Update Gradio_UI.py

Gradio_UI.py

@@ -20,7 +20,7 @@ import shutil
 from typing import Optional
 
 from smolagents.agent_types import AgentAudio, AgentImage, AgentText, handle_agent_output_types
-from smolagents.agents import ActionStep, MultiStepAgent
+from smolagents.agents import ActionStep, MultiStepAgent # Ensure MultiStepAgent is correctly referenced
 from smolagents.memory import MemoryStep
 from smolagents.utils import _is_package_available
 
@@ -51,10 +51,9 @@ def pull_messages_from_step(
         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)}"
+            parent_id = f"call_{len(step_log.tool_calls)}_{step_log.step_number or 'x'}" # Make parent_id more unique
 
             # 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)))
@@ -62,12 +61,14 @@ def pull_messages_from_step(
                 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"):
+                if not content.startswith("```python"): # Ensure it's a python block
                     content = f"```python\n{content}\n```"
+                else: # If it is, ensure newlines are correct
+                    content = content.replace("```python", "```python\n").replace("\n```", "\n```")
+
 
             parent_message_tool = gr.ChatMessage(
                 role="assistant",
@@ -80,12 +81,11 @@ def pull_messages_from_step(
             )
             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:
+                if log_content: # Only yield if there's actual content
                     log_content = re.sub(r"^Execution logs:\s*", "", log_content)
                     yield gr.ChatMessage(
                         role="assistant",
@@ -93,33 +93,34 @@ def pull_messages_from_step(
                         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"},
                 )
+            # This direct update might not work as expected as yield creates new objects.
+            # Status update is visual; actual logic might be more complex.
+            parent_message_tool.metadata["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:
+        elif hasattr(step_log, "error") and step_log.error is not None: # Standalone errors
             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"):
+        step_footnote_parts = [step_number]
+        if hasattr(step_log, "input_token_count") and step_log.input_token_count is not None and \
+           hasattr(step_log, "output_token_count") and step_log.output_token_count is not None:
             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}")
+            step_footnote_parts.append(token_str)
+        if hasattr(step_log, "duration") and step_log.duration is not None:
+            step_duration = f" | Duration: {round(float(step_log.duration), 2)}s"
+            step_footnote_parts.append(step_duration)
+        
+        step_footnote_text = "".join(filter(None, step_footnote_parts))
+        if step_footnote_text:
+             step_footnote = f"""<span style="color: #bbbbc2; font-size: 12px;">{step_footnote_text}</span> """
+             yield gr.ChatMessage(role="assistant", content=f"{step_footnote}")
         yield gr.ChatMessage(role="assistant", content="-----")
 
 
@@ -136,43 +137,33 @@ def stream_to_gradio(
         )
     import gradio as gr
 
-    total_input_tokens = 0
-    total_output_tokens = 0
+    # Reset interaction logs for the new run if the agent has this attribute
+    if hasattr(agent, 'interaction_logs'):
+        agent.interaction_logs.clear()
+        print("DEBUG: Cleared agent interaction_logs for new run.")
+
 
     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):
+        if hasattr(agent.model, "last_input_token_count") and agent.model.last_input_token_count is not None: # Check for None
+            if isinstance(step_log, ActionStep): # Only add token counts to ActionSteps
                 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,
-        ):
+        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)
+    # After the loop, step_log holds the final answer or the last step's log
+    final_answer_content = step_log 
+    final_answer_processed = handle_agent_output_types(final_answer_content)
 
-    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"},
-        )
+    if isinstance(final_answer_processed, AgentText):
+        yield gr.ChatMessage(role="assistant", content=f"**Final answer:**\n{final_answer_processed.to_string()}\n")
+    elif isinstance(final_answer_processed, AgentImage):
+        yield gr.ChatMessage(role="assistant", content={"path": final_answer_processed.to_string(), "mime_type": "image/png"})
+    elif isinstance(final_answer_processed, AgentAudio):
+        yield gr.ChatMessage(role="assistant", content={"path": final_answer_processed.to_string(), "mime_type": "audio/wav"})
     else:
-        yield gr.ChatMessage(role="assistant", content=f"**Final answer:** {str(final_answer)}")
+        yield gr.ChatMessage(role="assistant", content=f"**Final answer:** {str(final_answer_processed)}")
 
