

In [23]:

    
from lib import MongoBackend
from matplotlib import pyplot as plt
from matplotlib.pyplot import show
backend = MongoBackend('master')

Centralized Preproc



In [27]:

    
query = ({}, {'lines':1, 'time':1})
backend.plot_query("imr_preprocess", query, title="Centralized Preprocess",
                   ylabel='time (sec)', xlabel='#entries',
                   label="exec. time")

Centralized W2V Train



In [29]:

    
backend.plot_query("imr_w2v_train", query, '-x', title="W2V Centralized Train",
                   ylabel='time (sec)', xlabel='#entries',label="exec. time")

### alternative implementation ###
# res = backend.query("imr_w2v_train", query)
# res = zip(*res)[1:]

# import matplotlib.pyplot as plt
# plt.plot(res[0],res[1], alpha=0.3)
# plt.scatter(res[0],res[1], marker='D',s=50, color='red')


plt.show()

Centralized W2V Vectorize



In [30]:

    
backend.plot_query("imr_w2v_vectorize", query, title="W2V Centralized Vectorization",
                   ylabel='time (sec)', xlabel='# documents',
                   label="exec. time")

Centralized Classifier Train



In [31]:

    
backend.plot_query("imr_train", query, title="Train Centralized Classifier",
                   ylabel='time (sec)', xlabel='# documents',
                   label="exec. time")

Centralized Classification



In [ ]:

    
backend.plot_query("imr_predict", query, title="Predict with Centralized Classifier",
                   ylabel='time (sec)', xlabel='# documents',
                   label="exec. time")

Spark W2V Train



In [4]:

    
query = ({}, {'lines':1, 'time':1})
backend.plot_query("imr_w2v_train_spark", query, title="Spark W2V Train",
                   ylabel='time (sec)', xlabel='#entries',
                   label="exec. time")

Spark W2V Vectorize



In [8]:

    
backend.plot_query("imr_w2v_vectorize_spark", query, title="Spark W2V Vectorization",
                   ylabel='time (sec)', xlabel='# documents',
                   label="exec. time")

Spark Classifier Train



In [7]:

    
backend.plot_query("imr_lr_train_spark", query, title="Train Spark Classifier",
                   ylabel='time (sec)', xlabel='# documents',
                   label="exec. time")

Spark Classification



In [10]:

    
backend.plot_query("imr_classify_spark", query, title="Predict with Spark Classifier",
                   ylabel='time (sec)', xlabel='# documents',
                   label="exec. time")

Compararative

Classifier train



In [25]:

    
backend.plot_query("imr_train", query, title="Classifier Training",
                   ylabel='time (sec)', xlabel='# documents',
                   label="Centralized", show_plot=False)
backend.plot_query("imr_lr_train_spark", query,
                   ylabel='time (sec)', xlabel='# documents',
                   label="Spark")

Classify



In [26]:

    
backend.plot_query("imr_predict", query, title="Predict using Model",
                   ylabel='time (sec)', xlabel='# documents',
                   label="Centralized", show_plot=False)
backend.plot_query("imr_classify_spark", query,
                   ylabel='time (sec)', xlabel='# documents',
                   label="Spark")

W2V train



In [27]:

    
query = ({}, {'lines':1, 'time':1})
backend.plot_query("imr_w2v_train", query, '-x', title="W2V Train", show_plot=False,
                   ylabel='time (sec)', xlabel='#entries',label="Centralized")
backend.plot_query("imr_w2v_train_spark", query,
                   ylabel='time (sec)', xlabel='#entries',
                   label="Spark")

W2V Vectorize



In [28]:

    
backend.plot_query("imr_w2v_vectorize", query, title="W2V Vectorization",
                   ylabel='Centralized', xlabel='# documents', show_plot=False,
                   label="Centralized")


backend.plot_query("imr_w2v_vectorize_spark", query, 
                   ylabel='time (sec)', xlabel='# documents',
                   label="Spark")