Update Gradio_UI.py

Gradio_UI.py

@@ -25,19 +25,159 @@ from smolagents.memory import MemoryStep
 from smolagents.utils import _is_package_available
 
 
-def pull_messages_from_step(step_log: MemoryStep):
-    # Your existing pull_messages_from_step unchanged...
-    # (omitted here for brevity - keep your existing function as is)
-    ...
+def pull_messages_from_step(
+    step_log: MemoryStep,
+):
+    """Extract ChatMessage objects from agent steps with proper nesting"""
+    import gradio as gr
 
+    if isinstance(step_log, ActionStep):
+        # Output the step number
+        step_number = f"Step {step_log.step_number}" if step_log.step_number is not None else ""
+        yield gr.ChatMessage(role="assistant", content=f"**{step_number}**")
 
-def stream_to_gradio(agent, task: str, reset_agent_memory: bool = False, additional_args: Optional[dict] = None):
-    # Your existing stream_to_gradio unchanged...
-    # (omitted here for brevity - keep your existing function as is)
-    ...
+        # First yield the thought/reasoning from the LLM
+        if hasattr(step_log, "model_output") and step_log.model_output is not None:
+            # Clean up the LLM output
+            model_output = step_log.model_output.strip()
+            # Remove any trailing <end_code> and extra backticks, handling multiple possible formats
+            model_output = re.sub(r"```\s*<end_code>", "```", model_output)  # handles ```<end_code>
+            model_output = re.sub(r"<end_code>\s*```", "```", model_output)  # handles <end_code>```
+            model_output = re.sub(r"```\s*\n\s*<end_code>", "```", model_output)  # handles ```\n<end_code>
+            model_output = model_output.strip()
+            yield gr.ChatMessage(role="assistant", content=model_output)
+
+        # For tool calls, create a parent message
+        if hasattr(step_log, "tool_calls") and step_log.tool_calls is not None:
+            first_tool_call = step_log.tool_calls[0]
+            used_code = first_tool_call.name == "python_interpreter"
+            parent_id = f"call_{len(step_log.tool_calls)}"
+
+            # Tool call becomes the parent message with timing info
+            # First we will handle arguments based on type
+            args = first_tool_call.arguments
+            if isinstance(args, dict):
+                content = str(args.get("answer", str(args)))
+            else:
+                content = str(args).strip()
+
+            if used_code:
+                # Clean up the content by removing any end code tags
+                content = re.sub(r"```.*?\n", "", content)  # Remove existing code blocks
+                content = re.sub(r"\s*<end_code>\s*", "", content)  # Remove end_code tags
+                content = content.strip()
+                if not content.startswith("```python"):
+                    content = f"```python\n{content}\n```"
+
+            parent_message_tool = gr.ChatMessage(
+                role="assistant",
+                content=content,
+                metadata={
+                    "title": f"🛠️ Used tool {first_tool_call.name}",
+                    "id": parent_id,
+                    "status": "pending",
+                },
+            )
+            yield parent_message_tool
+
+            # Nesting execution logs under the tool call if they exist
+            if hasattr(step_log, "observations") and (
+                step_log.observations is not None and step_log.observations.strip()
+            ):  # Only yield execution logs if there's actual content
+                log_content = step_log.observations.strip()
+                if log_content:
+                    log_content = re.sub(r"^Execution logs:\s*", "", log_content)
+                    yield gr.ChatMessage(
+                        role="assistant",
+                        content=f"{log_content}",
+                        metadata={"title": "📝 Execution Logs", "parent_id": parent_id, "status": "done"},
+                    )
+
+            # Nesting any errors under the tool call
+            if hasattr(step_log, "error") and step_log.error is not None:
+                yield gr.ChatMessage(
+                    role="assistant",
+                    content=str(step_log.error),
+                    metadata={"title": "💥 Error", "parent_id": parent_id, "status": "done"},
+                )
+
+            # Update parent message metadata to done status without yielding a new message
+            parent_message_tool.metadata["status"] = "done"
+
+        # Handle standalone errors but not from tool calls
+        elif hasattr(step_log, "error") and step_log.error is not None:
+            yield gr.ChatMessage(role="assistant", content=str(step_log.error), metadata={"title": "💥 Error"})
+
+        # Calculate duration and token information
+        step_footnote = f"{step_number}"
+        if hasattr(step_log, "input_token_count") and hasattr(step_log, "output_token_count"):
+            token_str = (
+                f" | Input-tokens:{step_log.input_token_count:,} | Output-tokens:{step_log.output_token_count:,}"
+            )
+            step_footnote += token_str
+        if hasattr(step_log, "duration"):
+            step_duration = f" | Duration: {round(float(step_log.duration), 2)}" if step_log.duration else None
+            step_footnote += step_duration
+        step_footnote = f"""<span style="color: #bbbbc2; font-size: 12px;">{step_footnote}</span> """
+        yield gr.ChatMessage(role="assistant", content=f"{step_footnote}")
+        yield gr.ChatMessage(role="assistant", content="-----")
+
+
+def stream_to_gradio(
+    agent,
+    task: str,
+    reset_agent_memory: bool = False,
+    additional_args: Optional[dict] = None,
+):
+    """Runs an agent with the given task and streams the messages from the agent as gradio ChatMessages."""
+    if not _is_package_available("gradio"):
+        raise ModuleNotFoundError(
+            "Please install 'gradio' extra to use the GradioUI: `pip install 'smolagents[gradio]'`"
+        )
+    import gradio as gr
+
+    total_input_tokens = 0
+    total_output_tokens = 0
+
+    for step_log in agent.run(task, stream=True, reset=reset_agent_memory, additional_args=additional_args):
+        # Track tokens if model provides them
+        if hasattr(agent.model, "last_input_token_count"):
+            total_input_tokens += agent.model.last_input_token_count
+            total_output_tokens += agent.model.last_output_token_count
+            if isinstance(step_log, ActionStep):
+                step_log.input_token_count = agent.model.last_input_token_count
+                step_log.output_token_count = agent.model.last_output_token_count
+
+        for message in pull_messages_from_step(
+            step_log,
+        ):
+            yield message
+
+    final_answer = step_log  # Last log is the run's final_answer
+    final_answer = handle_agent_output_types(final_answer)
+
+    if isinstance(final_answer, AgentText):
+        yield gr.ChatMessage(
+            role="assistant",
+            content=f"**Final answer:**\n{final_answer.to_string()}\n",
+        )
+    elif isinstance(final_answer, AgentImage):
+        yield gr.ChatMessage(
+            role="assistant",
+            content={"path": final_answer.to_string(), "mime_type": "image/png"},
+        )
+    elif isinstance(final_answer, AgentAudio):
+        yield gr.ChatMessage(
+            role="assistant",
+            content={"path": final_answer.to_string(), "mime_type": "audio/wav"},
+        )
+    else:
+        yield gr.ChatMessage(role="assistant", content=f"**Final answer:** {str(final_answer)}")
 