 
 class GradioUI:
@@ -187,17 +178,51 @@ class GradioUI:
         self.file_upload_folder = 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):
+                os.makedirs(self.file_upload_folder, exist_ok=True) # Use makedirs
+        
+        self._latest_file_path_for_download = None # For download button state
+
+    def _check_for_created_file(self):
+        """Helper function to check interaction logs for a created file path."""
+        self._latest_file_path_for_download = None # Reset
+        if hasattr(self.agent, 'interaction_logs') and self.agent.interaction_logs:
+            print(f"DEBUG UI: Checking {len(self.agent.interaction_logs)} interaction log entries.")
+            for log_entry in self.agent.interaction_logs:
+                if log_entry.get("tool_name") == "create_document":
+                    tool_output_value = log_entry.get("tool_output")
+                    print(f"DEBUG UI: Log for 'create_document', output: {tool_output_value}")
+                    if tool_output_value and isinstance(tool_output_value, str):
+                        if not tool_output_value.strip().startswith("ERROR:"):
+                            normalized_path = os.path.normpath(tool_output_value)
+                            if os.path.exists(normalized_path):
+                                self._latest_file_path_for_download = normalized_path
+                                print(f"DEBUG UI: File path for download set: {self._latest_file_path_for_download}")
+                                return True # Found a valid file
+                            else:
+                                print(f"DEBUG UI: Path from log ('{normalized_path}') does not exist.")
+                        else:
+                            print(f"DEBUG UI: 'create_document' tool reported error: {tool_output_value}")
+        return False
+
+
+    def interact_with_agent(self, prompt, messages_history, download_btn_state, file_output_state):
         import gradio as gr
+        
+        messages_history.append(gr.ChatMessage(role="user", content=prompt))
+        yield messages_history, gr.update(visible=False), gr.update(value=None, visible=False) # Hide download items initially
 
-        messages.append(gr.ChatMessage(role="user", content=prompt))
-        yield messages
+        # Stream agent messages to chatbot
         for msg in stream_to_gradio(self.agent, task=prompt, reset_agent_memory=False):
-            messages.append(msg)
-            yield messages
-        yield messages
+            messages_history.append(msg)
+            yield messages_history, gr.update(visible=False), gr.update(value=None, visible=False) # Keep hidden during streaming
+
+        # After streaming all agent messages, check for created file
+        file_found = self._check_for_created_file()
+        
+        # Update UI based on whether a file was found
+        # Yielding final state for chatbot, download button, and file component
+        yield messages_history, gr.update(visible=file_found), gr.update(value=None, visible=False)
+
 
     def upload_file(
         self,
@@ -209,9 +234,6 @@ class GradioUI:
             "text/plain",
         ],
     ):
-        """
-        Handle file uploads, default allowed types are .pdf, .docx, and .txt
-        """
         import gradio as gr
 
         if file is None:
@@ -219,78 +241,135 @@ class GradioUI:
 
         try:
             mime_type, _ = mimetypes.guess_type(file.name)
+            if mime_type is None: # Fallback if guess_type returns None
+                mime_type = file.type # Gradio File object has a 'type' attribute
         except Exception as e:
             return gr.Textbox(f"Error: {e}", visible=True), file_uploads_log
-
+        
         if mime_type not in allowed_file_types:
-            return gr.Textbox("File type disallowed", visible=True), file_uploads_log
+            return gr.Textbox(f"File type '{mime_type}' disallowed", visible=True), file_uploads_log
 
-        # Sanitize file name
         original_name = os.path.basename(file.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)
-
-        # Save the uploaded file to the specified folder
-        file_path = os.path.join(self.file_upload_folder, os.path.basename(sanitized_name))
-        shutil.copy(file.name, file_path)
+        sanitized_name = re.sub(r"[^\w\-.]", "_", original_name)
+        
+        # Ensure correct extension based on mime type, if possible
+        base_name, current_ext = os.path.splitext(sanitized_name)
+        
+        type_to_ext_map = {
+            "application/pdf": ".pdf",
+            "application/vnd.openxmlformats-officedocument.wordprocessingml.document": ".docx",
+            "text/plain": ".txt",
+        }
+        expected_ext = type_to_ext_map.get(mime_type)
+        if expected_ext and current_ext.lower() != expected_ext:
+            sanitized_name = base_name + expected_ext
+        
+        file_path = os.path.join(self.file_upload_folder, sanitized_name)
+        shutil.copy(file.name, file_path) # file.name is the temp path of the uploaded file
 
         return gr.Textbox(f"File uploaded: {file_path}", visible=True), file_uploads_log + [file_path]
 
     def log_user_message(self, text_input, file_uploads_log):
-        return (
-            text_input
-            + (
+        # This function prepares the prompt that goes to the agent.
+        # It also clears the text_input box.
+        full_prompt = text_input
+        if file_uploads_log: # Check if list is not empty
+            full_prompt += (
                 f"\nYou have been provided with these files, which might be helpful or not: {file_uploads_log}"
-                if len(file_uploads_log) > 0
-                else ""
-            ),
-            "",
-        )
+            )
+        return full_prompt, "" # Return the full prompt and an empty string to clear input
+
+    def prepare_and_show_download_file(self):
+        import gradio as gr
+        if self._latest_file_path_for_download and os.path.exists(self._latest_file_path_for_download):
+            print(f"DEBUG UI: Preparing download for UI: {self._latest_file_path_for_download}")
+            return gr.File.update(value=self._latest_file_path_for_download, 
+                                  label=os.path.basename(self._latest_file_path_for_download), 
+                                  visible=True)
+        else:
+            print("DEBUG UI: No valid file path to prepare for download component.")
+            gr.Warning("No file available for download or path is invalid.")
+            return gr.File.update(visible=False)
 
