data_utils.py

import time
import os
import random
import numpy as np
import torch
import torch.utils.data

import commons
from mel_processing import spectrogram_torch
from utils import load_wav_to_torch, load_unit_audio_pairs


class UnitAudioLoader(torch.utils.data.Dataset):
    '''
        1) loads audio and speech units
        2) compute spectrograms
    '''

    def __init__(self, unit_audio_pairs, hparams, train=True):
        self.unit_audio_pairs = load_unit_audio_pairs(unit_audio_pairs)
        self.max_wav_value = hparams.max_wav_value
        self.sampling_rate = hparams.sampling_rate
        self.filter_length = hparams.filter_length
        self.hop_length = hparams.hop_length
        self.win_length = hparams.win_length
        self.sampling_rate = hparams.sampling_rate
        random.seed(1234)
        random.shuffle(self.unit_audio_pairs)
        self._filter()

    def _filter(self):
        lengths = []
        for audio_path, _ in self.unit_audio_pairs:
            lengths.append(os.path.getsize(audio_path) // (2 * self.hop_length))
        self.lengths = lengths

    def get_unit_audio_pair(self, unit_audio_pairs):
        audio_path, unit_path = unit_audio_pairs[0], unit_audio_pairs[1]
        unit = np.load(unit_path)
        unit = torch.FloatTensor(unit)
        # unit = torch.LongTensor(unit)
        spec, wav = self.get_audio(audio_path)
        return (unit, spec, wav)

    def get_audio(self, filename):
        audio, sampling_rate = load_wav_to_torch(filename)
        if sampling_rate != self.sampling_rate:
            raise ValueError("{} {} SR doesn't match target {} SR".format(
                sampling_rate, self.sampling_rate))
        audio_norm = audio / self.max_wav_value
        audio_norm = audio_norm.unsqueeze(0)
        spec_filename = filename.replace(".wav", ".spec.pt")
        if os.path.exists(spec_filename):
            spec = torch.load(spec_filename)
        else:
            spec = spectrogram_torch(audio_norm, self.filter_length,
                                     self.sampling_rate, self.hop_length, self.win_length,
                                     center=False)
            spec = torch.squeeze(spec, 0)
            torch.save(spec, spec_filename)
        return spec, audio_norm

    def __getitem__(self, index):
        return self.get_unit_audio_pair(self.unit_audio_pairs[index])

    def __len__(self):
        return len(self.unit_audio_pairs)


class UnitAudioCollate():
    def __init__(self, return_ids=False):
        self.return_ids = return_ids

    def __call__(self, batch):
        """Collate's training batch from normalized text and aduio
       PARAMS
       ------
       batch: [unit, spec_normalized, wav_normalized]
       """
        # Right zero-pad all one-hot text sequences to max input length
        _, ids_sorted_decreasing = torch.sort(
            torch.LongTensor([x[1].size(1) for x in batch]),
            dim=0, descending=True)

        max_unit_len = max([len(x[0]) for x in batch])
        max_spec_len = max([x[1].size(1) for x in batch])
        max_wav_len = max([x[2].size(1) for x in batch])

        unit_lengths = torch.LongTensor(len(batch))
        spec_lengths = torch.LongTensor(len(batch))
        wav_lengths = torch.LongTensor(len(batch))

        unit_padded = torch.FloatTensor(len(batch), max_unit_len, 256)
        spec_padded = torch.FloatTensor(len(batch), batch[0][1].size(0), max_spec_len)
        wav_padded = torch.FloatTensor(len(batch), 1, max_wav_len)
        unit_padded.zero_()
        spec_padded.zero_()
        wav_padded.zero_()
        for i in range(len(ids_sorted_decreasing)):
            row = batch[ids_sorted_decreasing[i]]

            unit = row[0]
            unit_padded[i, :unit.size(0)] = unit
            unit_lengths[i] = unit.size(0)

            spec = row[1]
            spec_padded[i, :, :spec.size(1)] = spec
            spec_lengths[i] = spec.size(1)

            wav = row[2]
            wav_padded[i, :, :wav.size(1)] = wav
            wav_lengths[i] = wav.size(1)

        if self.return_ids:
            return unit_padded, unit_lengths, spec_padded, spec_lengths, wav_padded, wav_lengths, ids_sorted_decreasing
        return unit_padded, unit_lengths, spec_padded, spec_lengths, wav_padded, wav_lengths

