Get list of phonemes from librispeech lexicon

[jerivl]

Issue: The librispeech lexicon contains a list of common english words and their composing phonemes. Need to know what phonemes exist

Task: Create function that returns a list of phonemes given the librispeech_lexicon.txt file

Desired input: Filepath to librispeech_lexicon.txt file

Desired output: Python list of all phonemes sorted alphabetically. (The natsort library might be handy if the standard python sort doesn't work)

The lexicon format follows [word] \t [phoneme_1] [phoneme_2] ... [phoneme_n] For example: ACCELERATE AE0 K S EH1 L ER0 EY2 T

List of syllables and composing Phonemes http://web.archive.org/web/20160822211027/http:/semarch.linguistics.fas.nyu.edu/barker/Syllables/index.txt

[deliat258]

"""
    Description:
        Create a function that will get all the phonemes that exist in the
        file 'librispeech-lexicon.txt'. The code will take in the file 'librispeech-lexicon.txt'
        and extract each line from the file. The file is structured as [word] \t [phoneme1] ... [phonemeN].
        After reading a line from the file, the first will be take out of the line leaving the phonemes in
        the line. The line will be split into each individual words/phonemes and be placed into a numpy
        array. Only phonemes which are in the numpy array will be added that way there will be no repetition
        of phonemes into the array. The expected output is a sorted array consisting on each phoneme that is
        in the file with no duplicates.
    Current inputs:
        librispeech-lexicon.txt
    Expected output:
        List of all the phonemes in the input file with no duplicates
    Error(s)/warning(s) to work on:
        FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will
        perform elementwise comparison if word not in test_arr:
"""
from pathlib import Path
import numpy as np

def phoneme_list():
    #Path of input file. Change at your will.
    text_path = Path('C:\\Deepcut\\SD1_Intervals_assignment\\librispeech-lexicon.txt')

    # creates an empty numpy array
    phoneme_arr = np.array([])

    # Opens file to read and will be referenced as file in the loop
    with open(text_path, 'r') as file:
        # takes one line from the file. Will be referenced as line
        for line in file:
            # discards the file word file each letter
            discard, space, phoneme_line = line.partition(' ')
            # this will split the line (the one without the first word)
            # and place the individual words into the numpy array
            # the split line will be referenced as word
            for word in phoneme_line.split():
                # if word is not in the numpy array then it will added in there
                if word not in phoneme_arr:
                    phoneme_arr = np.append(phoneme_arr, word)

    # closes file
    file.close()

    # prints a sorted array with no duplicates
    print(np.sort(phoneme_arr, kind='mergesort'))

if __name__ == '__main__':
    phoneme_list()

[jerivl]

This is a good solution. I'm may move the path to be an argument and have it return a list as the output, but this works. Thanks!