feature_extraction.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Sun May 31 11:15:47 2020

@author: krishna
"""

import os
import numpy as np
from utils import utils

def extract_features(audio_filepath):
    features = utils.feature_extraction(audio_filepath)
    return features
    
    

def FE_pipeline(feature_list,store_loc,mode):
    create_root = os.path.join(store_loc,mode)
    if not os.path.exists(create_root):
        os.makedirs(create_root)
    if mode=='train':
        fid = open('meta/training_feat.txt','w')
    elif mode=='test':
        fid = open('meta/testing_feat.txt','w')
    elif mode=='validation':
        fid = open('meta/validation_feat.txt','w')
    else:
        print('Unknown mode')
    
    for row in feature_list:
        filepath = row.split(' ')[0]
        lang_id = row.split(' ')[1]
        vid_folder = filepath.split('/')[-2]
        lang_folder = filepath.split('/')[-3]
        filename = filepath.split('/')[-1]
        create_folders = os.path.join(create_root,lang_folder,vid_folder)
        if not os.path.exists(create_folders):
            os.makedirs(create_folders)
        extract_feats = extract_features(filepath)
        dest_filepath = create_folders+'/'+filename[:-4]+'.npy'
        np.save(dest_filepath,extract_feats)
        to_write = dest_filepath+' '+lang_id
        fid.write(to_write+'\n')
    fid.close()
    


if __name__ == '__main__':
    store_loc = '/media/newhd/youtube_lid_data/Features'
    read_train = [line.rstrip('\n') for line in open('meta/training.txt')]
    FE_pipeline(read_train,store_loc,mode='train')
    
    read_test = [line.rstrip('\n') for line in open('meta/testing.txt')]
    FE_pipeline(read_test,store_loc,mode='test')
    
    read_val = [line.rstrip('\n') for line in open('meta/validation.txt')]
    FE_pipeline(read_val,store_loc,mode='validation')