def __init__(self,
adj=None,
gml_path=None,
csv_path=None,
leiden=True,
threshold=0,
gamma=1,
n_comms=5,
seed=None):
self.adj = adj
if adj is not None:
self.adj = get_adj(adj)
self.gml_path = gml_path
self.csv_path = csv_path
self.detector = leidenalg if leiden else louvain
self.threshold = threshold
self.gamma = gamma
self.n_comms = n_comms
self.seed = seed
self.graph = None
self.main_sub = None
self.partition = None
self.df = None
self.tempfile_path = None
if self.gml_path is None:
self.tempfile_path = Path(str(uuid.uuid4()) + '.gml')
def visualize_distro(adj, out_path, sample=None):
adj = get_adj(adj)
if isinstance(adj, pd.DataFrame):
adj = adj.values
if sample is not None:
if not 0 < sample <= 1:
raise ValueError('Value of sample must satisfy: 0 < sample <= 1')
size = int(len(adj) * sample)
rows = np.random.choice(len(adj), size, replace=False)
cols = np.random.choice(len(adj), size, replace=False)
adj = adj[rows][:, cols]
flat = adj.ravel()
mean = flat.mean()
std = flat.std()
std1 = mean + std
std2 = mean + (2 * std)
ax = sns.distplot(flat, label=f'Distro (n={flat.size})')
y_max = ax.get_ylim()[1]
plt.plot((mean, mean), (0, y_max / 2), label='Mean')
plt.plot((std1, std1), (0, y_max / 2), label='Mean + 1 std. dev.')
plt.plot((std2, std2), (0, y_max / 2), label='Mean + 2 std. dev.')
plt.xlabel('r-value')
plt.legend()
plt.savefig(out_path, bbox_inches='tight', dpi=600)
return mean, std
def test_get_adj_str_df(self, tmpdir):
adj_path = self._save_adj(tmpdir, binary=False)
adj = utils.get_adj(adj_path)
assert type(adj) == pd.DataFrame
def test_get_adj_str_ndarray(self, tmpdir):
adj_path = self._save_adj(tmpdir)
adj = utils.get_adj(adj_path)
assert type(adj) == np.ndarray
def test_get_adj_df(self):
expected = self._build_adj()
adj = utils.get_adj(expected)
assert type(adj) == pd.DataFrame
def test_get_adj_ndarray(self):
expected = self._build_adj().values
adj = utils.get_adj(expected)
assert type(adj) == np.ndarray