app.py

import os
import gradio as gr
import requests
import inspect
import pandas as pd
#import smolagents  #to test
from smolagents import CodeAgent, InferenceClientModel, DuckDuckGoSearchTool, HfApiModel, load_tool, tool
from huggingface_hub import InferenceClient
import json
from final_answer import FinalAnswerTool

api_url = "https://agents-course-unit4-scoring.hf.space"
questions_url = f"{api_url}/questions"
submit_url = f"{api_url}/submit"

class BasicAgent:
    def __init__(self):
        print("BasicAgent initialized.")
    def __call__(self, question: str) -> str:
        print(f"Agent received question (first 50 chars): {question[:50]}...")
        
        fixed_answer = "This is a default answer."
        print(f"Agent returning fixed answer: {fixed_answer}")
        return fixed_answer

def load_questions_from_file(filepath="questions.json"):
    try:
        with open(filepath, "r", encoding="utf-8") as f:
            questions_data = json.load(f)
            if not questions_data:
                print("Loaded file is empty.")
                return "Loaded file is empty.", None
            print(f"Loaded {len(questions_data)} questions from file.")
            return "Loaded questions successfully.", questions_data
    except FileNotFoundError:
        print("File not found. Please run the API fetch first.")
        return "File not found.", None
    except json.JSONDecodeError as e:
        print(f"Error decoding JSON: {e}")
        return f"Error decoding JSON: {e}", None
    except Exception as e:
        print(f"Unexpected error: {e}")
        return f"Unexpected error: {e}", None

#set up
#token

#Model
model = HfApiModel(
max_tokens=2096,
temperature=0.5,
model_id='Qwen/Qwen2.5-Coder-32B-Instruct',# it is possible that this model may be overloaded
custom_role_conversions=None,
)

final_answer = FinalAnswerTool()

#Agent
agent_codeagent = CodeAgent(
    model=model,
    tools=[final_answer], ## add your tools here (don't remove final answer)
    max_steps=6,
    verbosity_level=1,
    grammar=None,
    planning_interval=None,
    name=None,
    description=None
    #prompt_templates=prompt_templates
)


# #def run_and_submit_one():
#     # 1. Instantiate Agent ( modify this part to create your agent)
#     try:
#         #agent = BasicAgent()
#         agent = agent_codeagent
        
#     except Exception as e:
#         print(f"Error instantiating agent: {e}")
#         return f"Error initializing agent: {e}", None
    
#    # 2. Fetch Questions by loading from local json
#     status_message, questions_data = load_questions_from_file()

#     if questions_data is not None and len(questions_data) > 0:
#         first_question = questions_data[0]
#         print("First question object:", first_question)

#         #To test
#         question_text = first_question.get("question")
#         task_id = first_question.get("task_id")
#         print(f"\nTask ID: {task_id}")
#         print(f"Question: {question_text}")
#     else:
#         print("No data found.")

#     # 3. Run your Agent
#     results_log = []
#     answers_payload = []

#     try:
#         submitted_answer = agent(question_text)
#         answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer})
#         results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
#     except Exception as e:
#         print(f"Error running agent on task {task_id}: {e}")
#         results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": f"AGENT ERROR: {e}"})

#     if not answers_payload:
#         print("Agent did not produce any answers to submit.")
#         return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)

#run_and_submit_one()       

# Gradio handler that runs the agent
def run_once(state):
    if state is not None:
        return "Already run once. Refresh to rerun.", state

    status_message, questions_data = load_questions_from_file()
    if questions_data is None or len(questions_data) == 0:
        return "No questions found or failed to load.", None

    question = questions_data[0]
    question_text = question["question"]
    task_id = question["task_id"]

    try:
        answer = agent_codeagent(question_text)
        output = f"Answer to task {task_id}:\n{answer}"
        return output, output
    except Exception as e:
        return f"Error running agent: {e}", None

# Create Gradio interface
with gr.Blocks() as demo:
    gr.Markdown("## Run AI Agent Once")

    output_text = gr.Textbox(label="Agent Output", lines=10)
    run_button = gr.Button("Run Agent")
    state = gr.State()  # cache variable to prevent re-runs

    run_button.click(fn=run_once, inputs=state, outputs=[output_text, state])

# Launch the interface
demo.launch()