Hi
I have successfully installed Tulip from pip3. The C++ library provides a floydWarshall_SPAP function. Is it avalaible from Python ? Could you please provide a simple and minimal working example ? I'm completely new to Tulip. Thanks
Last edit: Pascal 2019-01-02
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
One usage would be:
tlp.selectShortestPaths(graph, tlp.node(1), tlp.node(2), tlp.AllPaths,
graph['viewMetric'], graph['viewSelection'])
the paths are stored in the edges (and connected nodes) of graph with
graph['viewSelection'] value equal to True.
For reference, here are the details:
ret = tlp.selectShortestPaths(graph, src, tgt, pathType, weights, selection)
:param graph: The graph to compute on. :type graph: :class:tlp.Graph
:param src: The source node of the paths :type node: :class:tlp.node
:param tgt: The target node of the paths :type node: :class:tlp.node
:param pathType: The type of path to consider :type pathType: tlp.OnePath,
tlp.OneDirectedPath, tlp.OneReversedPath, tlp.AllPaths,
tlp.AllDirectedPaths, tlp.AllReversedPaths (Reversed means directed with
reversed direction)
:param weights: A graph property giving the edges weight if weighted paths
have to be considered. Can be set to null to select unweighted paths. :type
weights:class:tlp.DoubleProperty
:param selection: The graph property to consider as selection. :type
selection:class:tlp.BooleanProperty
:rtype: boolean (True if a path is found)
:throws: an exception if one of the source or target nodes does not belong
to the graph.
Hi
I have successfully installed Tulip from pip3. The C++ library provides a
floydWarshall_SPAP function. Is it avalaible from Python ? Could you
provide a simple and minimal working example, I'm completely new to Tulip.
Thanks
Yes it is not yet in the last release, you'd probably have to compile Tulip
yourself or wait just a little bit for the next release to come out.
By the way, sourceforge is used only for releasing binaries, I would
suggest you to reach the Github repository for up-to-date information: https://github.com/Tulip-Dev/tulip
The version installed on my machine is 5.2.1 and doesn't support
selectShortestPaths:
from tulip import tlp>>> tlp.selectShortestPathsTraceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/home/po/.local/lib/python3.6/site-packages/tulip/init.py", line 41, in getattr
raise AttributeError(name)AttributeError: selectShortestPaths</module></stdin>
Upgrading via pip3 doesn't help.
By the way, does some docs provide a tutorial for Python users (I don't
use C++) in order to get familiar with Tulip fundamentals? Any advice?
Hi!
I have upgraded to version 5.3 and now selectShortestPaths is recognised, great.
Unfortunately, did'nt find many Python demo. Can someone kindly provide an example showing selectShortestPaths basic usage? For instance, given a directed weighted graph such as
wedges =[[1,2, 3.2], [2,3, 4.2], [1,3, 5.0]]
how can we retrieve the all pair shortest path matrix? Thanks in advance.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi!
I have upgraded to version 5.3 and now selectShortestPaths is recognised,
great.
Unfortunately, did'nt find many Python demo. Can someone kindly provide an
example showing selectShortestPaths basic usage? For instance, given a
directed weighted graph such as
wedges =[[1,2, 3.2], [2,3, 4.2], [1,3, 5.0]]
how can we retrieve the all pair shortest path matrix? Thanks in advance.
Thanks. The problem is that the link you provide is more or less equivalent to help(selectShortestPaths) and still requires a lot of digging before getting a concrete case to work. I was able to do this kind of stuff with other libraries (NetworkX, Networkit, Graph-Tool, Igraph, SageMath for instance) because their docs provide some Python examples. But they are very few in Tulip documentation. So, it would be very appreciated if somebody could kindly put in action the selectShortestPaths function against the directed graph with for instance wedges =[[1,2, 3.2], [2,3, 4.2], [1,3, 5.0]] ? Thanks in advance, Pascal
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I feel a bit of frustration over our documentation in your message. We are
very sorry, this is a new feature of the Python API here, so the doc may
not be fully explicit.
And the feature may be improved to a more pythonic way in the future
(please do not hesitate to PR if you wish to contribute).
A basic example of the feature is:
The first argument is the graph in which you want to compute your shortest
path,
the second is your source node,
the third is your target node,
the fourth is the kind of path you want to find:
- only one path considering the graph as undirected (tlp.OnePath),
- only one path considering the graph as directed (tlp.OneDirectedPath),
- only one path considering the graph as directed in the opposit direction (
tlp.OneReversedPath),
- all paths considering the graph as undirected (tlp.AllPath),
- all paths considering the graph as directed (tlp.AllDirectedPaths),
- all paths considering the graph as directed in the opposit direction (
tlp.AllReversedPaths),
the fifth argument is a weight property
the sixth argument is the selection (a boolean property) in which the
results will be stored
a boolean is returned to tell you if or not a path has been found
the results are stored in a boolean selection (graph['viewSelection']) for
which all nodes and edges statisfying your condition are set to True.
That means you can directly create a subgraph from the results of the
search.
However, it seems that you want to directly transpose the logic of a
different API to Tulip.
I am not sure what you mean by "wedges =[[1,2, 3.2], [2,3, 4.2], [1,3, 5.0]]",
but I suppose you would like to define the directed weighted graph in an
online manner.
The logic behind Tulip is a bit different to other libraries, because Tulip
handles hierarchies of graphs and subgraphs, with property inheritance, and
does so in an optimized fashion.
From your example, you could create the graph like this:
Then you can iterate over this graph, for example with a DFS search (maybe
that was the tricky part?):
create a subgraph from the result of the search
sg = addSubGraph(graph['shortest path'])
for n in sg.dfs(n1):
#do something
And you can do a lot more cool stuff from there, such as measuring
centrality in the results, laying out the corresponding subgraph etc. etc.
I hope this can help you!
If you have some feature request, sure you will, or other issues, please
try to post them on our github, as we have moved to this platform.
Thanks. The problem is that the link you provide is more or less
equivalent to help(selectShortestPaths) and still requires a lot of digging
before getting a concrete case to work. I was able to do this kind of stuff
with other libraries (NetworkX, Networkit, Graph-Tool, Igraph, SageMath for
instance) because their docs provide some Python examples. But they are
very few in Tulip documentation. So, it would be very appreciated if
somebody could kindly put in action the selectShortestPaths function
against the directed graph with for instance wedges =[[1,2, 3.2], [2,3,
4.2], [1,3, 5.0]] ? Thanks in advance, Pascal
I really appreciate your time Benjamin, your example is very helpfull, many thanks. You are right, Tulip approach is different but the library seems very powerfull and efficient. I will follow Tulip on Github!
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thanks a lot Pascal for the kind words. I'm very happy we could be of help.
If you have any further issue, please do not hesitate to contact us
(through Github, easier to track and does not mess with my email fonts :).
I hope you will enjoy all the fun of the tulip API!
--
Benjamin
I really appreciate your time Benjamin, your example is very helpfull,
many thanks. You are right, Tulip approach is different but the library
seems very powerfull and efficient. I will follow Tulip on Github!
Hi
I have successfully installed Tulip from pip3. The C++ library provides a floydWarshall_SPAP function. Is it avalaible from Python ? Could you please provide a simple and minimal working example ? I'm completely new to Tulip. Thanks
Last edit: Pascal 2019-01-02
Dear Pascal,
The python bindings have just been pushed recently (but I don't think it is
in the bundle yet)
https://github.com/Tulip-Dev/tulip/issues/39.
One usage would be:
tlp.selectShortestPaths(graph, tlp.node(1), tlp.node(2), tlp.AllPaths,
graph['viewMetric'], graph['viewSelection'])
the paths are stored in the edges (and connected nodes) of graph with
graph['viewSelection'] value equal to True.
For reference, here are the details:
ret = tlp.selectShortestPaths(graph, src, tgt, pathType, weights, selection)
:param graph: The graph to compute on. :type graph: :class:
tlp.Graph:param src: The source node of the paths :type node: :class:
tlp.node:param tgt: The target node of the paths :type node: :class:
tlp.node:param pathType: The type of path to consider :type pathType: tlp.OnePath,
tlp.OneDirectedPath, tlp.OneReversedPath, tlp.AllPaths,
tlp.AllDirectedPaths, tlp.AllReversedPaths (Reversed means directed with
reversed direction)
:param weights: A graph property giving the edges weight if weighted paths
have to be considered. Can be set to null to select unweighted paths. :type
weights:class:
tlp.DoubleProperty:param selection: The graph property to consider as selection. :type
selection:class:
tlp.BooleanProperty:rtype: boolean (True if a path is found)
:throws: an exception if one of the source or target nodes does not belong
to the graph.
Best,
Benjamin
On Wed, 2 Jan 2019 at 21:45, Pascal elastika@users.sourceforge.net wrote:
Hi Benjamin and thanks for your detailed answer
The version installed on my machine is 5.2.1 and doesn't support selectShortestPaths:
Upgrading via pip3 doesn't help.
