import gradio as gr
import os
os.system('cd monotonic_align && python setup.py build_ext --inplace && cd ..')

import json
import math
import torch
from torch import nn
from torch.nn import functional as F
from torch.utils.data import DataLoader

import commons
import utils
from data_utils import TextAudioLoader, TextAudioCollate, TextAudioSpeakerLoader, TextAudioSpeakerCollate
from models import SynthesizerTrn
from text.symbols import symbols
from text import text_to_sequence

from scipy.io.wavfile import write

def get_text(text, hps):
    text_norm = text_to_sequence(text, hps.data.text_cleaners)
    if hps.data.add_blank:
        text_norm = commons.intersperse(text_norm, 0)
    text_norm = torch.LongTensor(text_norm)
    return text_norm

def load_model(model_path, hps):
    net_g = SynthesizerTrn(
        len(symbols),
        hps.data.filter_length // 2 + 1,
        hps.train.segment_size // hps.data.hop_length,
        n_speakers=hps.data.n_speakers,
        **hps.model)
    _ = net_g.eval()
    _ = utils.load_checkpoint(model_path, net_g, None)
    return net_g

hps = utils.get_hparams_from_file("configs/vctk_base.json")

# Define a dictionary to store the model paths
model_paths = {
    "Model 1": "fr_wa_finetuned_pho/G_125000.pth",
    "Model 2": "fr_wa_finetuned/G_198000.pth",
    "Model 3": "path_to_model_3_checkpoint.pth",
    "Model 4": "path_to_model_4_checkpoint.pth"
}

# Load the initial model
net_g = load_model(model_paths["Model 1"], hps)

def tts(text, speaker_id, model_choice):
    global net_g
    net_g = load_model(model_paths[model_choice], hps)
    
    if len(text) > 2000:
        return "Error: Text is too long", None
    sid = torch.LongTensor([speaker_id])  # speaker identity
    stn_tst = get_text(text, hps)

    with torch.no_grad():
        x_tst = stn_tst.unsqueeze(0)
        x_tst_lengths = torch.LongTensor([stn_tst.size(0)])
        audio = net_g.infer(x_tst, x_tst_lengths, sid=sid, noise_scale=.667, noise_scale_w=0.8, length_scale=1)[0][
            0, 0].data.float().numpy()
    return "Success", (hps.data.sampling_rate, audio)

app = gr.Blocks()
with app:
    with gr.Tabs():
        for tab_name in ["Phonemes_finetuned", "Graphemes_finetuned", "Phonemes", "Graphemes"]:
            with gr.TabItem(tab_name):
                tts_input1 = gr.TextArea(label="Text in Walloon in phonemes IPA (2000 words limitation)", value="")
                tts_input2 = gr.Dropdown(label="Speaker", choices=["Male", "Female"], type="index", value="Male")
                model_choice = gr.Dropdown(label="Model", choices=list(model_paths.keys()), value="Model 1")
                tts_submit = gr.Button("Generate", variant="primary")
                tts_output1 = gr.Textbox(label="Message")
                tts_output2 = gr.Audio(label="Output")
                tts_submit.click(tts, [tts_input1, tts_input2, model_choice], [tts_output1, tts_output2])

    app.launch()