Use scaled-down Lamoda images

app.py

@@ -95,21 +95,56 @@ def get_additional_images(id_product_money: str, marketplace: str) -> List[str]:
     
     return []
 
-def download_and_encode_images(image_urls: List[str]) -> List[Dict]:
+def try_scaled_image_url(client: httpx.Client, url: str, marketplace: str, max_retries: int = 3) -> str:
+    """Try to get a scaled version of the image URL, fall back to original if not available."""
+    if marketplace == 'lamoda':
+        scaled_url = url.replace('product', 'img600x866')
+        for attempt in range(max_retries):
+            try:
+                response = client.get(scaled_url, timeout=5.0)
+                if response.status_code == 200:
+                    print(f"Using scaled image: {scaled_url}")
+                    return scaled_url
+                else:
+                    print(f"Scaled image not available (status {response.status_code}), using original: {url}")
+                    return url
+            except httpx.TimeoutException:
+                print(f"Timeout checking scaled image (attempt {attempt + 1}/{max_retries})")
+                if attempt == max_retries - 1:
+                    print(f"Max retries reached, using original: {url}")
+                    return url
+            except Exception as e:
+                print(f"Error checking scaled image: {type(e).__name__}: {str(e)}")
+                return url
+    return url
+
+def download_and_encode_images(image_urls: List[str], marketplace: str) -> List[Dict]:
     """Download images and convert them to base64 format for Gemini."""
     encoded_images = []
-    with httpx.Client() as client:
+    timeout = httpx.Timeout(10.0, connect=5.0)
+    with httpx.Client(timeout=timeout) as client:
         for url in image_urls:
-            try:
-                response = client.get(url)
-                response.raise_for_status()
-                encoded_image = base64.b64encode(response.content).decode('utf-8')
-                encoded_images.append({
-                    'mime_type': 'image/jpeg',  # Assuming JPEG format
-                    'data': encoded_image
-                })
-            except Exception as e:
-                print(f"Error downloading image {url}: {str(e)}")
+            max_retries = 3
+            for attempt in range(max_retries):
+                try:
+                    # Try to get scaled version if available
+                    final_url = try_scaled_image_url(client, url, marketplace)
+                    response = client.get(final_url)
+                    response.raise_for_status()
+                    encoded_image = base64.b64encode(response.content).decode('utf-8')
+                    encoded_images.append({
+                        'mime_type': 'image/jpeg',  # Assuming JPEG format
+                        'data': encoded_image
+                    })
+                    break  # Success, exit retry loop
+                except httpx.TimeoutException:
+                    print(f"Timeout downloading image (attempt {attempt + 1}/{max_retries}): {url}")
+                    if attempt == max_retries - 1:
+                        print(f"Max retries reached, skipping image: {url}")
+                except Exception as e:
+                    print(f"Error downloading image: {type(e).__name__}: {str(e)}")
+                    if attempt == max_retries - 1:
+                        print(f"Max retries reached, skipping image: {url}")
     return encoded_images
 
 def get_gemini_response(model_name: str, encoded_images: List[Dict], prompt: str) -> str:
@@ -151,7 +186,7 @@ def process_input(id_product_money: str, prompt: str) -> Tuple[List[str], str, s
         print(f"\nAll image URLs: {all_image_urls}")
         
         print("\nDownloading and encoding images...")
-        encoded_images = download_and_encode_images(all_image_urls)
+        encoded_images = download_and_encode_images(all_image_urls, marketplace)
         print(f"Number of encoded images: {len(encoded_images)}")
         
         if not encoded_images:

test.py

@@ -94,21 +94,56 @@ def get_additional_images(id_product_money: str, marketplace: str) -> List[str]:
     
     return []
 
-def download_and_encode_images(image_urls: List[str]) -> List[Dict]:
+def try_scaled_image_url(client: httpx.Client, url: str, marketplace: str, max_retries: int = 3) -> str:
+    """Try to get a scaled version of the image URL, fall back to original if not available."""
+    if marketplace == 'lamoda':
+        scaled_url = url.replace('product', 'img600x866')
+        for attempt in range(max_retries):
+            try:
+                response = client.get(scaled_url, timeout=5.0)
+                if response.status_code == 200:
+                    print(f"Using scaled image: {scaled_url}")
+                    return scaled_url
+                else:
+                    print(f"Scaled image not available (status {response.status_code}), using original: {url}")
+                    return url
+            except httpx.TimeoutException:
+                print(f"Timeout checking scaled image (attempt {attempt + 1}/{max_retries})")
+                if attempt == max_retries - 1:
+                    print(f"Max retries reached, using original: {url}")
+                    return url
+            except Exception as e:
+                print(f"Error checking scaled image: {type(e).__name__}: {str(e)}")
+                return url
+    return url
+
+def download_and_encode_images(image_urls: List[str], marketplace: str) -> List[Dict]:
     """Download images and convert them to base64 format for Gemini."""
     encoded_images = []
-    with httpx.Client() as client:
+    timeout = httpx.Timeout(10.0, connect=5.0)
+    with httpx.Client(timeout=timeout) as client:
         for url in image_urls:
-            try:
-                response = client.get(url)
-                response.raise_for_status()
-                encoded_image = base64.b64encode(response.content).decode('utf-8')
-                encoded_images.append({
-                    'mime_type': 'image/jpeg',  # Assuming JPEG format
-                    'data': encoded_image
-                })
-            except Exception as e:
-                print(f"Error downloading image {url}: {str(e)}")
+            max_retries = 3
+            for attempt in range(max_retries):
+                try:
+                    # Try to get scaled version if available
+                    final_url = try_scaled_image_url(client, url, marketplace)
+                    response = client.get(final_url)
+                    response.raise_for_status()
+                    encoded_image = base64.b64encode(response.content).decode('utf-8')
+                    encoded_images.append({
+                        'mime_type': 'image/jpeg',  # Assuming JPEG format
+                        'data': encoded_image
+                    })
+                    break  # Success, exit retry loop
+                except httpx.TimeoutException:
+                    print(f"Timeout downloading image (attempt {attempt + 1}/{max_retries}): {url}")
+                    if attempt == max_retries - 1:
+                        print(f"Max retries reached, skipping image: {url}")
+                except Exception as e:
+                    print(f"Error downloading image: {type(e).__name__}: {str(e)}")
+                    if attempt == max_retries - 1:
+                        print(f"Max retries reached, skipping image: {url}")
     return encoded_images
 
 def get_gemini_response(model_name: str, encoded_images: List[Dict], prompt: str) -> str:
@@ -150,7 +185,7 @@ def process_input(id_product_money: str, prompt: str) -> Tuple[List[str], str, s
         print(f"\nAll image URLs: {all_image_urls}")
         
         print("\nDownloading and encoding images...")
-        encoded_images = download_and_encode_images(all_image_urls)
+        encoded_images = download_and_encode_images(all_image_urls, marketplace)
         print(f"Number of encoded images: {len(encoded_images)}")
         
         if not encoded_images: