Update app.py

app.py

@@ -9,8 +9,7 @@ GEMINI_API_KEY = os.getenv("GEMINI_API_KEY")
 genai.configure(api_key=GEMINI_API_KEY)
 
 # we will be using the Gemini 2.0 Flash model with Thinking capabilities
-model = genai.GenerativeModel("gemini-2.0-flash-thinking-exp-1219")
-
+model = genai.GenerativeModel("gemini-2.0-flash-thinking-exp-1219") # Consider Gemini Pro Vision for Image input
 
 def format_chat_history(messages: list) -> list:
     """
@@ -26,27 +25,58 @@ def format_chat_history(messages: list) -> list:
             })
     return formatted_history
 
-def stream_gemini_response(user_message: str, messages: list) -> Iterator[list]:
+def stream_gemini_response(message_input: str|gr.File, messages: list) -> Iterator[list]:
     """
-    Streams thoughts and response with conversation history support.
+    Streams thoughts and response with conversation history support, handling text or file input.
     """
-    try:
-        print(f"\n=== New Request ===")
+    user_message = ""
+    input_file = None
+
+    if isinstance(message_input, str):
+        user_message = message_input
+        print(f"\n=== New Request (Text) ===")
         print(f"User message: {user_message}")
+    elif isinstance(message_input, gr.utils.SerializableData): #gr.File directly won't be recognized correctly after gr v4.x, use SerializableData instead.
+        input_file = message_input.name # Access the temporary file path
+        file_type = message_input.orig_name.split('.')[-1].lower() #Get original filename's extension
+        print(f"\n=== New Request (File) ===")
+        print(f"File uploaded: {input_file}, type: {file_type}")
+
+        try:
+            with open(input_file, "rb") as f: #Open file in binary mode for universal handling
+                file_data = f.read()
+
+            if file_type in ['png', 'jpg', 'jpeg', 'gif']: #Example Image Types - expand as needed
+                user_message = {"inline_data": {"mime_type": f"image/{file_type}", "data": file_data}} #Prepare image part for Gemini
+            elif file_type == 'csv':
+                user_message = {"inline_data": {"mime_type": "text/csv", "data": file_data}} #Prepare csv part
+
+        except Exception as e:
+            print(f"Error reading file: {e}")
+            messages.append(ChatMessage(role="assistant", content=f"Error reading file: {e}"))
+            yield messages
+            return
+    else:
+        messages.append(ChatMessage(role="assistant", content="Sorry, I cannot understand this input format."))
+        yield messages
+        return
+
 
+    try:
         # Format chat history for Gemini
         chat_history = format_chat_history(messages)
 
         # Initialize Gemini chat
         chat = model.start_chat(history=chat_history)
-        response = chat.send_message(user_message, stream=True)
+        response = chat.send_message(user_message, stream=True) #Send the message part as is
 
-        # Initialize buffers and flags
+        # Initialize buffers and flags - same as before
         thought_buffer = ""
         response_buffer = ""
         thinking_complete = False
 
-        # Add initial thinking message
+
+        # Add initial thinking message - same as before
         messages.append(
             ChatMessage(
                 role="assistant",
@@ -55,7 +85,7 @@ def stream_gemini_response(user_message: str, messages: list) -> Iterator[list]:
             )
         )
 
-        for chunk in response:
+        for chunk in response: #streaming logic - same as before
             parts = chunk.candidates[0].content.parts
             current_chunk = parts[0].text
 
@@ -108,6 +138,7 @@ def stream_gemini_response(user_message: str, messages: list) -> Iterator[list]:
 
         print(f"\n=== Final Response ===\n{response_buffer}")
 
+
     except Exception as e:
         print(f"\n=== Error ===\n{str(e)}")
         messages.append(
@@ -118,14 +149,14 @@ def stream_gemini_response(user_message: str, messages: list) -> Iterator[list]:
         )
         yield messages
 
