auxiliary_fns.py

import os
import requests
import subprocess
import pandas as pd
from PIL import Image
from io import BytesIO
import soundfile as sf

DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
IMAGE_FILES = ["png", "jpg", "tiff", "jpeg", "bmp"]
AUDIO_FILES = ["wav", "mp3", "aac", "ogg"]
TABULAR_FILES = ["csv", "xlsx"]

def read_audio_file(audio_bytes, file_extension):
    """
    Reads audio data from in-memory bytes.

    Args:
        audio_bytes (bytes): The audio data as bytes.
        file_extension (str): The extension of the audio file (e.g., 'wav', 'mp3').
    """
    try:
        audio_buffer = BytesIO(audio_bytes)
        format_string = file_extension.lower()
        data, samplerate = sf.read(audio_buffer, format=format_string)

        return (data, samplerate)

    except sf.LibsndfileError:
        print(f"Error: Could not read the audio data from memory with the specified format: {file_extension}")
    except Exception as e:
        print(f"An unexpected error occurred: {e}")


def read_tabular_data(file_bytes, file_extension):
    file_bytes.seek(0)
    if file_extension == "csv":
        return (pd.read_csv(file_bytes))
    elif file_extension == "xlsx":
        return (pd.read_excel(file_bytes))


def read_image_data(file_bytes, file_extension):
    return Image.open(file_bytes)


def write_and_execute_file(text):
    with open(f"file_to_execute.{file_extension}", "wb") as f:
        f.write(text)
    result = subprocess.run(['python', 'file_to_execute.py'], capture_output=True, text=True, check=True)
    return result.stdout


def file_handler(task_id, file_name):
    response = requests.get(f"{DEFAULT_API_URL}/files/{task_id}")
    response.raise_for_status()
    data = response.content

    ext = file_name.split(".")[-1]
    if ext in AUDIO_FILES:
        file_data = read_audio_file(data, ext)
    elif ext in TABULAR_FILES:
        file_data = read_tabular_file(data, ext)
    elif ext in IMAGE_FILES:
        file_data = read_image_file(data, ext)
    elif ext == "py":
        file_data = (data, ext)

    return file_data, ext