giuliorossetti / dynetx Goto Github PK

Hi!

Thank you for this great package, it is really to use and works for my temporal data. I have a few questions:

I have 7 graphs which each represent a time point. Many nodes in each graph are common.. but may lose connections with nodes/gain connections with other nodes as time progresses.

As such I would like to be able to visualise this in a software package such as a gephi. The code i use to generate this dynamic network is:

import dynetx as dn

networklistOne = [graph1, graph2 etc...]
dynamic_graph = dn.DynGraph(edge_removal=True)
for t, graph in enumerate(networklistOne):
    dynamic_graph.add_interactions_from(graph.edges(data=True), t=t)

My questions are:

is there a way by which this can be saved in a format that can be understood by gephi? it is slightly different to networkx so I am not sure about this.

The total sum of the nodes in the dynamic graph is not equal to the collective number of nodes in each network. I am guessing this because these are common nodes? and that the idea here is that we want to understand which node networks are added/removed over time?

Community detection, is there a way of carrying out community detection in the dynamic graphs?

thank you in advance!

See my question on stackoverflow about the efficiency of converting a list of static graphs to a dynamic graph.
In light of the size of the data, I believe the current duration of the process is unexplained.

I am trying to load edge lists(snapshots) using Dynetx from the file but i am getting this error, how to solve?
g = dn.read_snapshots('times2.txt', nodetype=int, timestamptype=int)
example of the edge lists
7 8001 10
7 8003 10
7 8004 10
7 8005 10
7 8006 10
7 8030 3
7 8034 10
7 8035 10
7 8036 10
7 8037 10
7 8038 10
7 8040 10
7 8041 10
7 8043 10
7 8044 1
7 8047 10
7 8065 10
7 8070 2

Hi,
I have been trying to create a sub class inheriting from DynGraph to implement some desired functionality:

import dynetx as dn
class myDynGraph(dn.DynGraph):
    # custom new method  
    def myMethod(self):
          pass

but when I instantiate the new class G = myDynGraph(), I have the following error:

  File "/home/glarrea/miniconda3/envs/this_is_a_test_dev/lib/python3.8/site-packages/IPython/core/interactiveshell.py", line 3417, in run_code
    exec(code_obj, self.user_global_ns, self.user_ns)
  File "<ipython-input-4-da889620d25d>", line 1, in <module>
    a = myGraph()
  File "/home/glarrea/miniconda3/envs/this_is_a_test_dev/lib/python3.8/site-packages/dynetx/classes/dyngraph.py", line 121, in __init__
    super(self.__class__, self).__init__(data, **attr)
  File "/home/glarrea/miniconda3/envs/this_is_a_test_dev/lib/python3.8/site-packages/dynetx/classes/dyngraph.py", line 121, in __init__
    super(self.__class__, self).__init__(data, **attr)
  File "/home/glarrea/miniconda3/envs/this_is_a_test_dev/lib/python3.8/site-packages/dynetx/classes/dyngraph.py", line 121, in __init__
    super(self.__class__, self).__init__(data, **attr)
  [Previous line repeated 2981 more times]
RecursionError: maximum recursion depth exceeded while calling a Python object

Modifying line 121 of dynetx/classes/dyngraph.py from :

121 super(self.__class__, self).__init__(data, **attr)
to:
121 super(DynGraph, self).__init__(data, **attr) worked for me.

Is there another way of inheriting without modifying the source code?

Thank you for the answer,

Gustavo

Hi, is there a way to add attributes to the edges in the graph other than t & e to the edge?

Hi,

Great job for this very useful library !

I have imported my data into a DynGraph object. I would like now to apply some networkx algorithm for each snapshot of the data separately, is there a way to easily transofrm each snapshot into a networkx graph ? Because even if a DynGraph object inherit from the nx.Graph class, the behaviours are not exactly the same.

