optimize cleanup search query shop api

src/Commerce/CommerceQueryParser.php

@@ -88,6 +88,7 @@ final readonly class CommerceQueryParser
         );
         $value = preg_replace($this->config->getPromptSanitizePattern(), ' ', $value) ?? $value;
         $value = preg_replace($this->config->getWhitespaceCollapsePattern(), ' ', $value) ?? $value;
+        $value = $this->applySearchTokenCorrections($value);
 
         return trim($value);
     }
@@ -222,7 +223,7 @@ final readonly class CommerceQueryParser
             fn(string $token): bool => mb_strlen($token) > $this->config->getMinSearchTokenLength()
         );
 
-        $tokens = $this->filterSearchTokens($tokens);
+        $tokens = $this->normalizeSearchTokens($tokens);
 
         return trim(implode(' ', $tokens));
     }
@@ -237,8 +238,7 @@ final readonly class CommerceQueryParser
             fn(string $token): bool => mb_strlen($token) >= $this->config->getMinDirectProductTokenLength()
         );
 
-        $tokens = $this->filterSearchTokens($tokens);
+        $tokens = $this->normalizeSearchTokens($tokens);
-        $tokens = array_values(array_unique($tokens));
 
         return trim(implode(' ', $tokens));
     }
@@ -282,11 +282,9 @@ final readonly class CommerceQueryParser
                     continue;
                 }
 
-                if ($this->isSearchControlToken($token)) {
+                foreach ($this->normalizeSearchTokens([$token]) as $normalizedToken) {
-                    continue;
+                    $tokens[$normalizedToken] = $normalizedToken;
                 }
-
-                $tokens[$token] = $token;
             }
         }
 
@@ -299,10 +297,52 @@ final readonly class CommerceQueryParser
      */
     private function filterSearchTokens(array $tokens): array
     {
-        return array_values(array_filter(
+        return $this->normalizeSearchTokens($tokens);
-            $tokens,
+    }
-            fn(string $token): bool => !$this->isSearchControlToken($token)
+
-        ));
+    /**
+     * @param string[] $tokens
+     * @return string[]
+     */
+    private function normalizeSearchTokens(array $tokens): array
+    {
+        $normalizedTokens = [];
+
+        foreach ($tokens as $token) {
+            $token = trim(mb_strtolower((string) $token, 'UTF-8'));
+
+            if ($token === '') {
+                continue;
+            }
+
+            $token = $this->config->getSearchTokenCorrections()[$token] ?? $token;
+            $token = $this->config->getSearchTokenCanonicalMap()[$token] ?? $token;
+
+            if ($this->isSearchControlToken($token)) {
+                continue;
+            }
+
+            $normalizedTokens[$token] = $token;
+        }
+
+        return array_values($normalizedTokens);
+    }
+
+    private function applySearchTokenCorrections(string $text): string
+    {
+        if ($text === '') {
+            return '';
+        }
+
+        foreach ($this->config->getSearchTokenCorrections() as $from => $to) {
+            $text = preg_replace(
+                '/\b' . preg_quote((string) $from, '/') . '\b/u',
+                (string) $to,
+                $text
+            ) ?? $text;
+        }
+
+        return preg_replace($this->config->getWhitespaceCollapsePattern(), ' ', $text) ?? $text;
     }
 
     private function isSearchControlToken(string $token): bool

src/Config/CommerceQueryParserConfig.php

@@ -28,6 +28,8 @@ final class CommerceQueryParserConfig
             'suche',
             'habt ihr',
             'gibt es',
+            'gebe mir',
+            'gib mir',
             'zeige mir',
             'welches gerät',
             'welche gerät',
@@ -36,6 +38,8 @@ final class CommerceQueryParserConfig
             'welches ist am besten',
             'alternative',
             'alternativen',
+            'unter anderem',
+            'u a',
             'welche',
             'welcher',
             'welches',
@@ -100,6 +104,66 @@ final class CommerceQueryParserConfig
             'verfuegbarkeit',
             'prüfe',
             'pruefe',
+            'den',
+            'die',
+            'das',
+            'der',
+            'dem',
+            'des',
+            'und',
+            'oder',
+            'sowie',
+            'seine',
+            'seinen',
+            'seiner',
+            'seinem',
+            'seines',
+            'siene',
+            'sienen',
+            'siener',
+            'sienem',
+            'sienes',
+            'gebe',
+            'gib',
+            'nenne',
+            'nenn',
+            'preis',
+            'preise',
+            'preisen',
+            'kostet',
+            'kosten',
+            'ua',
+        ];
+    }
+
+    /**
+     * @return array<string, string>
+     */
+    public function getSearchTokenCorrections(): array
+    {
+        return [
+            'siene' => 'seine',
+            'sienen' => 'seinen',
+            'siener' => 'seiner',
+            'sienem' => 'seinem',
+            'sienes' => 'seines',
+            'indicatoren' => 'indikatoren',
+        ];
+    }
+
+    /**
+     * @return array<string, string>
+     */
+    public function getSearchTokenCanonicalMap(): array
+    {
+        return [
+            'indikatoren' => 'indikator',
+            'indicators' => 'indikator',
+            'indicator' => 'indikator',
+            'reagenzien' => 'reagenz',
+            'reagents' => 'reagenz',
+            'reagent' => 'reagenz',
+            'produkte' => 'produkt',
         ];
     }