diff --git a/api.py b/api.py index f39c900..49acfe2 100755 --- a/api.py +++ b/api.py @@ -5,6 +5,7 @@ import uvicorn import aiofiles import configparser import asyncio +import time from typing import List from fastapi import FastAPI from fastapi.responses import JSONResponse @@ -136,7 +137,9 @@ async def get_latest_answer(): "answer": interaction.current_agent.last_answer, "agent_name": interaction.current_agent.agent_name if interaction.current_agent else "None", "success": "false", - "blocks": {f'{i}': block.jsonify() for i, block in enumerate(interaction.current_agent.get_blocks_result())} if interaction.current_agent else {} + "blocks": {f'{i}': block.jsonify() for i, block in enumerate(interaction.current_agent.get_blocks_result())} if interaction.current_agent else {}, + "status": interaction.current_agent.status_message if interaction.current_agent else "No status available", + "timestamp": str(time.time()) } query_resp_history.append(query_resp) return JSONResponse(status_code=200, content=query_resp) @@ -171,7 +174,9 @@ async def process_query(request: QueryRequest): answer="Waiting for agent...", agent_name="Waiting for agent...", success="false", - blocks={} + blocks={}, + status="Waiting for agent...", + timestamp=str(time.time()) ) if is_generating: logger.warning("Another query is being processed, please wait.") @@ -208,7 +213,9 @@ async def process_query(request: QueryRequest): "answer": query_resp.answer, "agent_name": query_resp.agent_name, "success": query_resp.success, - "blocks": query_resp.blocks + "blocks": query_resp.blocks, + "status": query_resp.status, + "timestamp": query_resp.timestamp } query_resp_history.append(query_resp_dict) diff --git a/sources/agents/agent.py b/sources/agents/agent.py index 79cab02..7fda4b3 100644 --- a/sources/agents/agent.py +++ b/sources/agents/agent.py @@ -45,6 +45,7 @@ class Agent(): self.tools = {} self.blocks_result = [] self.last_answer = "" + self.status_message = "Haven't started yet" self.verbose = verbose self.executor = ThreadPoolExecutor(max_workers=1) @@ -118,6 +119,7 @@ class Agent(): """ Asynchronously ask the LLM to process the prompt. """ + self.status_message = "Thinking..." loop = asyncio.get_event_loop() return await loop.run_in_executor(self.executor, self.sync_llm_request) diff --git a/sources/agents/browser_agent.py b/sources/agents/browser_agent.py index 221c9f9..395ca53 100644 --- a/sources/agents/browser_agent.py +++ b/sources/agents/browser_agent.py @@ -308,6 +308,7 @@ class BrowserAgent(Agent): pretty_print(f"Web agent requested exit.\n{reasoning}\n\n{ai_prompt}", color="failure") return ai_prompt, "" animate_thinking(f"Searching...", color="status") + self.status_message = "Searching..." search_result_raw = self.tools["web_search"].execute([ai_prompt], False) search_result = self.jsonify_search_results(search_result_raw)[:16] self.show_search_results(search_result) @@ -322,6 +323,7 @@ class BrowserAgent(Agent): extracted_form = self.extract_form(answer) if len(extracted_form) > 0: + self.status_message = "Filling web form..." pretty_print(f"Filling inputs form...", color="status") fill_success = self.browser.fill_form(extracted_form) page_text = self.browser.get_text() @@ -339,12 +341,14 @@ class BrowserAgent(Agent): link = self.select_link(links) if Action.REQUEST_EXIT.value in answer: + self.status_message = "Exiting web browser..." pretty_print(f"Agent requested exit.", color="status") complete = True break if (link == None and not len(extracted_form)) or Action.GO_BACK.value in answer or link in self.search_history: pretty_print(f"Going back to results. Still {len(unvisited)}", color="status") + self.status_message = "Going back to search results..." unvisited = self.select_unvisited(search_result) prompt = self.make_newsearch_prompt(user_prompt, unvisited) continue @@ -357,13 +361,16 @@ class BrowserAgent(Agent): page_text = self.browser.get_text() self.navigable_links = self.browser.get_navigable() prompt = self.make_navigation_prompt(user_prompt, page_text) + self.status_message = "Navigating..." self.browser.screenshot() pretty_print("Exited navigation, starting to summarize finding...", color="status") prompt = self.conclude_prompt(user_prompt) mem_last_idx = self.memory.push('user', prompt) + self.status_message = "Summarizing findings..." answer, reasoning = await self.llm_request() pretty_print(answer, color="output") + self.status_message = "Done" return answer, reasoning if __name__ == "__main__": diff --git a/sources/agents/casual_agent.py b/sources/agents/casual_agent.py index e5d1a13..0215575 100644 --- a/sources/agents/casual_agent.py +++ b/sources/agents/casual_agent.py @@ -23,6 +23,7 @@ class CasualAgent(Agent): animate_thinking("Thinking...", color="status") answer, reasoning = await self.llm_request() self.last_answer = answer + self.status_message = "Done" return answer, reasoning if __name__ == "__main__": diff --git a/sources/agents/code_agent.py b/sources/agents/code_agent.py index cd617d1..8c0e468 100644 --- a/sources/agents/code_agent.py +++ b/sources/agents/code_agent.py @@ -57,6 +57,7 @@ class CoderAgent(Agent): await asyncio.sleep(0) break animate_thinking("Executing code...", color="status") + self.status_message = "Executing code..." exec_success, _ = self.execute_modules(answer) answer = self.remove_blocks(answer) self.last_answer = answer @@ -65,8 +66,10 @@ class CoderAgent(Agent): break pretty_print("Execution failure", color="failure") pretty_print("Correcting code...", color="status") + self.status_message = "Correcting code..." self.show_answer() attempt += 1 + self.status_message = "Done" if attempt == max_attempts: return "I'm sorry, I couldn't find a solution to your problem. How would you like me to proceed ?", reasoning return answer, reasoning diff --git a/sources/agents/file_agent.py b/sources/agents/file_agent.py index c9338e5..348e12c 100644 --- a/sources/agents/file_agent.py +++ b/sources/agents/file_agent.py @@ -30,6 +30,7 @@ class FileAgent(Agent): exec_success, _ = self.execute_modules(answer) answer = self.remove_blocks(answer) self.last_answer = answer + self.status_message = "Done" return answer, reasoning if __name__ == "__main__": diff --git a/sources/agents/planner_agent.py b/sources/agents/planner_agent.py index b835345..75f845c 100644 --- a/sources/agents/planner_agent.py +++ b/sources/agents/planner_agent.py @@ -102,7 +102,7 @@ class PlannerAgent(Agent): ok = True return answer - async def start_agent_process(self, task: str, required_infos: dict | None) -> str: + async def start_agent_process(self, task: dict, required_infos: dict | None) -> str: agent_prompt = self.make_prompt(task['task'], required_infos) pretty_print(f"Agent {task['agent']} started working...", color="status") agent_answer, _ = await self.agents[task['agent'].lower()].process(agent_prompt, None) @@ -123,6 +123,7 @@ class PlannerAgent(Agent): if agents_tasks == (None, None): return "Failed to parse the tasks.", "" for task_name, task in agents_tasks: + self.status_message = "Starting agent process..." pretty_print(f"I will {task_name}.", color="info") pretty_print(f"Assigned agent {task['agent']} to {task_name}", color="info") if speech_module: speech_module.speak(f"I will {task_name}. I assigned the {task['agent']} agent to the task.") diff --git a/sources/schemas.py b/sources/schemas.py index 8a00f8c..04fb900 100644 --- a/sources/schemas.py +++ b/sources/schemas.py @@ -22,6 +22,8 @@ class QueryResponse(BaseModel): agent_name: str success: str blocks: dict + status: str + timestamp: str def __str__(self): return f"Done: {self.done}, Answer: {self.answer}, Agent Name: {self.agent_name}, Success: {self.success}, Blocks: {self.blocks}"