NetworkitBinaryReader tries to read edges multiple times

[fabratu]

The NetworkitBinaryReader tries to read edges multiple times for undirected graphs. The resulting graph is correct, since the multi-edge is suppressed. Happens both for C++ and Python.

Example (in Python):

import networkit as nk
G = nk.generators.ErdosRenyiGenerator(100, 0.5, False, False).generate();
nk.overview(G)

nk.setLogLevel("DEBUG")

nkWriter = nk.graphio.NetworkitBinaryWriter()
nkWriter.write(G, "./G_gen.nkbg")
nkReader = nk.graphio.NetworkitBinaryReader()
G = nkReader.read("./G_gen.nkbg")
nk.overview(G)

Output of nk.overview:

Network Properties:
nodes, edges			100, 2477
directed?			False
weighted?			False
isolated nodes			0
self-loops			0
density				0.500404
clustering coefficient		0.501031
min/max/avg degree		38, 67, 49.540000
degree assortativity		-0.029328
number of connected components	1
size of largest component	100 (100.00 %)

Network Properties:
nodes, edges			100, 2477
directed?			False
weighted?			False
isolated nodes			0
self-loops			0
density				0.500404
clustering coefficient		0.501031
min/max/avg degree		38, 67, 49.540000
degree assortativity		-0.029328
number of connected components	1
size of largest component	100 (100.00 %)

Warnings:

[INFO ][/project/networkit/cpp/io/NetworkitBinaryWriter.cpp, 29: virtual void NetworKit::NetworkitBinaryWriter::write(const NetworKit::Graph&, const string&)]: Written graph to ./G_gen.nkbg
[WARN ][/project/networkit/cpp/io/NetworkitBinaryReader.cpp, 274: NetworKit::NetworkitBinaryReader::readData<NetworKit::MemoryMappedFile>::<lambda(uint64_t)>]: Not adding edge 0-1 since it is already present.
[WARN ][/project/networkit/cpp/io/NetworkitBinaryReader.cpp, 274: NetworKit::NetworkitBinaryReader::readData<NetworKit::MemoryMappedFile>::<lambda(uint64_t)>]: Not adding edge 0-4 since it is already present.
[WARN ][/project/networkit/cpp/io/NetworkitBinaryReader.cpp, 274: NetworKit::NetworkitBinaryReader::readData<NetworKit::MemoryMappedFile>::<lambda(uint64_t)>]: Not adding edge 0-5 since it is already present.
...