-def user_message(msg: str, history: list) -> tuple[str, list]:
+def user_message(message_text, file_upload, history: list) -> tuple[str, None, list]:
     """Adds user message to chat history"""
-    history.append(ChatMessage(role="user", content=msg))
-    return "", history
-
+    msg = message_text if message_text else file_upload
+    history.append(ChatMessage(role="user", content=msg if isinstance(msg, str) else msg.name)) #Store message or filename in history.
+    return "", None, history #clear both input fields
 
 # Create the Gradio interface
-with gr.Blocks(theme=gr.themes.Soft(primary_hue="teal", secondary_hue="slate", neutral_hue="neutral")) as demo: # Using Soft theme with adjusted hues for a refined look
+with gr.Blocks(theme=gr.themes.Soft(primary_hue="teal", secondary_hue="slate", neutral_hue="neutral")) as demo:
     gr.Markdown("# Gemini 2.0 Flash 'Thinking' Chatbot 💭")
 
     chatbot = gr.Chatbot(
@@ -141,60 +172,65 @@ with gr.Blocks(theme=gr.themes.Soft(primary_hue="teal", secondary_hue="slate", n
             lines=1,
             label="Chat Message",
             placeholder="Type your message here...",
-            scale=4
+            scale=3
         )
+        file_upload = gr.File(label="Upload File", file_types=["image", "file"], scale=2) # Allow image and any file type
+
 
         clear_button = gr.Button("Clear Chat", scale=1)
 
     # Set up event handlers
     msg_store = gr.State("")  # Store for preserving user message
 
+
     input_box.submit(
-        lambda msg: (msg, msg, ""),  # Store message and clear input
-        inputs=[input_box],
-        outputs=[msg_store, input_box, input_box],
+        user_message,
+        inputs=[input_box, file_upload, chatbot],
+        outputs=[input_box, file_upload, chatbot],
         queue=False
     ).then(
-        user_message,  # Add user message to chat
-        inputs=[msg_store, chatbot],
-        outputs=[input_box, chatbot],
+        stream_gemini_response,
+        inputs=[input_box, chatbot], # Input either from text box or file, logic inside stream_gemini_response
+        outputs=chatbot
+    )
+
+    file_upload.upload(
+        user_message,
+        inputs=[input_box, file_upload, chatbot], # even textbox is input here so clearing both will work
+        outputs=[input_box, file_upload, chatbot],
         queue=False
     ).then(
-        stream_gemini_response,  # Generate and stream response
-        inputs=[msg_store, chatbot],
+        stream_gemini_response,
+        inputs=[file_upload, chatbot], # Input is now the uploaded file.
         outputs=chatbot
     )
 
+
     clear_button.click(
         lambda: ([], "", ""),
         outputs=[chatbot, input_box, msg_store],
         queue=False
     )
 
-    gr.Markdown(  # Description moved to the bottom
+    gr.Markdown(  # Description moved to the bottom - unchanged
         """
         <br><br><br>  <!-- Add some vertical space -->
         ---
         ### About this Chatbot
-
         This chatbot demonstrates the experimental 'thinking' capability of the **Gemini 2.0 Flash** model.
         You can observe the model's thought process as it generates responses, displayed with the "⚙️ Thinking" prefix.
-
         **Key Features:**
-
         *   Powered by Google's **Gemini 2.0 Flash** model.
         *   Shows the model's **thoughts** before the final answer (experimental feature).
         *   Supports **conversation history** for multi-turn chats.
+        *   Supports **Image and CSV file uploads** for analysis.
         *   Uses **streaming** for a more interactive experience.
-
         **Instructions:**
-
-        1.  Type your message in the input box below.
-        2.  Press Enter or click Submit to send.
+        1.  Type your message in the input box or Upload a file below.
+        2.  Press Enter/Submit or Upload to send.
         3.  Observe the chatbot's "Thinking" process followed by the final response.
         4.  Use the "Clear Chat" button to start a new conversation.
-
-        *Please note*: The 'thinking' feature is experimental and the quality of thoughts may vary.
+        *Please note*: The 'thinking' feature is experimental and the quality of thoughts may vary. File analysis capabilities may be limited depending on the model's experimental features.
         """
     )