def export_json_3d_graph():
path_to_csv = '../../data/literature/lesmiserables.csv'
graph_structure = read_graph_structure_from_csv(path_to_csv)
graph = WeightedGraph(graph_structure)
graph.build_graph()
graph.build_filtered_subgraphs()
graph.get_temporary_hubs_along_filtration()
graph.steady_hubs_persistence()
graph.ranging_hubs_persistence()
graph.plot_steady_persistence_diagram(export_json=True)
graph.plot_ranging_persistence_diagram(export_json=True)
plt.show()
def reproduce_thesis_figure(graph_structure_number=2):
above_max_diagonal_gap = True
path_to_csv_1 = './MATRICE DISTANZE.csv'
path_to_csv_2 = './MATRICE FREQUENZE.csv'
graph_structure = read_csv_distance_matrix(path_to_csv_1)
graph_structure_2 = read_csv_frequencies_matrix(path_to_csv_2)
graph_structure_1 = get_structure(graph_structure,
graph_structure_2,
product=False)
graph_structure_3 = get_structure(graph_structure_2,
graph_structure,
product=True)
graph_structures = [
graph_structure, graph_structure_1, graph_structure_2,
graph_structure_3
]
graph = WeightedGraph(graph_structures[graph_structure_number])
graph.build_graph()
graph.build_filtered_subgraphs(weight_transform=identity, sublevel=True)
graph.get_temporary_hubs_along_filtration()
# Steady
graph.steady_hubs_persistence(
above_max_diagonal_gap=above_max_diagonal_gap, gap_number=0)
fig, ax = plt.subplots()
graph.steady_pd.plot_gudhi(
ax, persistence_to_plot=graph.steady_pd.persistence_to_plot)
if hasattr(graph.steady_pd, 'proper_cornerpoints_above_gap'):
graph.steady_pd.plot_nth_widest_gap(ax_handle=ax, n=graph.gap_number)
graph.steady_pd.mark_points_above_diagonal_gaps(ax)
plt.show()
print('steady hubs above gap:',
graph.steady_pd.proper_cornerpoints_above_gap)
from pprint import pprint
pprint([(c.vertex, c.persistence)
for c in graph.steady_pd.proper_cornerpoints])
# Ranging
graph.ranging_hubs_persistence(
above_max_diagonal_gap=above_max_diagonal_gap)
fig, ax = plt.subplots()
graph.steady_pd.plot_gudhi(
ax, persistence_to_plot=graph.ranging_pd.persistence_to_plot)
if hasattr(graph.ranging_pd, 'proper_cornerpoints_above_gap'):
graph.ranging_pd.plot_nth_widest_gap(ax_handle=ax, n=graph.gap_number)
graph.ranging_pd.mark_points_above_diagonal_gaps(ax)
plt.show()
print('ranging hubs above gap:',
graph.ranging_pd.proper_cornerpoints_above_gap)
pprint([(c.vertex, c.persistence)
for c in graph.ranging_pd.proper_cornerpoints])
def reproduce_thesis_figure():
above_max_diagonal_gap = True
path_to_csv = '../../data/languages/SSWL5_norm.csv'
graph_structure = read_csv_distance_matrix(path_to_csv)
graph = WeightedGraph(graph_structure)
graph.build_graph()
graph.build_filtered_subgraphs(weight_transform=max_, sublevel=True)
graph.get_temporary_hubs_along_filtration()
# Steady
graph.steady_hubs_persistence(
above_max_diagonal_gap=above_max_diagonal_gap, gap_number=2)
graph.plot_steady_persistence_diagram(show=True)
fig, ax = plt.subplots()
graph.steady_pd.plot_gudhi(
ax, persistence_to_plot=graph.steady_pd.persistence_to_plot)
if hasattr(graph.steady_pd, 'proper_cornerpoints_above_gap'):
graph.steady_pd.plot_nth_widest_gap(ax_handle=ax, n=graph.gap_number)
graph.steady_pd.mark_points_above_diagonal_gaps(ax)
plt.savefig('steady_ling_gap2' + '.pdf', dpi=300)
print('Steady hubs:', [c.vertex for c in graph.steady_cornerpoints])
print('steady hubs above gap:',
graph.steady_pd.proper_cornerpoints_above_gap)
pprint([(c.birth, c.death, c.vertex)
for c in graph.steady_pd.proper_cornerpoints_above_gap])
# Ranging
graph.ranging_hubs_persistence(
above_max_diagonal_gap=above_max_diagonal_gap, gap_number=2)
graph.plot_ranging_persistence_diagram(show=True)
fig, ax = plt.subplots()
graph.ranging_pd.plot_gudhi(
ax, persistence_to_plot=graph.ranging_pd.persistence_to_plot)
if hasattr(graph.ranging_pd, 'proper_cornerpoints_above_gap'):
