default;
+        if (!array_key_exists($key, $this->config)) {
+            throw new \InvalidArgumentException(sprintf('Missing required RetrieX stopwords config key "%s".', $key));
+        }
+
+        $value = $this->config[$key];
 
         if (!is_array($value)) {
-            return $default;
+            throw new \InvalidArgumentException(sprintf('RetrieX stopwords config key "%s" must be a list.', $key));
         }
 
         $out = [];
@@ -76,6 +58,10 @@ final class StopWordsConfig
             }
         }
 
-        return $out !== [] ? $out : $default;
+        if ($out === []) {
+            throw new \InvalidArgumentException(sprintf('RetrieX stopwords config key "%s" must not be empty.', $key));
+        }
+
+        return $out;
     }
 }