parser_example.py

import sounddevice as sd
import numpy as np
import librosa
import os

# This script serves as example to parsing the txt alignments. It will iterate over all 
# utterances of the dataset, split utterances on silences that are longer 0.4s and play them.
# (this does not change your files)

librispeech_root = "E:/Datasets/Librispeech"    # Replace with yours

def split_on_silences(audio_fpath, words, end_times):
    # Load the audio waveform
    sample_rate = 16000     # Sampling rate of LibriSpeech 
    wav, _ = librosa.load(audio_fpath, sample_rate)
    
    words = np.array(words)
    start_times = np.array([0.0] + end_times[:-1])
    end_times = np.array(end_times)
    assert len(words) == len(end_times) == len(start_times)
    assert words[0] == '' and words[-1] == ''
    
    # Break the sentence on pauses that are longer than 0.4s
    mask = (words == '') & (end_times - start_times >= 0.4)
    mask[0] = mask[-1] = True
    breaks = np.where(mask)[0]
    segment_times = [[end_times[s], start_times[e]] for s, e in zip(breaks[:-1], breaks[1:])]
    segment_times = (np.array(segment_times) * sample_rate).astype(np.int)
    wavs = [wav[segment_time[0]:segment_time[1]] for segment_time in segment_times]
    texts = [' '.join(words[s + 1:e]).replace("  ", " ") for s, e in zip(breaks[:-1], breaks[1:])]
    
    # Play the audio segments
    print("From %s" % audio_fpath)
    if len(wavs) > 1:
        print("This sentence was split in %d segments:" % len(wavs))
    else:
        print("There are no silences long enough for this sentence to be split:")
    for wav, text in zip(wavs, texts):
        # Pad the waveform with 1 second of silence because sounddevice tends to cut them early
        # when playing them. You shouldn't need to do that in your parsers.
        wav = np.concatenate((wav, [0] * sample_rate))
        print("\t%s" % text)
        sd.play(wav, sample_rate, blocking=True)
    print("")
    
    return wavs, texts

if __name__ == '__main__':
    # Select sets (e.g. dev-clean, train-other-500, ...)
    for set_name in os.listdir(librispeech_root):
        set_dir = os.path.join(librispeech_root, set_name)
        if not os.path.isdir(set_dir):
            continue
        
        # Select speakers
        for speaker_id in os.listdir(set_dir):
            speaker_dir = os.path.join(set_dir, speaker_id)
            
            # Select books
            for book_id in os.listdir(speaker_dir):
                book_dir = os.path.join(speaker_dir, book_id)
                
                # Get the alignment file
                alignment_fpath = os.path.join(book_dir, "%s-%s.alignment.txt" % 
                                               (speaker_id, book_id))
                if not os.path.exists(alignment_fpath):
                    raise Exception("Alignment file not found. Did you download and merge the txt "
                                    "alignments with your LibriSpeech dataset?")
                
                # Parse each utterance present in the file
                alignment_file = open(alignment_fpath, "r")
                for line in alignment_file:
                    
                    # Retrieve the utterance id, the words as a list and the end_times as a list
                    utterance_id, words, end_times = line.strip().split(' ')
                    words = words.replace('\"', '').split(',')
                    end_times = [float(e) for e in end_times.replace('\"', '').split(',')]
                    audio_fpath = os.path.join(book_dir, utterance_id + '.flac')
                    
                    # Split utterances on silences
                    wavs, texts = split_on_silences(audio_fpath, words, end_times)
                    
                alignment_file.close()