 
 class GradioUI:
+    """A one-line interface to launch your agent in Gradio"""
+
     def __init__(self, agent: MultiStepAgent, file_upload_folder: str | None = None):
         if not _is_package_available("gradio"):
             raise ModuleNotFoundError(
@@ -45,28 +185,13 @@ class GradioUI:
             )
         self.agent = agent
         self.file_upload_folder = file_upload_folder
-        if self.file_upload_folder is not None and not os.path.exists(file_upload_folder):
-            os.mkdir(file_upload_folder)
+        if self.file_upload_folder is not None:
+            if not os.path.exists(file_upload_folder):
+                os.mkdir(file_upload_folder)
 
-    def interact_with_agent(self, prompt, messages, pdf_file_path=None):
+    def interact_with_agent(self, prompt, messages):
         import gradio as gr
 
-        # If PDF file and prompt/question are provided, use document_qna_tool
-        if pdf_file_path and prompt:
-            # Run document QnA tool directly
-            for tool in self.agent.tools:
-                # Assuming your document QnA tool is named "document_qna_tool"
-                if hasattr(tool, "name") and tool.name == "document_qna_tool":
-                    try:
-                        answer = tool.run(pdf_path=pdf_file_path, question=prompt)
-                    except Exception as e:
-                        answer = f"Error running Document QnA tool: {str(e)}"
-                    messages.append(gr.ChatMessage(role="user", content=prompt))
-                    messages.append(gr.ChatMessage(role="assistant", content=answer))
-                    yield messages
-                    return
-
-        # Otherwise fallback to normal chat interaction
         messages.append(gr.ChatMessage(role="user", content=prompt))
         yield messages
         for msg in stream_to_gradio(self.agent, task=prompt, reset_agent_memory=False):
@@ -84,6 +209,9 @@ class GradioUI:
             "text/plain",
         ],
     ):
+        """
+        Handle file uploads, default allowed types are .pdf, .docx, and .txt
+        """
         import gradio as gr
 
         if file is None:
@@ -97,43 +225,46 @@ class GradioUI:
         if mime_type not in allowed_file_types:
             return gr.Textbox("File type disallowed", visible=True), file_uploads_log, None
 
