fseiffarth / graphdatatonetworkxgraphlist

GraphDataToNetworkxGraphList

Reads graph data from:

https://ls11-www.cs.tu-dortmund.de/staff/morris/graphkerneldatasets

and outputs them as a list of python networkx graphs, see the details below.

Functions:

1. graph_data_to_graph_list(path, db) Takes database path {path} and database-name {db} and outputs the graph data in form of a triple: (graph_list, graph_labels, graph_attributes)

   -graph_list: #python list of networkx graphs with graph information given by the database

   -graph_labels: #python list of integers as the labels of the graphs according to the database

   -graph_attributes: #python list with additional attributes of the graphs according to the database or empty list if there are none

2. node_label_vector(graph, node_id) #outputs vector of node_label of node with {node_id} in graph {graph}

3. node_attribute_vector(graph, node_id) #outputs vector of node_attributes of node with {node_id} in graph {graph}

4. nodes_label_matrix(graph) #same as above but outputs all labels of nodes of graph {graph}

5. nodes_attribute_matrix(graph) #same as above but outputs all attributes of nodes of graph {graph}

6. edge_label(graph, node_i, node_j) #outputs vector of edge_labels of edge (node_i, node_j) in graph {graph}

7. edge_attribute_matrix(graph, node_i, node_j) #outputs vector of edge_attributes of edge (node_i, node_j) in graph {graph}

======= Python example:

path = "path_to_dbs/"
db = "MUTAG"
graph_data = graph_data_to_graph_list(path, db)
graph_list, graph_labels, graph_attributes = graph_data

graphdatatonetworkxgraphlist's Issues

Parsing Issue with Fingerprint Dataset

The Fingerprint dataset seems to be incorrectly processed by the converter function.
The following code snippet can be used to reproduce the problem, that is: all graphs from index 2149 in the dataset will have zero vertices and edges (which is not the case).

import GraphDataToNetworkxGraphList.GraphDataToGraphList as g2l
graphs = g2l.graph_data_to_graph_list(db='Fingerprint', path='./DS_all/')

problembaer = graphs[0][2149]
print(problembaer.number_of_nodes())