app.py

import os
import gradio as gr
import requests
import inspect
import pandas as pd
from typing import Any

# (Keep Constants as is)
# --- Constants ---
DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"

# --- Advanced Modular Agent Implementation ---
import json
import logging
import mimetypes
import openpyxl
import numpy as np
from datetime import datetime
from io import BytesIO
from PIL import Image
import subprocess
import tempfile
from huggingface_hub import InferenceClient
import cv2
import torch
from bs4 import BeautifulSoup

logging.basicConfig(filename='gaia_agent.log', level=logging.INFO, format='%(asctime)s %(levelname)s:%(message)s')
logger = logging.getLogger(__name__)
HF_TOKEN = os.environ.get("HF_TOKEN", "")

def llama3_chat(prompt):
    try:
        client = InferenceClient(provider="fireworks-ai", api_key=HF_TOKEN)
        completion = client.chat.completions.create(
            model="meta-llama/Llama-3.1-8B-Instruct",
            messages=[{"role": "user", "content": prompt}],
        )
        return completion.choices[0].message.content
    except Exception as e:
        logging.error(f"llama3_chat error: {e}")
        return f"LLM error: {e}"

def mixtral_chat(prompt):
    try:
        client = InferenceClient(provider="hf-inference", api_key=HF_TOKEN)
        completion = client.chat.completions.create(
            model="mistralai/Mixtral-8x7B-Instruct-v0.1",
            messages=[{"role": "user", "content": prompt}],
        )
        return completion.choices[0].message.content
    except Exception as e:
        logging.error(f"mixtral_chat error: {e}")
        return f"LLM error: {e}"

def extractive_qa(question, context):
    try:
        client = InferenceClient(provider="hf-inference", api_key=HF_TOKEN)
        answer = client.question_answering(
            question=question,
            context=context,
            model="deepset/roberta-base-squad2",
        )
        return answer["answer"]
    except Exception as e:
        logging.error(f"extractive_qa error: {e}")
        return f"QA error: {e}"

def table_qa(query, table):
    try:
        client = InferenceClient(provider="hf-inference", api_key=HF_TOKEN)
        answer = client.table_question_answering(
            query=query,
            table=table,
            model="google/tapas-large-finetuned-wtq",
        )
        return answer["answer"]
    except Exception as e:
        logging.error(f"table_qa error: {e}")
        return f"Table QA error: {e}"

def asr_transcribe(audio_path):
    try:
        import torchaudio
        from transformers import pipeline
        asr = pipeline("automatic-speech-recognition", model="openai/whisper-base.en")
        result = asr(audio_path)
        return result["text"]
    except Exception as e:
        logging.error(f"asr_transcribe error: {e}")
        return f"ASR error: {e}"

def image_caption(image_path):
    try:
        from transformers import BlipProcessor, BlipForConditionalGeneration
        from PIL import Image
        processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
        model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base")
        raw_image = Image.open(image_path).convert('RGB')
        inputs = processor(raw_image, return_tensors="pt")
        out = model.generate(**inputs)
        return processor.decode(out[0], skip_special_tokens=True)
    except Exception as e:
        logging.error(f"image_caption error: {e}")
        return f"Image captioning error: {e}"

def code_analysis(py_path):
    try:
        with open(py_path) as f:
            code = f.read()
        with tempfile.NamedTemporaryFile(mode='w', suffix='.py', delete=False) as tmp:
            tmp.write(code)
            tmp_path = tmp.name
        try:
            result = subprocess.run([
                "python3", tmp_path
            ], capture_output=True, text=True, timeout=5)
            if result.returncode == 0:
                output = result.stdout.strip().split('\n')
                return output[-1] if output else ''
            else:
                logging.error(f"code_analysis subprocess error: {result.stderr}")
                return f"Code error: {result.stderr}"
        except subprocess.TimeoutExpired:
            logging.error("code_analysis timeout")
            return "Code execution timed out"
        finally:
            os.remove(tmp_path)
    except Exception as e:
        logging.error(f"code_analysis error: {e}")
        return f"Code analysis error: {e}"

def youtube_video_qa(youtube_url, question):
    import subprocess
    import tempfile
    import os
    from transformers import pipeline
    try:
        with tempfile.TemporaryDirectory() as tmpdir:
            # Download video
            video_path = os.path.join(tmpdir, "video.mp4")
            cmd = ["yt-dlp", "-f", "mp4", "-o", video_path, youtube_url]
            subprocess.run(cmd, check=True)
            # Extract audio for ASR
            audio_path = os.path.join(tmpdir, "audio.mp3")
            cmd_audio = ["yt-dlp", "-f", "bestaudio", "--extract-audio", "--audio-format", "mp3", "-o", audio_path, youtube_url]
            subprocess.run(cmd_audio, check=True)
            # Transcribe audio
            asr = pipeline("automatic-speech-recognition", model="openai/whisper-base.en")
            result = asr(audio_path)
            transcript = result["text"]
            # Extract frames for vision QA
            cap = cv2.VideoCapture(video_path)
            frame_count = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
            fps = int(cap.get(cv2.CAP_PROP_FPS))
            frames = []
            for i in range(0, frame_count, max(1, fps*5)):
                cap.set(cv2.CAP_PROP_POS_FRAMES, i)
                ret, frame = cap.read()
                if not ret:
                    break
                img = Image.fromarray(cv2.cvtColor(frame, cv2.COLOR_BGR2RGB))
                frames.append(img)
            cap.release()
            # Object detection (YOLOv8)
            try:
                from ultralytics import YOLO
                yolo = YOLO("yolov8n.pt")
                detections = []
                for img in frames:
                    results = yolo(np.array(img))
                    for r in results:
                        for c in r.boxes.cls:
                            detections.append(yolo.model.names[int(c)])
                detection_summary = {}
                for obj in detections:
                    detection_summary[obj] = detection_summary.get(obj, 0) + 1
            except Exception as e:
                logging.error(f"YOLOv8 error: {e}")
                detection_summary = {}
            # Image captioning (BLIP)
            try:
                from transformers import BlipProcessor, BlipForConditionalGeneration
                processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
                model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base")
                captions = []
                for img in frames:
                    inputs = processor(img, return_tensors="pt")
                    out = model.generate(**inputs)
                    captions.append(processor.decode(out[0], skip_special_tokens=True))
            except Exception as e:
                logging.error(f"BLIP error: {e}")
                captions = []
            context = f"Transcript: {transcript}\nCaptions: {' | '.join(captions)}\nDetections: {detection_summary}"
            answer = extractive_qa(question, context)
            return answer
    except Exception as e:
        logging.error(f"YouTube video QA error: {e}")
        return f"Video analysis error: {e}"

TOOL_REGISTRY = {
    "llama3_chat": llama3_chat,
    "mixtral_chat": mixtral_chat,
    "extractive_qa": extractive_qa,
    "table_qa": table_qa,
    "asr_transcribe": asr_transcribe,
    "image_caption": image_caption,
    "code_analysis": code_analysis,
    "youtube_video_qa": youtube_video_qa,
}

class ModularGAIAAgent:
    def __init__(self, api_url=DEFAULT_API_URL, tool_registry=TOOL_REGISTRY):
        self.api_url = api_url
        self.tools = tool_registry
        self.reasoning_trace = []
        self.file_cache = set(os.listdir('.'))

    def fetch_questions(self, from_api=True, questions_path="Hugging Face Questions"):
        if from_api:
            r = requests.get(f"{self.api_url}/questions")
            r.raise_for_status()
            return r.json()
        else:
            with open(questions_path) as f:
                data = f.read()
            start = data.find("[")
            end = data.rfind("]") + 1
            questions = json.loads(data[start:end])
            return questions

    def download_file(self, file_id, file_name=None):
        if not file_name:
            file_name = file_id
        if file_name in self.file_cache:
            return file_name
        url = f"{self.api_url}/files/{file_id}"
        r = requests.get(url)
        if r.status_code == 200:
            with open(file_name, "wb") as f:
                f.write(r.content)
            self.file_cache.add(file_name)
            return file_name
        else:
            self.reasoning_trace.append(f"Failed to download file {file_id} (status {r.status_code})")
            return None

    def detect_file_type(self, file_name):
        ext = os.path.splitext(file_name)[-1].lower()
        if ext in ['.mp3', '.wav', '.flac']:
            return 'audio'
        elif ext in ['.png', '.jpg', '.jpeg', '.bmp']:
            return 'image'
        elif ext in ['.py']:
            return 'code'
        elif ext in ['.xlsx']:
            return 'excel'
        elif ext in ['.csv']:
            return 'csv'
        elif ext in ['.json']:
            return 'json'
        elif ext in ['.txt', '.md']:
            return 'text'
        else:
            return 'unknown'

    def analyze_file(self, file_name, file_type):
        if file_type == 'audio':
            transcript = self.tools['asr_transcribe'](file_name)
            self.reasoning_trace.append(f"Transcribed audio: {transcript[:100]}...")
            return transcript
        elif file_type == 'image':
            caption = self.tools['image_caption'](file_name)
            self.reasoning_trace.append(f"Image caption: {caption}")
            return caption
        elif file_type == 'code':
            result = self.tools['code_analysis'](file_name)
            self.reasoning_trace.append(f"Code analysis result: {result}")
            return result
        elif file_type == 'excel':
            wb = openpyxl.load_workbook(file_name)
            ws = wb.active
            data = list(ws.values)
            headers = data[0]
            table = [dict(zip(headers, row)) for row in data[1:]]
            self.reasoning_trace.append(f"Excel table loaded: {table[:2]}...")
            return table
        elif file_type == 'csv':
            df = pd.read_csv(file_name)
            table = df.to_dict(orient='records')
            self.reasoning_trace.append(f"CSV table loaded: {table[:2]}...")
            return table
        elif file_type == 'json':
            with open(file_name) as f:
                data = json.load(f)
            self.reasoning_trace.append(f"JSON loaded: {str(data)[:100]}...")
            return data
        elif file_type == 'text':
            with open(file_name) as f:
                text = f.read()
            self.reasoning_trace.append(f"Text loaded: {text[:100]}...")
            return text
        else:
            self.reasoning_trace.append(f"Unknown file type: {file_name}")
            return None