"""Multi speaker version"""
class UnitAudioSpeakerLoader(torch.utils.data.Dataset):
    """
        1) loads audio, speaker_id, speech unit pairs
        2) computes spectrograms from audio files.
    """
    def __init__(self, unit_sid_audio_pairs, hparams):
        self.unit_sid_audio_pairs = load_unit_audio_pairs(unit_sid_audio_pairs)
        self.max_wav_value = hparams.max_wav_value
        self.sampling_rate = hparams.sampling_rate
        self.filter_length  = hparams.filter_length
        self.hop_length     = hparams.hop_length
        self.win_length     = hparams.win_length
        self.sampling_rate  = hparams.sampling_rate

        random.seed(1234)
        random.shuffle(self.unit_sid_audio_pairs)
        self._filter()

    def _filter(self):
        lengths = []
        for audio_path, _, _ in self.unit_sid_audio_pairs:
            lengths.append(os.path.getsize(audio_path) // (2 * self.hop_length))
        self.lengths = lengths

    def get_unit_sid_audio_pair(self, unit_sid_audio_pair):
        # separate filename, speaker_id and text
        audio_path, sid, unit_path = unit_sid_audio_pair[0], unit_sid_audio_pair[1], unit_sid_audio_pair[2]
        unit = np.load(unit_path)
        unit = torch.FloatTensor(unit)
        # unit = torch.LongTensor(unit)
        spec, wav = self.get_audio(audio_path)
        sid = self.get_sid(sid)
        return (unit, spec, wav, sid)

    def get_audio(self, filename):
        audio, sampling_rate = load_wav_to_torch(filename)
        if sampling_rate != self.sampling_rate:
            raise ValueError("{} SR doesn't match target {} SR".format(
                sampling_rate, self.sampling_rate))
        audio_norm = audio / self.max_wav_value
        audio_norm = audio_norm.unsqueeze(0)
        spec_filename = filename.replace(".wav", ".spec.pt")
        if os.path.exists(spec_filename):
            spec = torch.load(spec_filename)
        else:
            spec = spectrogram_torch(audio_norm, self.filter_length,
                self.sampling_rate, self.hop_length, self.win_length,
                center=False)
            spec = torch.squeeze(spec, 0)
            torch.save(spec, spec_filename)
        return spec, audio_norm

    def get_sid(self, sid):
        sid = torch.LongTensor([int(sid)])
        return sid

    def __getitem__(self, index):
        return self.get_unit_sid_audio_pair(self.unit_sid_audio_pairs[index])

    def __len__(self):
        return len(self.unit_sid_audio_pairs)

class UnitAudioSpeakerCollate():
    """ Zero-pads model inputs and targets
    """
    def __init__(self, return_ids=False):
        self.return_ids = return_ids

    def __call__(self, batch):
        """Collate's training batch from normalized text, audio and speaker identities
        PARAMS
        ------
        batch: [unit, spec_normalized, wav_normalized, sid]
        """
        # Right zero-pad all one-hot text sequences to max input length
        _, ids_sorted_decreasing = torch.sort(
            torch.LongTensor([x[1].size(1) for x in batch]),
            dim=0, descending=True)

        max_unit_len = max([len(x[0]) for x in batch])
        max_spec_len = max([x[1].size(1) for x in batch])
        max_wav_len = max([x[2].size(1) for x in batch])

        unit_lengths = torch.LongTensor(len(batch))
        spec_lengths = torch.LongTensor(len(batch))
        wav_lengths = torch.LongTensor(len(batch))
        sid = torch.LongTensor(len(batch))

        unit_padded = torch.FloatTensor(len(batch), max_unit_len, 256)
        spec_padded = torch.FloatTensor(len(batch), batch[0][1].size(0), max_spec_len)
        wav_padded = torch.FloatTensor(len(batch), 1, max_wav_len)
        unit_padded.zero_()
        spec_padded.zero_()
        wav_padded.zero_()
        for i in range(len(ids_sorted_decreasing)):
            row = batch[ids_sorted_decreasing[i]]

            unit = row[0]
            unit_padded[i, :unit.size(0)] = unit
            unit_lengths[i] = unit.size(0)

