app.py

import argparse
from multiprocessing import cpu_count

import common
import config
import dub
import fandom
import finetune
import voice_fetch
import pathlib
from config import sources_to_fetch_voice, muted_characters, dataset_manifest_file_dest, source_text_to_dub, \
    dub_manifest_dest
import emotion
import importlib


def do_voice_collection():
    collections: dict[str, list[tuple[str, str]]] = {}
    
    for i in sources_to_fetch_voice:
        if i.startswith('custom:'):
            providerName = i[7:i.index(':', 7)]
            url = i[i.index(':')+1:]
            provider = importlib.import_module(f'customDataProviders.{providerName}')
            collections = fandom.merge_voice_collections([collections, provider.fetch_vo_urls(url, muted_characters)])
        else:
            quests = common.request_retry_wrapper(lambda: fandom.fetch_quest_entries(i))
            

            for quest in quests:
                collection = fandom.fetch_target_vo_from_quest_page(quest, muted_characters)

                collections = fandom.merge_voice_collections([collections, collection])

    # download voices
    voice_fetch.reduce_collection(collections)
    voice_fetch.fetch_collection(collections)

    # with open(dataset_manifest_file_dest, 'w') as f:
    #     f.write(voice_fetch.serialize_collection(collections))

    pathlib.Path(dataset_manifest_file_dest).write_text(voice_fetch.serialize_collection(collections))
    
    voice_fetch.generate_text_list()
    

def do_subtitle_collection():
    collections: dict[str, list[str]] = {}
    for i in source_text_to_dub:
        quests = [i[i.index(':')+1:]] if i.startswith('quest:') else fandom.fetch_quest_entries(i)
        for quest in quests:
            collection = fandom.fetch_target_subtitles(quest, muted_characters)
            collections = fandom.merge_subtitle_collections([collections, collection])

    # with open(dub_manifest_dest, 'w') as f:
    #     f.write(dub.serialize_collection(collections))
    pathlib.Path(dub_manifest_dest).write_text(dub.serialize_collection(collections))

def do_finetune():
    finetune.start()
    
    
def find_missing_voices():
    result = []
    for i in source_text_to_dub:
        quests = [i[i.index(':')+1:]] if i.startswith('quest:') else fandom.fetch_quest_entries(i)
        common.log(f'Checking for missing voices in {quests}')
        for quest in quests:
            r = fandom.find_potentially_missing_voice_over_chars(quest, muted_characters)
            result.extend(r)
    
    return [i for i in set(result)]

if __name__ == '__main__':
    # use argparse to parse command line arguments

    parser = argparse.ArgumentParser()
    parser.description = 'AI Dubbing for Genshin Impact Natlan Region due to the EN VA Strike'
    parser.add_argument('--voice', action='store_true', help='Fetch voices for dataset')
    parser.add_argument('--subtitle', action='store_true', help='Fetch subtitles for dubbing')
    parser.add_argument('--inference-server', action='store_true', help='Start GPT-SoVITs inference server for dubbing')
    parser.add_argument('--dub-all', action='store_true', help='Dub all the subtitles in the manifest file')
    parser.add_argument('--emotion-classification', action='store_true', help='Pre-process the audio files and generate emotion analsysis configuration for dubbing')
    parser.add_argument('--dataset-overview', action='store_true', help='Check the dataset overview')
    parser.add_argument('--missing-voices', action='store_true', help='Find potentially missing voices for the subtitles')
    parser.add_argument('--use-middleware-logic', action='store_true', help='Use middleware logic for initializing muted characters and model paths')
    args = parser.parse_args()
    global models_path
    global muted_characters
    
    if args.use_middleware_logic:
        models_path = common.get_available_model_path()
        muted_characters = common.get_muted_chars()
        common.log(f'Using middleware logic for models path: {models_path} and muted characters: {muted_characters}')
    
    if args.voice:
        do_voice_collection()
    elif args.subtitle:
        do_subtitle_collection()
    elif args.dub_all:
        dub.dub_all()
    elif args.inference_server:
        dub.run_gpt_sovits_api_server()
    elif args.emotion_classification:
        emotion.do_classification()
    elif args.dataset_overview:
        common.dataset_overview()
    elif args.missing_voices:
        common.log(f'Potentially missing voices: {find_missing_voices()}')
    else:
        parser.print_help()