LoadBalancer.py

import os
import json
from indexer import indexer
import re
from tvdb import fetch_and_cache_json
from threading import Event, Thread
import time
import logging
from utils import convert_to_gb
from api import InstancesAPI

CACHE_DIR = os.getenv("CACHE_DIR")

download_progress = {}

class LoadBalancer:
    def __init__(self, cache_dir, index_file, token, repo, polling_interval=4, max_retries=3, initial_delay=1):
        self.version = "0.0.2.10 V Beta"
        self.instances = []
        self.instances_health = {}
        self.polling_interval = polling_interval
        self.max_retries = max_retries
        self.initial_delay = initial_delay
        self.stop_event = Event()
        self.instances_api = InstancesAPI(self.instances)
        self.CACHE_DIR = cache_dir
        self.INDEX_FILE = index_file
        self.TOKEN = token
        self.REPO = repo
        self.FILM_STORE = {}
        self.TV_STORE = {}
        self.file_structure = None
        self.index_file_last_modified = None

        # Ensure CACHE_DIR exists
        if not os.path.exists(self.CACHE_DIR):
            os.makedirs(self.CACHE_DIR)

        # Index the file structure initially
        indexer()

        # Load the file structure JSON
        self.load_file_structure()

        # Start polling and file checking in separate threads
        polling_thread = Thread(target=self.start_polling)
        polling_thread.daemon = True
        polling_thread.start()

        file_checking_thread = Thread(target=self.check_file_updates)
        file_checking_thread.daemon = True
        file_checking_thread.start()

    def load_file_structure(self):
        if not os.path.exists(self.INDEX_FILE):
            raise FileNotFoundError(f"{self.INDEX_FILE} not found. Please make sure the file exists.")

        with open(self.INDEX_FILE, 'r') as f:
            self.file_structure = json.load(f)
        logging.info("File structure loaded successfully.")

    def check_file_updates(self):
        while not self.stop_event.is_set():
            if self.index_file_last_modified != os.path.getmtime(self.INDEX_FILE):
                logging.info(f"{self.INDEX_FILE} has been updated. Re-indexing...")
                indexer()  # Re-run the indexer
                self.load_file_structure()  # Reload the file structure
                self.index_file_last_modified = os.path.getmtime(self.INDEX_FILE)

                # Restart prefetching thread
                if hasattr(self, 'prefetch_thread') and self.prefetch_thread.is_alive():
                    self.prefetch_thread.join()
                
                self.prefetch_thread = Thread(target=self.start_prefetching)
                self.prefetch_thread.daemon = True
                self.prefetch_thread.start()

            time.sleep(120)  # Check every 2 minutes

    def register_instance(self, instance_url):
        if instance_url not in self.instances:
            self.instances.append(instance_url)
            logging.info(f"Registered instance {instance_url}")
        else:
            logging.info(f"Instance {instance_url} is already registered.")

    def remove_instance(self, instance_url):
        if instance_url in self.instances:
            self.instances.remove(instance_url)
            self.instances_health.pop(instance_url, None)
            logging.info(f"Removed instance {instance_url}")
        else:
            logging.info(f"Instance {instance_url} not found for removal.")
        
    def get_reports(self):
        reports = self.instances_api.fetch_reports()

        # Initialize temporary JSON data holders
        temp_film_store = {}
        temp_tv_store = {}

        for instance_url in self.instances[:]:  # Copy list to avoid modification during iteration
            if instance_url in reports:
                report = reports[instance_url]
                logging.info(f"Report from {instance_url}: {report}")
                self.process_report(instance_url, report, temp_film_store, temp_tv_store)
            else:
                logging.error(f"Failed to get report from {instance_url}. Removing instance.")
                self.remove_instance(instance_url)

        self.FILM_STORE = temp_film_store
        self.TV_STORE = temp_tv_store

    def process_report(self, instance_url, report, temp_film_store, temp_tv_store):
        film_store = report.get('film_store', {})
        tv_store = report.get('tv_store', {})
        cache_size = report.get('cache_size')

        logging.info(f"Processing report from {instance_url}")

        # Update temporary film store
        for title, path in film_store.items():
            url = f"{instance_url}/api/film/{title.replace(' ', '%20')}"
            temp_film_store[title] = url

        # Update temporary TV store
        for title, seasons in tv_store.items():
            if title not in temp_tv_store:
                temp_tv_store[title] = {}
            for season, episodes in seasons.items():
                if season not in temp_tv_store[title]:
                    temp_tv_store[title][season] = {}
                for episode, path in episodes.items():
                    url = f"{instance_url}/api/tv/{title.replace(' ', '%20')}/{season.replace(' ', '%20')}/{episode.replace(' ', '%20')}"
                    temp_tv_store[title][season][episode] = url

        logging.info("Film and TV Stores processed successfully.")
        self.update_instances_health(instance=instance_url, cache_size=cache_size)

    def start_polling(self):
        logging.info("Starting polling.")
        while not self.stop_event.is_set():
            self.get_reports()
            time.sleep(self.polling_interval)
        logging.info("Polling stopped.")

    def stop_polling(self):
        logging.info("Stopping polling.")
        self.stop_event.set()

    def start_prefetching(self):
        """Start the metadata prefetching in a separate thread."""
        self.prefetch_metadata()

#################################################################

    def update_instances_health(self, instance, cache_size):
        self.instances_health[instance] = {"used":cache_size["cache_size"],
                                           "total": "50 GB"}
        logging.info(f"Updated instance {instance} with cache size {cache_size}")


    def download_film_to_best_instance(self, title):
        """
        Downloads a film to the first instance that has more free space on the self.instance_health list variable. 
        The instance_health looks like this:
        {
            "https://unicone-studio-instance1.hf.space": {
                "total": "50 GB",
                "used": "3.33 GB"
            }
        }
        Args:
            title (str): The title of the film.
        """
        best_instance = None
        max_free_space = -1
        