    def answer_question(self, question_obj):
        self.reasoning_trace = []
        q = question_obj["question"]
        file_name = question_obj.get("file_name", "")
        file_content = None
        file_type = None
        # YouTube video question detection
        if "youtube.com" in q or "youtu.be" in q:
            url = None
            for word in q.split():
                if "youtube.com" in word or "youtu.be" in word:
                    url = word.strip().strip(',')
                    break
            if url:
                answer = self.tools['youtube_video_qa'](url, q)
                self.reasoning_trace.append(f"YouTube video analyzed: {url}")
                self.reasoning_trace.append(f"Final answer: {answer}")
                return self.format_answer(answer), self.reasoning_trace
        if file_name:
            file_id = file_name.split('.')[0]
            local_file = self.download_file(file_id, file_name)
            if local_file:
                file_type = self.detect_file_type(local_file)
                file_content = self.analyze_file(local_file, file_type)
        # Plan: choose tool based on question and file
        if file_type == 'audio' or file_type == 'text':
            if file_content:
                answer = self.tools['extractive_qa'](q, file_content)
            else:
                answer = self.tools['llama3_chat'](q)
        elif file_type == 'excel' or file_type == 'csv':
            if file_content:
                answer = self.tools['table_qa'](q, file_content)
            else:
                answer = self.tools['llama3_chat'](q)
        elif file_type == 'image':
            if file_content:
                answer = self.tools['llama3_chat'](f"{q}\nImage description: {file_content}")
            else:
                answer = self.tools['llama3_chat'](q)
        elif file_type == 'code':
            answer = file_content
        else:
            answer = self.tools['llama3_chat'](q)
        self.reasoning_trace.append(f"Final answer: {answer}")
        return self.format_answer(answer), self.reasoning_trace

    def format_answer(self, answer):
        if isinstance(answer, str):
            answer = answer.strip().rstrip('.')
            for prefix in ['answer:', 'result:', 'the answer is', 'final answer:', 'response:']:
                if answer.lower().startswith(prefix):
                    answer = answer[len(prefix):].strip()
            import re
            answer = re.sub(r'\b(the|a|an)\b ', '', answer, flags=re.IGNORECASE)
            answer = answer.strip().rstrip('.')
        return answer

# --- Basic Agent Definition (now wraps ModularGAIAAgent) ---
class BasicAgent:
    def __init__(self):
        print("BasicAgent (GAIA Modular Agent) initialized.")
        self.agent = ModularGAIAAgent()
    def __call__(self, question: str, file_name: str = "") -> str:
        print(f"Agent received question (first 50 chars): {question[:50]}...")
        try:
            answer, trace = self.agent.answer_question({"task_id": "manual", "question": question, "file_name": file_name})
            print(f"Agent returning answer: {answer}")
            return answer
        except Exception as e:
            print(f"Agent error: {e}")
            return f"AGENT ERROR: {e}"

