Update app.py

app.py

@@ -64,24 +64,27 @@ def cache_pil_image(image: Image.Image) -> str:
     image.save(image_path, "JPEG")
     return image_path
 
-def upload(files: Optional[List[str]], chatbot: List[dict]) -> List[dict]:
-    for file in files:
-        image = Image.open(file).convert("RGB")
-        image_preview = preprocess_image(image)
-        if image_preview:
-            gr.Image(image_preview).render()
-        image_path = cache_pil_image(image)
-        chatbot.append({"role": "user", "content": f"Uploaded image: {image_path}"})
+def upload(files: Optional[List[gr.File]], chatbot: List[dict]) -> List[dict]:
+    if files:
+        for file in files:
+            image = Image.open(file.name).convert("RGB")
+            image_preview = preprocess_image(image)
+            if image_preview:
+                image_path = cache_pil_image(image)
+                chatbot.append({"role": "user", "content": f"Uploaded image: {image_path}"})
     return chatbot
 
 def advanced_response(
-    files: Optional[List[str]],
+    text_prompt: str,
+    files: Optional[List[gr.File]],
     model_choice: str,
     system_instruction: str,
     chatbot: List[dict],
 ):
-    if not files:
-        return chatbot
+    if not text_prompt.strip() and not files:
+        chatbot.append({"role": "assistant", "content": "Please provide a prompt or upload an image."})
+        yield chatbot
+        return
 
     model = genai.GenerativeModel(
         model_name=model_choice,
@@ -95,13 +98,17 @@ def advanced_response(
     chat = model.start_chat(history=chatbot)
     chat.system_instruction = system_instruction
 
-    images = [cache_pil_image(preprocess_image(Image.open(file))) for file in files]
-    response = chat.generate_content(images, stream=True)
+    if text_prompt:
+        chatbot.append({"role": "user", "content": text_prompt})
+    if files:
+        images = [cache_pil_image(preprocess_image(Image.open(file.name))) for file in files]
+        chatbot.append({"role": "user", "content": f"Uploaded images: {', '.join(images)}"})
 
-    chatbot.append({"role": "assistant", "content": ""})
-    for chunk in response:
-        chatbot[-1]["content"] += chunk.text
-        yield chatbot
+    response = chat.send_message(text_prompt)
+    response.resolve()
+
+    chatbot.append({"role": "assistant", "content": response.text})
+    yield chatbot
 
 # Construcción de la interfaz con las dos pestañas originales
 with gr.Blocks() as demo:
@@ -152,7 +159,7 @@ with gr.Blocks() as demo:
 
         run_button_2.click(
             fn=advanced_response,
-            inputs=[upload_button, model_dropdown_2, system_instruction_2, chatbot_2],
+            inputs=[text_input_2, upload_button, model_dropdown_2, system_instruction_2, chatbot_2],
             outputs=[chatbot_2],
         )
         upload_button.upload(