Fix bugs.

Problems.md

@@ -0,0 +1,21 @@
+# About graph kenrels.
+
+## (Random walk) Sylvester equation kernel.
+
+### ImportError: cannot import name 'frange' from 'matplotlib.mlab'
+
+You are using an outdated `control` with a recent `matplotlib`. `mlab.frange` was removed in `matplotlib-3.1.0`, and `control` removed the call in `control-0.8.2`.
+
+Update your `control` package.
+
+### Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so.
+
+The Intel Math Kernel Library (MKL) is missing or not properly set. I assume the MKL is required by `control` module.
+
+Install MKL. Then add the following to your path:
+
+```
+export PATH=/opt/intel/bin:$PATH
+
+export LD_LIBRARY_PATH=/opt/intel/lib/intel64:/opt/intel/mkl/lib/intel64:$LD_LIBRARY_PATH
+```

gklearn/experiments/papers/PRL_2020/synthesized_graphs_num_nl.py

@@ -0,0 +1,54 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+Created on Mon Sep 21 10:34:26 2020
+
+@author: ljia
+"""
+from utils import Graph_Kernel_List_VSym, compute_graph_kernel
+
+
+def generate_graphs(num_nl_alp):
+	from gklearn.utils.graph_synthesizer import GraphSynthesizer
+	gsyzer = GraphSynthesizer()
+	graphs = gsyzer.unified_graphs(num_graphs=100, num_nodes=20, num_edges=40, num_node_labels=num_nl_alp, num_edge_labels=0, seed=None, directed=False)
+	return graphs
+
+
+def xp_synthesied_graphs_num_node_label_alphabet():
+		
+	# Run and save.
+	import pickle
+	import os
+	save_dir = 'outputs/synthesized_graphs_num_node_label_alphabet/'
+	if not os.path.exists(save_dir):
+		os.makedirs(save_dir)
+
+	run_times = {}
+	
+	for kernel_name in Graph_Kernel_List_VSym:
+		print()
+		print('Kernel:', kernel_name)
+		
+		run_times[kernel_name] = []
+		for num_nl_alp in [0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20]: # [0, 4, 8, 12, 16, 20, 24, 28, 32, 36, 40]:
+			print()
+			print('Number of node label alphabet:', num_nl_alp)
+			
+			# Generate graphs.
+			graphs = generate_graphs(num_nl_alp)
+
+			# Compute Gram matrix.
+			gram_matrix, run_time = compute_graph_kernel(graphs, kernel_name)
+			run_times[kernel_name].append(run_time)
+			
+			pickle.dump(run_times, open(save_dir + 'run_time.' + kernel_name + '.' + str(num_nl_alp) + '.pkl', 'wb'))
+		
+	# Save all.	
+	pickle.dump(run_times, open(save_dir + 'run_times.pkl', 'wb'))	
+	
+	return
+
+
+if __name__ == '__main__':
+	xp_synthesied_graphs_num_node_label_alphabet()

gklearn/experiments/papers/PRL_2020/utils.py

@@ -5,10 +5,22 @@ Created on Tue Sep 22 11:33:28 2020
 
 @author: ljia
 """
-Graph_Kernel_List = ['PathUpToH', 'WLSubtree', 'SylvesterEquation', 'Marginalized', 'ShortestPath', 'Treelet', 'ConjugateGradient', 'FixedPoint', 'SpectralDecomposition', 'StructuralSP', 'CommonWalk'] 
+Graph_Kernel_List = ['PathUpToH', 'WLSubtree', 'SylvesterEquation', 'Marginalized', 'ShortestPath', 'Treelet', 'ConjugateGradient', 'FixedPoint', 'SpectralDecomposition', 'StructuralSP', 'CommonWalk']
 # Graph_Kernel_List = ['CommonWalk', 'Marginalized', 'SylvesterEquation', 'ConjugateGradient', 'FixedPoint', 'SpectralDecomposition', 'ShortestPath', 'StructuralSP', 'PathUpToH', 'Treelet', 'WLSubtree']
 
