Update app.py

app.py

@@ -45,16 +45,19 @@ collection_name = 'celeb_images'
 celeb_images = client[db_name][collection_name]
 
 # Function to generate image description using Claude 3 Sonnet
-def generate_image_description_with_claude(image_base64):
+def generate_image_description_with_claude(images_base64_strs, image_base64):
     claude_body = json.dumps({
         "anthropic_version": "bedrock-2023-05-31",
         "max_tokens": 1000,
-        "system": "Please respond as a facial similarity detection expert.",
+        "system": "Please act as face comperison analyzer.",
         "messages": [{
             "role": "user",
             "content": [
                 {"type": "image", "source": {"type": "base64", "media_type": "image/jpeg", "data": image_base64}},
-                {"type": "text", "text": "Who is in this image?"}
+                 {"type": "image", "source": {"type": "base64", "media_type": "image/jpeg", "data": images_base64_strs[0]}},
+                 {"type": "image", "source": {"type": "base64", "media_type": "image/jpeg", "data": images_base64_strs[1]}},
+                 {"type": "image", "source": {"type": "base64", "media_type": "image/jpeg", "data": images_base64_strs[2]}},
+                {"type": "text", "text": "Please let the user know how his first image is similar to the other 3 and which one is the most similar?"}
             ]
         }]
     })
@@ -94,16 +97,18 @@ def start_image_search(image, text):
         }, {"$project": {"image": 1}}
     ]))
     
-    images_with_descriptions = []
+    images = []
+    images_base64_strs = []
     for image_doc in doc:
         pil_image = Image.open(io.BytesIO(base64.b64decode(image_doc['image'])))
         img_byte = io.BytesIO()
         pil_image.save(img_byte, format='JPEG')
         img_base64 = base64.b64encode(img_byte.getvalue()).decode('utf-8')
-        description = generate_image_description_with_claude(img_base64)
-        images_with_descriptions.append((pil_image, description))
+        images_base64_strs.append(img_base64)
+        images.append(pil_image)
     
-    return images_with_descriptions
+    description = generate_image_description_with_claude(images_base64_strs, img_base64_str)
+    return images_with_descriptions, description
 
 # Gradio Interface
 with gr.Blocks() as demo:
@@ -116,8 +121,8 @@ with gr.Blocks() as demo:
     """)
     gr.Interface(fn=start_image_search, 
     inputs=[gr.Image(type="pil", label="Upload an image"), gr.Textbox(label="Enter an adjustment to the image")],
-        outputs=gr.Gallery(label="Located images with AI-generated descriptions", show_label=True, elem_id="gallery",
-                           columns=[3], rows=[1], object_fit="contain", height="auto")
+        outputs=[gr.Gallery(label="Located images for AI-generated descriptions", show_label=False, elem_id="gallery",
+                           columns=[3], rows=[1], object_fit="contain", height="auto"),gr.TextBox(label="AI Based vision description")]
     )
 
 demo.launch()