        # Calculate free space for each instance
        for instance_url, space_info in self.instances_health.items():
            total_space = convert_to_gb(space_info['total'])
            used_space = convert_to_gb(space_info['used'])
            free_space = total_space - used_space
            
            if free_space > max_free_space:
                max_free_space = free_space
                best_instance = instance_url
        
        if best_instance:
            result = self.instances_api.download_film(best_instance, title)
            film_id = result["film_id"]
            status = result["status"]
            progress_url = f'{best_instance}/api/progress/{film_id}'
            response = {
                "film_id":film_id,
                "status":status,
                "progress_url":progress_url
            }

            return response
        else:
            logging.error("No suitable instance found for downloading the film.")
            return {"error": "No suitable instance found for downloading the film."}

    def download_episode_to_best_instance(self, title, season, episode):
        """
        Downloads a episode to the first instance that has more free space on the self.instance_health list variable. 
        The instance_health looks like this:
        {
            "https://unicone-studio-instance1.hf.space": {
                "total": "50 GB",
                "used": "3.33 GB"
            }
        }
        Args:
            title (str): The title of the Tv show.
            season (str): The season of the Tv show.
            episode (str): The title of the Tv show.
        """
        best_instance = None
        max_free_space = -1
        
        # Calculate free space for each instance
        for instance_url, space_info in self.instances_health.items():
            total_space = convert_to_gb(space_info['total'])
            used_space = convert_to_gb(space_info['used'])
            free_space = total_space - used_space
            
            if free_space > max_free_space:
                max_free_space = free_space
                best_instance = instance_url
        
        if best_instance:
            result = self.instances_api.download_episode(best_instance, title, season, episode)
            episode_id = result["episode_id"]
            status = result["status"]
            progress_url = f'{best_instance}/api/progress/{episode_id}'
            response = {
                "episode_id":episode_id,
                "status":status,
                "progress_url":progress_url
            }

            return response
        else:
            logging.error("No suitable instance found for downloading the film.")
            return {"error": "No suitable instance found for downloading the film."}

#################################################################
    def find_movie_path(self, title):
        """Find the path of the movie in the JSON data based on the title."""
        for directory in self.file_structure:
            if directory['type'] == 'directory' and directory['path'] == 'films':
                for sub_directory in directory['contents']:
                    if sub_directory['type'] == 'directory':
                        for item in sub_directory['contents']:
                            if item['type'] == 'file' and title.lower() in item['path'].lower():
                                return item['path']
        return None

    def find_tv_path(self, title):
        """Find the path of the TV show in the JSON data based on the title."""
        for directory in self.file_structure:
            if directory['type'] == 'directory' and directory['path'] == 'tv':
                for sub_directory in directory['contents']:
                    if sub_directory['type'] == 'directory' and title.lower() in sub_directory['path'].lower():
                        return sub_directory['path']
        return None

    def get_tv_structure(self, title):
        """Find the path of the TV show in the JSON data based on the title."""
        for directory in self.file_structure:
            if directory['type'] == 'directory' and directory['path'] == 'tv':
                for sub_directory in directory['contents']:
                    if sub_directory['type'] == 'directory' and title.lower() in sub_directory['path'].lower():
                        return sub_directory
        return None

    def get_film_id(self, title):
        """Generate a film ID based on the title."""
        return title.replace(" ", "_").lower()

    def prefetch_metadata(self):
        """Prefetch metadata for all items in the file structure."""
        for item in self.file_structure:
            if 'contents' in item:
                for sub_item in item['contents']:
                    original_title = sub_item['path'].split('/')[-1]
                    media_type = 'series' if item['path'].startswith('tv') else 'movie'
                    title = original_title
                    year = None

                    # Extract year from the title if available
                    match = re.search(r'\((\d{4})\)', original_title)
                    if match:
                        year_str = match.group(1)
                        if year_str.isdigit() and len(year_str) == 4:
                            title = original_title[:match.start()].strip()
                            year = int(year_str)
                    else:
                        parts = original_title.rsplit(' ', 1)
                        if len(parts) > 1 and parts[-1].isdigit() and len(parts[-1]) == 4:
                            title = parts[0].strip()
                            year = int(parts[-1])

                    fetch_and_cache_json(original_title, title, media_type, year)

    def get_all_tv_shows(self):
        """Get all TV shows from the indexed cache structure JSON file."""
        tv_shows = {}
        for directory in self.file_structure:
            if directory['type'] == 'directory' and directory['path'] == 'tv':
                for sub_directory in directory['contents']:
                    if sub_directory['type'] == 'directory':
                        show_title = sub_directory['path'].split('/')[-1]
                        tv_shows[show_title] = []
                        for season_directory in sub_directory['contents']:
                            if season_directory['type'] == 'directory':
                                season = season_directory['path'].split('/')[-1]
                                for episode in season_directory['contents']:
                                    if episode['type'] == 'file':
                                        tv_shows[show_title].append({
                                            "season": season,
                                            "episode": episode['path'].split('/')[-1],
                                            "path": episode['path']
                                        })
        return tv_shows

    def get_all_films(self):
        """Get all films from the indexed cache structure JSON file."""
        films = []
        for directory in self.file_structure:
            if directory['type'] == 'directory' and directory['path'] == 'films':
                for sub_directory in directory['contents']:
                    if sub_directory['type'] == 'directory':
                        films.append(sub_directory['path'])
        return films