sample code from web for machine learning

demo1.py

# coding: utf-8

# In[1]:


# Start of code for librosa walkthrough
# add autocomplete with TAB
# get_ipython().run_line_magic('config', 'IPCompleter.greedy=True')


# In[3]:


import IPython.display as ipd


# In[4]:


#ipd.Audio('../data/Train/2022.wav')

#
# OSX does not have splay but it does have 
# afplay
ipd.Audio('../samples/UrbanSound/data/gun_shot/93139.wav')


# In[9]:


ipd.Audio('../samples/UrbanSound/data/gun_shot/87562.wav')


# In[10]:


ipd.Audio('../samples/UrbanSound/data/gun_shot/77247.wav')


# In[11]:


ipd.Audio('../samples/UrbanSound/data/gun_shot/7060.wav')


# In[5]:


import librosa


# In[6]:


data, sampling_rate = librosa.load('../samples/Demo/train/Train/2022.wav')


# In[7]:


data
#sampling_rate
# sampling_rate is 22050


# In[9]:


get_ipython().run_line_magic('pylab', 'inline')
import os
import pandas as pd
import librosa
import librosa.display
import glob 

plt.figure(figsize=(12, 4))
librosa.display.waveplot(data, sr=sampling_rate)


# In[14]:


train = pd.read_csv('../samples/Demo/train/train.csv') 


# In[84]:


#train


# In[17]:


i = random.choice(train.index)
i


# In[21]:


data_dir = '../samples/Demo/'


# In[19]:


audio_name = train.ID[i]


# In[23]:


path = os.path.join(data_dir, 'Train', str(audio_name) + '.wav')
path


# In[24]:


print('Class: ', train.Class[i])


# In[32]:


# orig
# x, sr = librosa.load('../data/Train/' + str(train.ID[i]) + '.wav')
the_filename = '../samples/Demo/train/Train/' + str(train.ID[i]) + '.wav'
x, sr = librosa.load(the_filename)


# In[30]:


plt.figure(figsize=(12, 4))


# In[31]:


librosa.display.waveplot(x, sr=sr)


# In[33]:


ipd.Audio(the_filename)


# In[34]:


train.Class.value_counts()


# In[35]:


test = pd.read_csv('../samples/Demo/test/test.csv')


# In[36]:


test['Class'] = 'jackhammer'


# In[38]:


test.to_csv('sub01.csv', index=False)


# In[42]:


data_dir = '../samples/Demo/train'


# In[60]:


def parser(row):
    # function to load files and extract features
    file_name = os.path.join(os.path.abspath(data_dir), 'Train', str(row.ID) + '.wav')
    # print(file_name)
    print('.', end='')

    # handle exception to check if there isn't a file which is corrupted
    try:
        # here kaiser_fast is a technique used for faster extraction
        X, sample_rate = librosa.load(file_name, res_type='kaiser_fast') 
        # we extract mfcc feature from data
        mfccs = np.mean(librosa.feature.mfcc(y=X, sr=sample_rate, n_mfcc=40).T,axis=0) 
    except Exception as e:
        print("Error encountered while parsing file: ", file)
        return None, None

    feature = mfccs
    label = row.Class
 
    return [feature, label]


# In[61]:


temp = train.apply(parser, axis=1)


# In[52]:


# train
# is a dataframe, here is a preview
#     ID   Class
# 0    0    siren
# 1    1    street_music
# 2    2    drilling


# In[85]:


#str(temp)


# In[64]:


#temp
type(temp)


# In[76]:


temp.columns = ['feature', 'label']


# In[83]:


temp.label


# In[77]:


temp.head()


# In[78]:


from sklearn.preprocessing import LabelEncoder


# In[79]:


X = np.array(temp.feature.tolist())


# In[75]:


y = np.array(temp.label.tolist())


# In[ ]:


lb = LabelEncoder()


# In[ ]:


y = np_utils.to_categorical(lb.fit_transform(y))