rm CachedRetriever.php

add second shopsearch

src/Commerce/CommerceQueryParser.php

@@ -26,35 +26,43 @@ final readonly class CommerceQueryParser
         string $historyContext = ''
     ): CommerceSearchQuery {
         $normalizedPrompt = $this->normalize($originalPrompt);
+        $isDirectProductQuery = $this->isDirectProductQuery($normalizedPrompt);
 
         [$priceMin, $priceMax] = $this->extractPriceRange($normalizedPrompt);
         $sizes = $this->extractSizes($normalizedPrompt);
         $brand = $this->extractBrand($normalizedPrompt);
 
         $searchText = $this->buildSearchText(
-            $normalizedPrompt,
-            $sizes,
-            $brand,
-            $priceMin,
-            $priceMax
+            prompt: $normalizedPrompt,
+            sizes: $sizes,
+            brand: $brand,
+            priceMin: $priceMin,
+            priceMax: $priceMax,
+            preserveDirectProductQuery: $isDirectProductQuery
         );
 
-        if ($historyContext !== '' && $this->shouldUseHistoryContext($normalizedPrompt)) {
+        if (
+            !$isDirectProductQuery
+            && $historyContext !== ''
+            && $this->shouldUseHistoryContext($normalizedPrompt)
+        ) {
             $latestHistoryQuestion = $this->extractLatestQuestionFromHistory($historyContext);
 
             if ($latestHistoryQuestion !== '') {
                 $normalizedHistoryPrompt = $this->normalize($latestHistoryQuestion);
+                $isDirectHistoryProductQuery = $this->isDirectProductQuery($normalizedHistoryPrompt);
 
                 [$historyPriceMin, $historyPriceMax] = $this->extractPriceRange($normalizedHistoryPrompt);
                 $historySizes = $this->extractSizes($normalizedHistoryPrompt);
                 $historyBrand = $this->extractBrand($normalizedHistoryPrompt);
 
                 $historySearchText = $this->buildSearchText(
-                    $normalizedHistoryPrompt,
-                    $historySizes,
-                    $historyBrand,
-                    $historyPriceMin,
-                    $historyPriceMax
+                    prompt: $normalizedHistoryPrompt,
+                    sizes: $historySizes,
+                    brand: $historyBrand,
+                    priceMin: $historyPriceMin,
+                    priceMax: $historyPriceMax,
+                    preserveDirectProductQuery: $isDirectHistoryProductQuery
                 );
 
                 $searchText = $this->mergeSearchTexts($historySearchText, $searchText);
@@ -98,7 +106,7 @@ final readonly class CommerceQueryParser
      */
     private function extractPriceRange(string $prompt): array
     {
-        $priceMin = 0;
+        $priceMin = null;
         $priceMax = null;
 
         if (preg_match('/\bzwischen\s+(\d+(?:[.,]\d+)?)\s+und\s+(\d+(?:[.,]\d+)?)\s+euro\b/u', $prompt, $m) === 1) {
@@ -165,19 +173,34 @@ final readonly class CommerceQueryParser
         array $sizes,
         ?string $brand,
         ?float $priceMin,
-        ?float $priceMax
+        ?float $priceMax,
+        bool $preserveDirectProductQuery = false
     ): string {
+        if ($preserveDirectProductQuery) {
+            return $this->buildDirectProductSearchText($prompt);
+        }
+
         $text = ' ' . $prompt . ' ';
 
         foreach ($this->config->getPhrasesToRemove() as $phrase) {
-            $text = str_replace($phrase, ' ', $text);
+            $normalizedPhrase = $this->normalize((string) $phrase);
+            if ($normalizedPhrase === '') {
+                continue;
+            }
+
+            $text = str_replace(' ' . $normalizedPhrase . ' ', ' ', $text);
         }
 
         foreach ($sizes as $size) {
-            $text = preg_replace('/\b' . preg_quote($size, '/') . '\b/u', ' ', $text) ?? $text;
+            $normalizedSize = $this->normalize((string) $size);
+            if ($normalizedSize === '') {
+                continue;
+            }
+
+            $text = preg_replace('/\b' . preg_quote($normalizedSize, '/') . '\b/u', ' ', $text) ?? $text;
         }
 
-        if ($brand !== null && $brand !== '') {
+        if ($brand !== null && $brand !== '' && !$this->isBrandPartOfModelPhrase($prompt, $brand)) {
             $text = preg_replace('/\b' . preg_quote($brand, '/') . '\b/u', ' ', $text) ?? $text;
         }
 
@@ -200,6 +223,22 @@ final readonly class CommerceQueryParser
         return trim(implode(' ', $tokens));
     }
 
+    private function buildDirectProductSearchText(string $prompt): string
+    {
+        $text = $prompt;
+        $text = preg_replace('/\s+/u', ' ', $text) ?? $text;
+        $text = trim($text, " \t\n\r\0\x0B-.,");
+
+        $tokens = array_filter(
+            explode(' ', $text),
+            static fn(string $token): bool => mb_strlen($token) > 0
+        );
+
+        $tokens = array_values(array_unique($tokens));
+
+        return trim(implode(' ', $tokens));
+    }
+
     private function shouldUseHistoryContext(string $prompt): bool
     {
         return preg_match(
@@ -263,6 +302,57 @@ final readonly class CommerceQueryParser
         ));
     }
 
+    private function isDirectProductQuery(string $prompt): bool
+    {
+        if ($prompt === '') {
+            return false;
+        }
+
+        if ($this->containsModelLikePhrase($prompt)) {
+            return true;
+        }
+
+        if ($this->containsAccessoryLikePhrase($prompt)) {
+            return true;
+        }
+
+        $tokens = preg_split('/\s+/u', $prompt, -1, PREG_SPLIT_NO_EMPTY) ?: [];
+
+        if (count($tokens) <= 4 && preg_match('/\d/u', $prompt) === 1) {
+            return true;
+        }
+
+        return false;
+    }
+
+    private function containsModelLikePhrase(string $text): bool
+    {
+        return preg_match(
+                '/\b[a-zäöüß][a-zäöüß®\-]*(?:\s+[a-zäöüß][a-zäöüß®\-]*){0,2}\s+\d{2,5}[a-z0-9\-]*\b/u',
+                $text
+            ) === 1;
+    }
+
+    private function containsAccessoryLikePhrase(string $text): bool
+    {
+        return preg_match(
+                '/\b(?:indikator|indicator|reagenz|reagent|kit|set)\s+\d{1,5}[a-z0-9\-]*\b/u',
+                $text
+            ) === 1;
+    }
+
+    private function isBrandPartOfModelPhrase(string $prompt, string $brand): bool
+    {
+        if ($brand === '') {
+            return false;
+        }
+
+        return preg_match(
+                '/\b' . preg_quote($brand, '/') . '\s+\d{2,5}[a-z0-9\-]*\b/u',
+                $prompt
+            ) === 1;
+    }
+
     private function toFloat(string $value): ?float
     {
         $value = str_replace(',', '.', trim($value));