graph.ranging_pd.plot_nth_widest_gap(ax_handle=ax, n=graph.gap_number)
graph.ranging_pd.mark_points_above_diagonal_gaps(ax)
plt.savefig('ranging_ling_gap2' + '.pdf', dpi=300)
print('Ranging hubs:', [c.vertex for c in graph.ranging_cornerpoints])
print('ranging hubs above gap:',
graph.ranging_pd.proper_cornerpoints_above_gap)
pprint([(c.birth, c.death, c.vertex)
for c in graph.ranging_pd.proper_cornerpoints_above_gap])
def reproduce_thesis_figure():
above_max_diagonal_gap = True
path_to_csv = '../../data/literature/lesmiserables.csv'
graph_structure = read_graph_structure_from_csv(path_to_csv)
graph = WeightedGraph(graph_structure)
graph.build_graph()
graph.build_filtered_subgraphs()
graph.get_temporary_hubs_along_filtration()
# Steady
graph.steady_hubs_persistence(
above_max_diagonal_gap=above_max_diagonal_gap, gap_number=1)
graph.plot_steady_persistence_diagram()
fig, ax = plt.subplots()
graph.steady_pd.plot_gudhi(
ax, persistence_to_plot=graph.steady_pd.persistence_to_plot)
if hasattr(graph.steady_pd, 'proper_cornerpoints_above_gap'):
graph.steady_pd.plot_nth_widest_gap(ax_handle=ax, n=graph.gap_number)
graph.steady_pd.mark_points_above_diagonal_gaps(ax)
plt.savefig('gap2_steady' + '.pdf', dpi=300)
print('steady hubs above gap:',
graph.steady_pd.proper_cornerpoints_above_gap)
pprint([(c.birth, c.death, c.vertex)
for c in graph.steady_pd.proper_cornerpoints_above_gap])
# Ranging
graph.ranging_hubs_persistence(
above_max_diagonal_gap=above_max_diagonal_gap, gap_number=1)
graph.plot_ranging_persistence_diagram()
fig, ax = plt.subplots()
graph.ranging_pd.plot_gudhi(
ax, persistence_to_plot=graph.ranging_pd.persistence_to_plot)
if hasattr(graph.ranging_pd, 'proper_cornerpoints_above_gap'):
graph.ranging_pd.plot_nth_widest_gap(ax_handle=ax, n=graph.gap_number)
graph.ranging_pd.mark_points_above_diagonal_gaps(ax)
plt.savefig('gap2_ranging' + '.pdf', dpi=300)
print('ranging hubs above gap:',
graph.ranging_pd.proper_cornerpoints_above_gap)
pprint([(c.birth, c.death, c.vertex)
for c in graph.ranging_pd.proper_cornerpoints_above_gap])
from __future__ import absolute_import, division, print_function
from sys import platform
if platform == "darwin":
import matplotlib
matplotlib.use("TkAgg")
import sys
sys.path.append('../../')
import matplotlib.pyplot as plt
from perscomb.read_data_from_csv import read_graph_structure_from_csv
from perscomb.weighted_graph import WeightedGraph
def opp(x):
return -x
def identity(x):
return x
path_to_csv = '../../data/neuroscience/celegans_connectome.csv'
graph_structure = read_graph_structure_from_csv(path_to_csv)
graph = WeightedGraph(graph_structure)
graph.build_graph()
graph.build_filtered_subgraphs(weight_transform=opp, sublevel=True)
graph.get_temporary_hubs_along_filtration()
graph.ranging_hubs_persistence()
graph.plot_ranging_persistence_diagram(export_json='./c_elegans')
def get_structure(self, graph, index):
if self.persistence_function == "steady":
nodes_color_map,\
nodelist,\
node_size, cornerpoints = graph.plot_steady_persistence_diagram(return_attr = True)
elif self.persistence_function == "ranging":
nodes_color_map,\
nodelist,\
node_size, cornerpoints = graph.plot_ranging_persistence_diagram(return_attr = True)
return graph.export_graph_and_hubs_as_dict(nodes_color_map,
nodelist,
node_size,
cornerpoints,
observation=index)
if __name__ == "__main__":
path_to_folder = 'data/literature/GOT/'
csv_files = ['1', '2', '3', '45']
graphs = []
for file_name in csv_files:
path_to_csv = os.path.join(path_to_folder, file_name + '.csv')
graph_structure = read_graph_structure_from_csv(path_to_csv)
graphs.append(WeightedGraph(graph_structure))
tvg = TimeVaryingHubs(graphs)
tvg.compute_persistence(persistence_function="ranging")
tvg.export_csv_for_plot(save_as='./got_dynamical_ranging.csv')