+        # Sanitize file name
         original_name = os.path.basename(file.name)
-        sanitized_name = re.sub(r"[^\w\-.]", "_", original_name)
+        sanitized_name = re.sub(
+            r"[^\w\-.]", "_", original_name
+        )  # Replace any non-alphanumeric, non-dash, or non-dot characters with underscores
 
         type_to_ext = {}
         for ext, t in mimetypes.types_map.items():
             if t not in type_to_ext:
                 type_to_ext[t] = ext
 
+        # Ensure the extension correlates to the mime type
         sanitized_name = sanitized_name.split(".")[:-1]
         sanitized_name.append("" + type_to_ext[mime_type])
         sanitized_name = "".join(sanitized_name)
 
-        file_path = os.path.join(self.file_upload_folder, os.path.basename(sanitized_name))
-        shutil.copy(file.name, file_path)
-
-        return gr.Textbox(f"File uploaded: {file_path}", visible=True), file_uploads_log + [file_path], file_path
-
-    def log_user_message(self, text_input, file_uploads_log):
-        # Append info about uploaded files to the user message context
-        return (
-            text_input
-            + (
-                f"\nYou have provided these files: {file_uploads_log}"
-                if len(file_uploads_log) > 0
-                else ""
-            ),
-            "",
-        )
-
+        # Save the uploaded file to the specified folder
+        output_path = os.path.join(self.file_upload_folder, sanitized_name)
+        shutil.copyfile(file.name, output_path)
+        file_uploads_log.append(output_path)
+    
+        return gr.Textbox(f"Uploaded: {sanitized_name}", visible=True), file_uploads_log, output_path
+    
+    def log_user_message(self, message, file_uploads_log, messages):
+        import gradio as gr
+    
+        if not messages:
+            messages = []
+    
+        messages.append(gr.ChatMessage(role="user", content=message))
+        return messages, ""
+    
     def launch(self, **kwargs):
         import gradio as gr
-
+    
         with gr.Blocks(fill_height=True) as demo:
             stored_messages = gr.State([])
             file_uploads_log = gr.State([])
-            current_pdf_path = gr.State(None)
-
+            current_file_path = gr.State(None)
+    
             chatbot = gr.Chatbot(
                 label="Agent",
                 type="messages",
@@ -144,37 +275,44 @@ class GradioUI:
                 resizeable=True,
                 scale=1,
             )
-
+    
             if self.file_upload_folder is not None:
-                upload_file = gr.File(label="Upload a PDF file", file_types=[".pdf"])
+                upload_file = gr.File(label="Upload a file")
                 upload_status = gr.Textbox(label="Upload Status", interactive=False, visible=False)
-
+    
                 upload_file.change(
                     self.upload_file,
-                    inputs=[upload_file, file_uploads_log],
-                    outputs=[upload_status, file_uploads_log, current_pdf_path],
+                    [upload_file, file_uploads_log],
+                    [upload_status, file_uploads_log, current_file_path],
                 )
-
-            text_input = gr.Textbox(lines=1, label="Ask a question about the document or chat")
-            submit_btn = gr.Button("Send")
-
-            # On submit, pass current_pdf_path and text input to interact_with_agent
-            def user_interact(text, messages, pdf_path):
-                if not text:
-                    return messages
-                return self.interact_with_agent(text, messages, pdf_path)
-
-            submit_btn.click(
-                user_interact,
-                inputs=[text_input, stored_messages, current_pdf_path],
-                outputs=chatbot,
-            )
-
-            # Also allow pressing Enter in textbox to submit
+    
+                # New button to send the uploaded document to the agent
+                send_doc_btn = gr.Button("Send Document")
+    
+                # Handler function for document sending
+                def send_document(file_path, messages):
+                    if not file_path:
+                        messages.append(gr.ChatMessage(role="assistant", content="No document uploaded."))
+                        return messages
+                    # Construct prompt with file path
+                    prompt = f"Please process this document: {file_path}"
+                    messages.append(gr.ChatMessage(role="user", content=prompt))
+                    for msg in stream_to_gradio(self.agent, task=prompt, reset_agent_memory=False):
+                        messages.append(msg)
+                        yield messages
+                    yield messages
+    
+                send_doc_btn.click(
+                    send_document,
+                    inputs=[current_file_path, stored_messages],
+                    outputs=chatbot,
+                )
+    
+            text_input = gr.Textbox(lines=1, label="Chat Message")
             text_input.submit(
-                user_interact,
-                inputs=[text_input, stored_messages, current_pdf_path],
-                outputs=chatbot,
-            )
-
+                self.log_user_message,
+                [text_input, file_uploads_log, stored_messages],
+                [stored_messages, text_input],
+            ).then(self.interact_with_agent, [stored_messages, chatbot], [chatbot])
+    
         demo.launch(debug=True, share=True, **kwargs)