def run_and_submit_all(profile: gr.OAuthProfile | None):
    """
    Fetches all questions, runs the BasicAgent on them, submits all answers,
    and displays the results.
    """
    space_id = os.getenv("SPACE_ID")
    if profile:
        username = f"{profile.username}"
        print(f"User logged in: {username}")
    else:
        print("User not logged in.")
        return "Please Login to Hugging Face with the button.", None
    api_url = DEFAULT_API_URL
    questions_url = f"{api_url}/questions"
    submit_url = f"{api_url}/submit"
    try:
        agent = BasicAgent()
    except Exception as e:
        print(f"Error instantiating agent: {e}")
        return f"Error initializing agent: {e}", None
    agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
    print(agent_code)
    print(f"Fetching questions from: {questions_url}")
    try:
        response = requests.get(questions_url, timeout=15)
        response.raise_for_status()
        questions_data = response.json()
        if not questions_data:
            print("Fetched questions list is empty.")
            return "Fetched questions list is empty or invalid format.", None
        print(f"Fetched {len(questions_data)} questions.")
    except requests.exceptions.RequestException as e:
        print(f"Error fetching questions: {e}")
        return f"Error fetching questions: {e}", None
    except requests.exceptions.JSONDecodeError as e:
        print(f"Error decoding JSON response from questions endpoint: {e}")
        print(f"Response text: {response.text[:500]}")
        return f"Error decoding server response for questions: {e}", None
    except Exception as e:
        print(f"An unexpected error occurred fetching questions: {e}")
        return f"An unexpected error occurred fetching questions: {e}", None
    results_log = []
    answers_payload = []
    print(f"Running agent on {len(questions_data)} questions...")
    for item in questions_data:
        task_id = item.get("task_id")
        question_text = item.get("question")
        file_name = item.get("file_name", "")
        if not task_id or question_text is None:
            print(f"Skipping item with missing task_id or question: {item}")
            continue
        try:
            submitted_answer = agent(question_text, file_name)
            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)
    submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload}
    status_update = f"Agent finished. Submitting {len(answers_payload)} answers for user '{username}'..."
    print(status_update)
    print(f"Submitting {len(answers_payload)} answers to: {submit_url}")
    try:
        response = requests.post(submit_url, json=submission_data, timeout=60)
        response.raise_for_status()
        result_data = response.json()
        final_status = (
            f"Submission Successful!\n"
            f"User: {result_data.get('username')}\n"
            f"Overall Score: {result_data.get('score', 'N/A')}% "
            f"({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')} correct)\n"
            f"Message: {result_data.get('message', 'No message received.')}")
        print("Submission successful.")
        results_df = pd.DataFrame(results_log)
        return final_status, results_df
    except requests.exceptions.HTTPError as e:
        error_detail = f"Server responded with status {e.response.status_code}."
        try:
            error_json = e.response.json()
            error_detail += f" Detail: {error_json.get('detail', e.response.text)}"
        except requests.exceptions.JSONDecodeError:
            error_detail += f" Response: {e.response.text[:500]}"
        status_message = f"Submission Failed: {error_detail}"
        print(status_message)
        results_df = pd.DataFrame(results_log)
        return status_message, results_df
    except requests.exceptions.Timeout:
        status_message = "Submission Failed: The request timed out."
        print(status_message)
        results_df = pd.DataFrame(results_log)
        return status_message, results_df
    except requests.exceptions.RequestException as e:
        status_message = f"Submission Failed: Network error - {e}"
        print(status_message)
        results_df = pd.DataFrame(results_log)
        return status_message, results_df
    except Exception as e:
        status_message = f"An unexpected error occurred during submission: {e}"
        print(status_message)
        results_df = pd.DataFrame(results_log)
        return status_message, results_df

# --- Build Gradio Interface using Blocks ---
with gr.Blocks() as demo:
    gr.Markdown("# Basic Agent Evaluation Runner")
    gr.Markdown(
        """
        **Instructions:**

        1.  Please clone this space, then modify the code to define your agent's logic, the tools, the necessary packages, etc ...
        2.  Log in to your Hugging Face account using the button below. This uses your HF username for submission.
        3.  Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score.

        ---
        **Disclaimers:**
        Once clicking on the "submit button, it can take quite some time ( this is the time for the agent to go through all the questions).
        This space provides a basic setup and is intentionally sub-optimal to encourage you to develop your own, more robust solution. For instance for the delay process of the submit button, a solution could be to cache the answers and submit in a seperate action or even to answer the questions in async.
        """
    )

    gr.LoginButton()

    run_button = gr.Button("Run Evaluation & Submit All Answers")

    status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False)
    # Removed max_rows=10 from DataFrame constructor
    results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True)

    run_button.click(
        fn=run_and_submit_all,
        outputs=[status_output, results_table]
    )

if __name__ == "__main__":
    print("\n" + "-"*30 + " App Starting " + "-"*30)
    # Check for SPACE_HOST and SPACE_ID at startup for information
    space_host_startup = os.getenv("SPACE_HOST")
    space_id_startup = os.getenv("SPACE_ID") # Get SPACE_ID at startup

    if space_host_startup:
        print(f"✅ SPACE_HOST found: {space_host_startup}")
        print(f"   Runtime URL should be: https://{space_host_startup}.hf.space")
    else:
        print("ℹ️  SPACE_HOST environment variable not found (running locally?).")

    if space_id_startup: # Print repo URLs if SPACE_ID is found
        print(f"✅ SPACE_ID found: {space_id_startup}")
        print(f"   Repo URL: https://huggingface.co/spaces/{space_id_startup}")
        print(f"   Repo Tree URL: https://huggingface.co/spaces/{space_id_startup}/tree/main")
    else:
        print("ℹ️  SPACE_ID environment variable not found (running locally?). Repo URL cannot be determined.")

    print("-"*(60 + len(" App Starting ")) + "\n")

    print("Launching Gradio Interface for Basic Agent Evaluation...")
    demo.launch(debug=True, share=False)