     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([])
-            chatbot = gr.Chatbot(
-                label="Agent",
-                type="messages",
-                avatar_images=(
-                    None,
-                    "https://huggingface.co/datasets/agents-course/course-images/resolve/main/en/communication/Alfred.png",
-                ),
-                resizeable=True,
-                scale=1,
-            )
-            # If an upload folder is provided, enable the upload feature
-            if self.file_upload_folder is not None:
-                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,
-                    [upload_file, file_uploads_log],
-                    [upload_status, file_uploads_log],
-                )
-            text_input = gr.Textbox(lines=1, label="Chat Message")
+        with gr.Blocks(fill_height=True, theme=gr.themes.Soft()) as demo: # Added a theme
+            # --- State Variables ---
+            # stored_messages is used to build the prompt for the agent, not directly for chatbot display here.
+            # The chatbot takes messages directly from interact_with_agent.
+            # We'll use chat_history_state for the chatbot's message list.
+            chat_history_state = gr.State([]) 
+            file_uploads_log = gr.State([]) # Tracks paths of uploaded files
+
+            # --- UI Layout ---
+            gr.Markdown("# Smol Talk with your Agent") # Title
+
+            with gr.Row():
+                with gr.Column(scale=3): # Main chat area
+                    chatbot = gr.Chatbot(
+                        label="Agent Interaction",
+                        # Bubble full width can make text hard to read, try default
+                        # bubble_full_width=False, 
+                        avatar_images=(
+                            None, # User avatar
+                            "https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo-round.png" # Agent avatar
+                        ),
+                        height=600 
+                    )
+                    text_input = gr.Textbox(
+                        lines=1, 
+                        label="Your Message to the Agent", 
+                        placeholder="Type your message and press Enter..."
+                    )
+                
+                with gr.Column(scale=1): # Sidebar for uploads and downloads
+                    if self.file_upload_folder is not None:
+                        gr.Markdown("### File Upload")
+                        upload_file_component = gr.File(label="Upload a supporting file")
+                        upload_status_display = gr.Textbox(label="Upload Status", interactive=False, visible=True, lines=2) # Make visible by default
+                        upload_file_component.upload( # Use 'upload' event for gr.File
+                            self.upload_file,
+                            [upload_file_component, file_uploads_log],
+                            [upload_status_display, file_uploads_log],
+                        )
+                    
+                    gr.Markdown("### Generated File")
+                    # This button becomes visible if a file is created by the agent
+                    download_btn = gr.Button("Download Generated File", visible=False) 
+                    # This gr.File component becomes visible and populated when the button above is clicked
+                    file_output_display = gr.File(label="Downloadable Document", visible=False, interactive=False) 
+
+            # --- Event Handling ---
+            
+            # When user submits text_input:
+            # 1. log_user_message: prepares the prompt (text + file info), clears text_input.
+            #    The output 'prepared_prompt' is then passed to interact_with_agent.
+            # 2. interact_with_agent: streams agent's responses to chatbot, updates download button.
+            
+            # We need a state to hold the prepared prompt temporarily if log_user_message is separate
+            prepared_prompt_state = gr.State("")
+
             text_input.submit(
                 self.log_user_message,
                 [text_input, file_uploads_log],
-                [stored_messages, text_input],
-            ).then(self.interact_with_agent, [stored_messages, chatbot], [chatbot])
+                [prepared_prompt_state, text_input] # prepared_prompt_state gets the full prompt, text_input is cleared
+            ).then(
+                self.interact_with_agent,
+                [prepared_prompt_state, chat_history_state, download_btn, file_output_display], # Pass current UI states
+                [chat_history_state, download_btn, file_output_display] # Update these UI states
+            )
+
+            # When download_btn is clicked:
+            download_btn.click(
+                self.prepare_and_show_download_file, 
+                [], # No inputs needed from UI for this action
+                [file_output_display] # Update the file_output_display component
+            )
 
-        demo.launch(debug=True, share=True, **kwargs)
+        # Launch the Gradio app
+        # Set share=False if running locally or on Spaces where share=True might be an issue
+        demo.launch(debug=True, share=kwargs.get("share", False), **kwargs)
 
 
 __all__ = ["stream_to_gradio", "GradioUI"]