import os
import glob
import random
import torch
import numpy as np
from torch.utils import data
from suno_utils.audio import Audio
from torchaudio_augmentations import (
    RandomApply,
    Noise,
    Gain,
    PitchShift,
    Compose,
)


class LyricSimDataset(data.Dataset):
    def __init__(
        self,
        metadata,
        task_indices,
        split="train",
        input_length_s=30.0,
        input_text_length=300,
        sample_rate=24000,
        num_samples=-1,
    ):
        assert split in ["train", "valid"]
        self.metadata = metadata
        self.indices = task_indices["lyric_sim"][split]
        self.input_length_s = input_length_s
        self.input_text_length = input_text_length
        self.sample_rate = sample_rate
        self.num_samples = num_samples
        self.split = split

        # get augmentation
        if split == "train":
            self._get_augmentations()

        random.seed(207)
        random.shuffle(self.indices)

        print(f"{len(self.indices)} files are available for lyric_sim {split} set")

    def _get_augmentations(self):
        # Stochastic data augmentation
        transforms = [
            RandomApply([Noise(min_snr=0.1, max_snr=0.5)], p=0.3),
            RandomApply([Gain()], p=0.2),
            RandomApply(
                [
                    PitchShift(
                        n_samples=int(self.input_length_s * self.sample_rate),
                        sample_rate=self.sample_rate,
                        pitch_shift_min=-3.0,
                        pitch_shift_max=3.0,
                    )
                ],
                p=0.4,
            ),
        ]
        self.augmentation = Compose(transforms=transforms)

    def get_lyrics(self, index):
        lyrics = self.metadata[index]["lyrics"]

        if self.split == "train":
            if len(lyrics) > self.input_text_length:
                start_ix = random.randint(0, len(lyrics) - self.input_text_length - 1)
                lyrics = lyrics[start_ix : start_ix + self.input_text_length]
        elif self.split == "valid":
            lyrics = lyrics[: self.input_text_length]

        lyrics = "[CLS]" + "[Lyrics]" + lyrics

        return lyrics

    def __getitem__(self, index):
        # read data
        if self.split == "train":
            random_ix = random.choice(self.indices)
            audio_path = self.metadata[random_ix]["filepath"]
            lyrics = self.get_lyrics(random_ix)
        elif self.split == "valid":
            audio_path = self.metadata[self.indices[index]]["filepath"]
            lyrics = self.get_lyrics(self.indices[index])
        audio = Audio.from_file(audio_path, sample_rate=self.sample_rate)
        if self.split == "train":
            # random crop
            try:
                start_ms = random.randint(
                    0, (audio.duration_ms - int(1000 * self.input_length_s) - 1)
                )
                start_s = start_ms / 1000
            except ValueError:
                start_s = 0.0
            wav = audio.get_segment(
                from_s=start_s, to_s=start_s + self.input_length_s
            ).array_float

            # augmentation
            wav = (
                self.augmentation(torch.from_numpy(wav).unsqueeze(0)).squeeze(0).numpy()
            )
        elif self.split == "valid":
            # crop first and last 30s
            start_s = 0.0
            wav = audio.get_segment(
                from_s=start_s, to_s=start_s + self.input_length_s
            ).array_float

        # zero padding
        if len(wav) < int(self.sample_rate * self.input_length_s):
            pad = int(self.sample_rate * self.input_length_s) - len(wav)
            wav = np.pad(wav, (0, pad), mode="constant", constant_values=0)

        return wav, lyrics

    def __len__(self):
        if self.num_samples > 0:
            return self.num_samples
        else:
            return len(self.indices)