hello.py

from scikits.audiolab import wavread
from scikits.talkbox.features import mfcc
from scikits.talkbox.linpred.levinson_lpc import lpc
import os
import numpy as np
from test.LPC import LPCExtractor
from test.MFCC import get_mfcc_extractor
from collections import defaultdict
from sklearn.mixture import GMM
from collections import Counter
import pickle
import operator
import random
from random import randint

gmm_labels = []
gmm_instances = []
gmm_tables = {}
random_str = str(1290152)
f = open("pickle_dump/"+ random_str + "/ubm-pickle.txt","r")
ubm_model = pickle.load(f)
f.close()
base_dir = "/Users/adhithyar/Desktop/ece6255/"
l = LPCExtractor(16000)
mfcc_extractor = get_mfcc_extractor()
counter = 0
for file_name in os.listdir(base_dir + "pickle_dump/" + random_str + "/"):
    print file_name
    if file_name != random_str:
        counter = counter + 1
        f = open("pickle_dump/" + random_str + "/" + file_name)
        gmm_instance = pickle.load(f)
        f.close()
        k = file_name.split("-")[0]
        gmm_labels.append(k)
        gmm_tables[k] = gmm_instance
        gmm_instances.append(gmm_instance)

results = {}
correct_scores = {}
print gmm_labels

f = open("testing_list"+random_str+".txt","r")
testing_set_filenames = pickle.load(f)
f.close()

base_dir = "/Users/adhithyar/Desktop/ece6255/converted_data/"
audio_files = os.listdir(base_dir)
counter = 0
for file_name in audio_files:
    print counter + 1
    if file_name not in testing_set_filenames:
        counter = counter + 1
        flits = file_name.split("-")
        if flits[0] not in correct_scores:
            correct_scores[flits[0]] = []
        data, fs = wavread(base_dir + file_name)[:2]
        ceps = mfcc_extractor.extract(data)
        lpcs = l.extract(data)
        b = np.concatenate((ceps, lpcs), axis=1)
        corresponding_gmm_instance = gmm_tables[flits[0]]
        correct_scores[flits[0]].append(np.sum(corresponding_gmm_instance.score(b)))

f = open("pickle_dump/" + random_str +  "/correct_scores.txt", 'w')
pickle.dump(correct_scores, f)
f.close()