def test_compute_prob_matrix(self):
"""
Test `compute_prob_matrix` function
"""
clickstream = gen_random_clickstream(n_of_streams=100, n_of_pages=12)
markov_clickstream = MarkovClickstream(clickstream_list=clickstream)
markov_clickstream.compute_prob_matrix()
prob_matrix = markov_clickstream.prob_matrix
for row in range(prob_matrix.shape[0]):
self.assertAlmostEqual(np.sum(prob_matrix[row, :]), 1)
def test_get_unique_pages(self):
"""
Tests the `get_unique_pages` function
"""
clickstream = gen_random_clickstream(n_of_streams=100, n_of_pages=12)
markov_clickstream = MarkovClickstream(clickstream_list=clickstream)
pages = markov_clickstream.get_unique_pages()
self.assertEqual(len(pages), len(set(pages)))
clickstream_pages = flatten_list(clickstream)
self.assertEqual(set(pages), set(clickstream_pages))
def test_initialise_count_matrix(self):
"""
Test `initialise_count_matrix` function
"""
clickstream = gen_random_clickstream(n_of_streams=100, n_of_pages=12)
markov_clickstream = MarkovClickstream(clickstream_list=clickstream)
markov_clickstream.initialise_count_matrix()
count_matrix = markov_clickstream.count_matrix
pages = set(flatten_list(clickstream))
expected = np.zeros((len(pages), len(pages)))
self.assertTrue((count_matrix == expected).all())
def test_calculate_pagerank(self):
"""
Test the 'calculate_pagerank` function.
"""
clickstream = gen_random_clickstream(n_of_streams=100, n_of_pages=12)
markov_clickstream = MarkovClickstream(clickstream_list=clickstream)
n_pages = len(markov_clickstream.pages)
max_edges = random.randrange(1, n_pages)
digraph, pagerank_scores = markov_clickstream.calculate_pagerank(
max_nodes=max_edges)
self.assertIsInstance(digraph, nx.DiGraph)
self.assertIsInstance(pagerank_scores, dict)
self.assertEqual(digraph.number_of_nodes(), n_pages)
self.assertLess(digraph.number_of_nodes(), (max_edges * n_pages) + 1)
self.assertGreater(digraph.number_of_edges(), n_pages - 1)
self.assertEqual(len(pagerank_scores.values()), n_pages)
def test_populate_count_matrix(self):
"""
Tests the `populate_count_matrix` function
"""
clickstream = [['P1', 'P2', 'P2', 'P2'], ['P3', 'P4', 'P5']]
markov_clickstream = MarkovClickstream(clickstream_list=clickstream)
markov_clickstream.populate_count_matrix()
expected_count_matrix = np.array([[0., 1., 0., 0., 0.],
[0., 2., 0., 0., 0.],
[0., 0., 0., 1., 0.],
[0., 0., 0., 0., 1.],
[0., 0., 0., 0., 0.]])
self.assertTrue(
(expected_count_matrix == markov_clickstream.count_matrix).all())
def test_normalise_row(self):
"""
Tests `normalise_row` function
"""
test_cases = [
(np.array([5.0, 0.0]), np.array([1.0, 0.0])),
(np.array([0.0, 0.0]), np.array([0.0, 0.0])),
]
for row, expected in test_cases:
normalised = MarkovClickstream.normalise_row(row)
print(row, expected, normalised)
self.assertTrue((normalised == expected).all())
df.head()
# In[18]:
data = []
for i in range(100):
x = df.values[i]
x = x[~pd.isnull(x)]
data.append(x)
data
# In[19]:
from markovclick.models import MarkovClickstream
m = MarkovClickstream(data)
# In[20]:
fig = plt.figure(figsize=(18, 16), dpi=100, facecolor='r', edgecolor='k')
sns.heatmap(m.prob_matrix,
xticklabels=m.pages,
yticklabels=m.pages,
cmap="YlGnBu")
# In[21]:
from markovclick.viz import visualise_markov_chain
graph = visualise_markov_chain(m)
# In[22]: