def test_networkx(n_testing_node_pairs: int = 10**5, verbose: bool = True):
if verbose is True:
log('Testing networkx.')
n_nodes = 2000
g = nx.erdos_renyi_graph(n_nodes, p=5 / n_nodes)
graph_info(g)
if verbose is True:
log('Calculating distances.')
distance_mx = calculate_distances(g)
if verbose is True:
log(f'Finished calculating distances. Validating results on {n_testing_node_pairs} random node pairs.'
)
nodes = list(g.nodes)
for _ in range(n_testing_node_pairs):
u, v = np.random.choice(nodes, size=2, replace=False)
assert _get_networkx_distance(g, u, v) == distance_mx[u, v]
if verbose is True:
log('Done.')
def test_spatial(verbose: bool = True):
if verbose is True:
log('Testing spatial.')
lat_lng1 = np.array([[0, 0]])
lat_lng2 = np.array([[0, 180], [90, 0]])
distances_real = np.array([20015.086796020572, 10007.543398010284])
distances = get_earth_distances(lat_lng1, lat_lng2)
assert distances.shape == (lat_lng1.shape[0], lat_lng2.shape[0])
assert (distances.flatten() == distances_real).all()
if verbose is True:
log('Done.')
def test_pickle(verbose: bool = True):
if verbose is True:
log('Testing pickle.')
temp_fpath = _get_temp_fpath('pkl')
obj = nx.Graph()
obj.add_edge(0, 1)
save_pickle(obj, temp_fpath)
load_pickle(temp_fpath)
temp_fpath.unlink()
if verbose is True:
log('Done.')
def test_json(verbose: bool = True):
if verbose is True:
log('Testing json.')
temp_fpath = _get_temp_fpath('json')
obj = {'key1': [1, 2, 3], 'key2': 'string'}
save_json(obj, temp_fpath)
load_json(temp_fpath)
temp_fpath.unlink()
load_json_or_create(temp_fpath, dict)
if verbose is True:
log('Done.')
def test_text(verbose: bool = True):
if verbose is True:
log('Testing text.')
documents = [
'My cat is a very strange creature.', 'Да ну, это какая-то фигня...'
]
tokenized_documents_real = [['cat', 'strange', 'creature'],
['это', 'какой-то', 'фигня']]
tokenized_documents = tokenize_documents(documents)
assert len(tokenized_documents) == len(tokenized_documents_real)
assert all(tokenized_document == tokenized_document_real
for tokenized_document, tokenized_document_real in zip(
tokenized_documents, tokenized_documents_real))
if verbose is True:
log('Done.')
def test_log(verbose: bool = True):
log('Testing', 'log', 1, '...')
if verbose:
log('Testing logging to file.')
log_fpath = _get_temp_fpath('log')
log_message = 'Test message'
log(log_message, to_file=log_fpath)
with log_fpath.open() as file:
contents = file.readlines()
assert (len(contents) == 1) and (contents[0].strip().endswith(log_message))
log_fpath.unlink()
if verbose:
log('Done.')
def test_sparse(verbose: bool = True):
if verbose is True:
log('Testing sparse.')
mx = np.random.randint(0, 2, size=(15, 20))
sparse_mx = sparse.csr_matrix(mx)
for axis in {0, 1}:
log(f'Testing for axis {axis}.')
mx_with_normalized_vectors = mx / np.linalg.norm(
mx, axis=axis, keepdims=True)
mx_with_normalized_vectors[np.isnan(mx_with_normalized_vectors)] = 0
sparse_mx_with_normalized_vectors = normalize_vectors(sparse_mx,
axis=axis)
assert (sparse_mx_with_normalized_vectors == mx_with_normalized_vectors
).all()
if verbose is True:
log('Done.')
documents = [
'My cat is a very strange creature.', 'Да ну, это какая-то фигня...'
]
tokenized_documents_real = [['cat', 'strange', 'creature'],
['это', 'какой-то', 'фигня']]
tokenized_documents = tokenize_documents(documents)
assert len(tokenized_documents) == len(tokenized_documents_real)
assert all(tokenized_document == tokenized_document_real
for tokenized_document, tokenized_document_real in zip(
tokenized_documents, tokenized_documents_real))
if verbose is True:
log('Done.')
if __name__ == '__main__':
test_log()
test_json()
test_pickle()
test_networkx()
test_sparse()
test_spatial()
test_text()
log('Note that the function `get_layout` from `myutils.networkx` was not tested since it requires '
'`python-igraph`, which is not always trivial to install. Seems to work though.'
)
