status_checker.py

import os
from pathlib import Path
from threading import Thread
import gradio as gr
from huggingface_hub import upload_folder, HfFolder, delete_repo
import sys
import time
import subprocess


HfFolder().save_token(os.getenv("HF_TOKEN"))

output_dataset_id = "nateraw/asdf123"
# Where user will write outputs from their script
outputs_dir = Path('outputs')

process_pid = os.getenv('USER_SCRIPT_PID', None)
print(f"USER SCRIPT PID? - {process_pid} - type {type(process_pid)}")


def process_is_complete():
    # Attempt at solving the todo shown below
    p = subprocess.Popen(['ps', '-p', process_pid], stdout=subprocess.PIPE)
    out = p.communicate()[0].decode('utf-8').strip().split('\n')
    return len(out) == 1

    
# A file that marks that the script has finished running
# TODO - maybe check process like this instead of with done.txt: https://stackoverflow.com/a/2944076
output_file = Path('done.txt')

def status_checker():
    print("Waiting to find output_file to check if script is done running")
    while True:
        print(f"Process is complete? {process_is_complete()}")
        if output_file.exists():
            print("Found the output file - Uploading the outputs from the script")
            upload_folder(repo_id=output_dataset_id, folder_path=str(outputs_dir), path_in_repo='.', repo_type='dataset')
            print("Finished uploading outputs from script. Done now!")
            return
        print("Didn't find it...sleeping for 5 seconds.")
        time.sleep(5)

with gr.Blocks() as demo:
    gr.Markdown(Path('about.md').read_text())

if not output_file.exists():
    print("Didnt find output file on init, so starting thread to watch for it!")
    thread = Thread(target=status_checker, daemon=True)
    thread.start()

demo.launch()