Thanks

I am trying to add MultiGraph or MultiDiGraph in dynetx:

import dynetx as dn
G = dn.MultiGraph(edge_removal=True)

but getting the error:
AttributeError: module 'dynetx' has no attribute 'MultiGraph'

Hi @GiulioRossetti! Now that #151 is merged and DyNetX now works with NetworkX 3.0, could you create a new release and deploy new wheels to PyPI? This way when using pip install dynetx it will install a version that works with NetworkX 3.0.

Hi,
I was running the code for the example of the time_slice method, and the output does not match the one in the documentation.
Given:

G = dn.DynDiGraph()
G.add_interaction(0,1, t=0)
G.add_interaction(1,2, t=0)
G.add_interaction(2,3, t=0)
G.add_interaction(0,4, t=1)
G.add_interaction(4,5, t=1)
G.add_interaction(5,6, t=1)
G.add_interaction(7,1, t=2)
G.add_interaction(1,2, t=2)
G.add_interaction(2,3, t=2)
H = G.time_slice(0)
H.interactions()

The output should be: [(0, 1), (1, 2), (1, 3)], but instead, it is [(0, 1, {'t': [[0, -1]]})]. Notice also the -1 in the t. Nonetheless, for [e for e in G1.time_slice(0).edges()] the output is correct (except for the t, which continue to show -1).

Could it be related to how seen is populated in interactions_iter?

Self edges do not show up when I called dn.interactions(G) to get the list of all edges. However, G.has_interaction() indicates that such interaction exists.

>>> import dynetx as dn
>>> G = dn.DynGraph()
>>> G.add_interaction(0, 1, t=0)
>>> G.add_interaction(0, 2, t=0)
>>> G.add_interaction(0, 0, t=0)
>>> G.add_interaction(1, 1, t=0)
>>> G.add_interaction(2, 2, t=0)
>>> G.interactions(t=0)
[(0, 1, {'t': [0]}), (0, 2, {'t': [0]})]
>>> G.has_interaction(0, 0, t=0)
True
>>>

Hello,

How I can draw the network ??

Thanks very much

Hi,

Great package to extend networkx for a sorely needed functionality of graph representation! I am currently trying to PR the notion of a MixedEdgeGraph into networkx. With that, I am also interested in developing a time-series graph that can contain different types of edges. I am wondering if you can help elaborate on the inner data-structures of the DynGraphs and the "API contract".

1. For example, what is time_to_edge attribute a dictionary comprised of? Are there any constraints?
2. Same for snapshots attribute.
3. How does the edge_removal attribute work? There seems to be no API to remove interactions?

I can help make a PR with inline comments and also documentation to help future users afterwards.

While I'm at it, I also have a question regarding the general API. I noticed that edge and node removals are not supported. Is there any limitations preventing this from being added? If so, is there anything I can do to help?

According to the LICENSE file, this package is under the BSD 2-clause license, but setup.py mentions GPLv3. Can you clarify which license applies? Thanks!

Hi.

I am getting this message every time that I want to add an interaction to an empty graph. Here is a snapshot of my code:

import dynetx as dn
g = dn.DynGraph(edge_removal=True)
g.add_interaction(u=1, v=2, t=0)

And the error:

AttributeError: 'DynGraph' object has no attribute '_node'

I installed the dynetx with "pip install dynetx".

As per documentation, the nodes(t=x) function of the class DynGraph, should be returning a list of the nodes in the graph at the snapshot x, as of now it is returning the list of degrees of each node at the snapshot x.

The example in the documentation is as below:

>>> import dynetx as dn
>>> G = dn.DynGraph()   # or DiGraph, MultiGraph, MultiDiGraph, etc
>>> G.add_path([0,1,2], 0)
>>> G.nodes(t=0)
[0, 1, 2]

whereas, on execution, it works as below:

>>> import dynetx as dn
>>> G = dn.DynGraph()
>>> G.add_path([0,1,2], 0)
>>> G.nodes(t=0)
[1, 2, 1]