            spec = row[1]
            spec_padded[i, :, :spec.size(1)] = spec
            spec_lengths[i] = spec.size(1)

            wav = row[2]
            wav_padded[i, :, :wav.size(1)] = wav
            wav_lengths[i] = wav.size(1)

            sid[i] = row[3]

        if self.return_ids:
            return unit_padded, unit_lengths, spec_padded, spec_lengths, wav_padded, wav_lengths, sid, ids_sorted_decreasing
        return unit_padded, unit_lengths, spec_padded, spec_lengths, wav_padded, wav_lengths, sid

class DistributedBucketSampler(torch.utils.data.distributed.DistributedSampler):
    """
    Maintain similar input lengths in a batch.
    Length groups are specified by boundaries.
    Ex) boundaries = [b1, b2, b3] -> any batch is included either {x | b1 < length(x) <=b2} or {x | b2 < length(x) <= b3}.

    It removes samples which are not included in the boundaries.
    Ex) boundaries = [b1, b2, b3] -> any x s.t. length(x) <= b1 or length(x) > b3 are discarded.
    """

    def __init__(self, dataset, batch_size, boundaries, num_replicas=None, rank=None, shuffle=True):
        super().__init__(dataset, num_replicas=num_replicas, rank=rank, shuffle=shuffle)
        self.lengths = dataset.lengths
        self.batch_size = batch_size
        self.boundaries = boundaries

        self.buckets, self.num_samples_per_bucket = self._create_buckets()
        self.total_size = sum(self.num_samples_per_bucket)
        self.num_samples = self.total_size // self.num_replicas

    def _create_buckets(self):
        buckets = [[] for _ in range(len(self.boundaries) - 1)]
        for i in range(len(self.lengths)):
            length = self.lengths[i]
            idx_bucket = self._bisect(length)
            if idx_bucket != -1:
                buckets[idx_bucket].append(i)

        for i in range(len(buckets) - 1, 0, -1):
            if len(buckets[i]) == 0:
                buckets.pop(i)
                self.boundaries.pop(i + 1)

        num_samples_per_bucket = []
        for i in range(len(buckets)):
            len_bucket = len(buckets[i])
            total_batch_size = self.num_replicas * self.batch_size
            rem = (total_batch_size - (len_bucket % total_batch_size)) % total_batch_size
            num_samples_per_bucket.append(len_bucket + rem)
        return buckets, num_samples_per_bucket

    def __iter__(self):
        # deterministically shuffle based on epoch
        g = torch.Generator()
        g.manual_seed(self.epoch)

        indices = []
        if self.shuffle:
            for bucket in self.buckets:
                indices.append(torch.randperm(len(bucket), generator=g).tolist())
        else:
            for bucket in self.buckets:
                indices.append(list(range(len(bucket))))

        batches = []
        for i in range(len(self.buckets)):
            bucket = self.buckets[i]
            len_bucket = len(bucket)
            ids_bucket = indices[i]
            num_samples_bucket = self.num_samples_per_bucket[i]

            # add extra samples to make it evenly divisible
            rem = num_samples_bucket - len_bucket
            ids_bucket = ids_bucket + ids_bucket * (rem // len_bucket) + ids_bucket[:(rem % len_bucket)]

            # subsample
            ids_bucket = ids_bucket[self.rank::self.num_replicas]

            # batching
            for j in range(len(ids_bucket) // self.batch_size):
                batch = [bucket[idx] for idx in ids_bucket[j * self.batch_size:(j + 1) * self.batch_size]]
                batches.append(batch)

        if self.shuffle:
            batch_ids = torch.randperm(len(batches), generator=g).tolist()
            batches = [batches[i] for i in batch_ids]
        self.batches = batches

        assert len(self.batches) * self.batch_size == self.num_samples
        return iter(self.batches)

    def _bisect(self, x, lo=0, hi=None):
        if hi is None:
            hi = len(self.boundaries) - 1

        if hi > lo:
            mid = (hi + lo) // 2
            if self.boundaries[mid] < x and x <= self.boundaries[mid + 1]:
                return mid
            elif x <= self.boundaries[mid]:
                return self._bisect(x, lo, mid)
            else:
                return self._bisect(x, mid + 1, hi)
        else:
            return -1

    def __len__(self):
        return self.num_samples // self.batch_size