 
+Graph_Kernel_List_VSym = ['PathUpToH', 'WLSubtree', 'Marginalized', 'ShortestPath', 'Treelet', 'ConjugateGradient', 'FixedPoint', 'StructuralSP', 'CommonWalk'] 
+
+
+Graph_Kernel_List_ESym = ['PathUpToH', 'Marginalized', 'Treelet', 'ConjugateGradient', 'FixedPoint', 'StructuralSP', 'CommonWalk']
+
+
+Graph_Kernel_List_VCon = ['ShortestPath', 'ConjugateGradient', 'FixedPoint', 'StructuralSP']
+
+
+Graph_Kernel_List_ECon = ['ConjugateGradient', 'FixedPoint', 'StructuralSP']
+
+
 def compute_graph_kernel(graphs, kernel_name):
 	import multiprocessing
 	
@@ -86,6 +98,7 @@ def compute_graph_kernel(graphs, kernel_name):
 		estimator = weisfeilerlehmankernel
 		params = {'base_kernel': 'subtree', 'height': 5}
 		
+# 	params['parallel'] = None
 	params['n_jobs'] = multiprocessing.cpu_count()
 	params['verbose'] = True
 	results = estimator(graphs, **params)

gklearn/kernels/untilHPathKernel.py

@@ -649,7 +649,7 @@ def paths2labelseqs(plist, G, ds_attrs, node_label, edge_label):
             #     path_strs.append(tuple(strlist))
         else:
             path_strs = [
-                tuple([G.node[node][node_label] for node in path])
+                tuple([G.nodes[node][node_label] for node in path])
                 for path in plist
             ]
         return path_strs

gklearn/utils/graph_synthesizer.py

@@ -20,9 +20,9 @@ class GraphSynthesizer(object):
 	def random_graph(self, num_nodes, num_edges, num_node_labels=0, num_edge_labels=0, seed=None, directed=False, max_num_edges=None, all_edges=None):
 		g = nx.Graph()
 		if num_node_labels > 0:
+			node_labels = np.random.randint(0, high=num_node_labels, size=num_nodes)
 			for i in range(0, num_nodes):
-				node_labels = np.random.randint(0, high=num_node_labels, size=num_nodes)
-				g.add_node(str(i), node_label=node_labels[i])
+				g.add_node(str(i), atom=node_labels[i]) # @todo: update "atom".
 		else:
 			for i in range(0, num_nodes):
 				g.add_node(str(i))
@@ -31,11 +31,11 @@ class GraphSynthesizer(object):
 			edge_labels = np.random.randint(0, high=num_edge_labels, size=num_edges)				
 			for i in random.sample(range(0, max_num_edges), num_edges):
 				node1, node2 = all_edges[i]
-				g.add_edge(node1, node2, edge_label=edge_labels[i])
+				g.add_edge(str(node1), str(node2), bond_type=edge_labels[i])  # @todo: update "bond_type".
 		else:
 			for i in random.sample(range(0, max_num_edges), num_edges):
 				node1, node2 = all_edges[i]
-				g.add_edge(node1, node2)
+				g.add_edge(str(node1), str(node2))
 		
 		return g
 	

requirements.txt

@@ -1,6 +1,6 @@
 numpy>=1.16.2
 scipy>=1.1.0
-matplotlib>=3.0.0
+matplotlib>=3.1.0
 networkx>=2.2
 scikit-learn>=0.20.0
 tabulate>=0.8.2

requirements_pypi.txt

@@ -1,6 +1,6 @@
 numpy>=1.16.2
 scipy>=1.1.0
-matplotlib>=3.0.0
+matplotlib>=3.1.0
 networkx>=2.2
 scikit-learn>=0.20.0
 tabulate>=0.8.2