import librosa
import os
import soundfile
from tqdm import tqdm, tqdm_notebook

base_dir = "./data_sakuramiko_senbetsu"
output_dir = "./plachta/VITS-fast-fine-tuning/custom_character_voice/sakuramiko"
all_dir = [f for f in os.listdir(base_dir) if not os.path.isfile(os.path.join(base_dir, f))]

file_list = []

skip_dir = ["301_dousa",
            "801_eng_suuji",
            "801_eng_jikan",
            "803_eng_others",
            "912_alphabet",
            "912_alphabet2",
            "913_web",
            "sample"]

total_file_write = 0

def recursive_til_audio_file_found(path):
    listed_dir = [f for f in os.listdir(path)]
    if len(listed_dir) == 0:
        return
    test_path_first = os.path.join(path, listed_dir[0])
    
    # continue through the directory if not a file
    if not os.path.isfile(test_path_first):
        for next_dir in listed_dir:
            next_path = os.path.join(path, next_dir)
            # skip any directory specify in skip_dir
            for skip in skip_dir:
                if next_path.find(skip) != -1:
                    break
            else:
                recursive_til_audio_file_found(next_path)
        return

    #for new_dir in tqdm_notebook(listed_dir, desc=f"Processing : {path}"):
    for new_dir in listed_dir:
        new_path = os.path.join(path, new_dir)
        
        #if it is file, convert the audio to 16k and write to output directory
#        output_path_base = path.replace(base_dir, output_dir)
#        if not os.path.exists(output_path_base):
#            os.makedirs(output_path_base, exist_ok=True)

        # not an audio file
        if new_path.find(".wav") == -1 and new_path.find(".mp3") == -1:
            continue

        global total_file_write
#        audio, rate = librosa.load(new_path, sr=16000)
        audio, rate = librosa.load(new_path, sr=22050)
#        output_path = os.path.join(output_path_base, new_dir)
        output_path = os.path.join(output_dir, "sakuramiko_" + str(total_file_write) + ".wav")
#        output_path = os.path.join(output_dir, new_dir[0:-4] + ".wav")
        soundfile.write(output_path, audio, rate, format='wav', subtype="PCM_16")
        file_list.append(new_dir)
        
        total_file_write += 1
        pbar.update(1)
        #print(f"\rWrite file{output_path}", end="")
        
with tqdm(total=24778) as pbar:
    recursive_til_audio_file_found(base_dir)
print(f"Total audio file written : {total_file_write}")

import json
out_json = {}
for val in file_list:
    out_json[val] = {"path":val, "kana":""}
    
with open("./amitaro.json", "w") as outfile:
    outfile.write(json.dumps(out_json))