By the way, does some docs provide a tutorial for Python users (I don't use C++) in order to get familiar with Tulip fundamentals? Any advice?
Yes it is not yet in the last release, you'd probably have to compile Tulip
yourself or wait just a little bit for the next release to come out.
By the way, sourceforge is used only for releasing binaries, I would
suggest you to reach the Github repository for up-to-date information:
https://github.com/Tulip-Dev/tulip
As for all the python documentation, you can find it here, alongside with
quite a lot of examples:
http://tulip.labri.fr/Documentation/current/tulip-python/html/index.html
--
Benjamin
On Fri, 4 Jan 2019 at 19:35, Pascal elastika@users.sourceforge.net wrote:
Ok, I will wait for the next release. And thanks for the pointer to the Python docs.
Hi!
I have upgraded to version 5.3 and now
selectShortestPathsis recognised, great.Unfortunately, did'nt find many Python demo. Can someone kindly provide an example showing selectShortestPaths basic usage? For instance, given a directed weighted graph such as
wedges =[[1,2, 3.2], [2,3, 4.2], [1,3, 5.0]]how can we retrieve the all pair shortest path matrix? Thanks in advance.
Hi Pascal,
You may refer to the github post here:
https://github.com/Tulip-Dev/tulip/issues/39#issuecomment-451337822
Best,
Benjamin
On Fri, 28 Jun 2019 at 03:59, Pascal elastika@users.sourceforge.net wrote:
Thanks. The problem is that the link you provide is more or less equivalent to
help(selectShortestPaths)and still requires a lot of digging before getting a concrete case to work. I was able to do this kind of stuff with other libraries (NetworkX, Networkit, Graph-Tool, Igraph, SageMath for instance) because their docs provide some Python examples. But they are very few in Tulip documentation. So, it would be very appreciated if somebody could kindly put in action the selectShortestPaths function against the directed graph with for instance wedges =[[1,2, 3.2], [2,3, 4.2], [1,3, 5.0]] ? Thanks in advance, PascalHi Pascal,
I feel a bit of frustration over our documentation in your message. We are
very sorry, this is a new feature of the Python API here, so the doc may
not be fully explicit.
And the feature may be improved to a more pythonic way in the future
(please do not hesitate to PR if you wish to contribute).
A basic example of the feature is:
tlp.selectShortestPaths(graph, tlp.node(1), tlp.node(2), tlp.AllPaths,
graph['viewMetric'], graph['viewSelection'])
The first argument is the graph in which you want to compute your shortest
path,
the second is your source node,
the third is your target node,
the fourth is the kind of path you want to find:
- only one path considering the graph as undirected (tlp.OnePath),
- only one path considering the graph as directed (tlp.OneDirectedPath),
- only one path considering the graph as directed in the opposit direction (
tlp.OneReversedPath),
- all paths considering the graph as undirected (tlp.AllPath),
- all paths considering the graph as directed (tlp.AllDirectedPaths),
- all paths considering the graph as directed in the opposit direction (
tlp.AllReversedPaths),
the fifth argument is a weight property
the sixth argument is the selection (a boolean property) in which the
results will be stored
a boolean is returned to tell you if or not a path has been found
the results are stored in a boolean selection (graph['viewSelection']) for
which all nodes and edges statisfying your condition are set to True.
That means you can directly create a subgraph from the results of the
search.
However, it seems that you want to directly transpose the logic of a
different API to Tulip.
I am not sure what you mean by "wedges =[[1,2, 3.2], [2,3, 4.2], [1,3, 5.0]]",
but I suppose you would like to define the directed weighted graph in an
online manner.
The logic behind Tulip is a bit different to other libraries, because Tulip
handles hierarchies of graphs and subgraphs, with property inheritance, and
does so in an optimized fashion.
From your example, you could create the graph like this:
graph = tlp.newGraph()
n1 = graph.addNode()
n2 = graph.addNode()
n3 = graph.addNode()
e1 = graph.addEdge(n1, n2)
e2 = graph.addEdge(n2, n3)
e3 = graph.addEdge(n1, n3)
graph['weight'][e1] = 3.2
graph['weight'][e2] = 4.2
graph['weight'][e3] = 5.0
Given this example, you could search the shortest path:
this property will store the results
graph.getBooleanProperty('shortest path')
tlp.selectShortestPaths(graph, n1, n2, tlp.OnePath, graph['weight'],
graph['shortest path'])
Then you can iterate over this graph, for example with a DFS search (maybe
that was the tricky part?):
create a subgraph from the result of the search
sg = addSubGraph(graph['shortest path'])
for n in sg.dfs(n1):
#do something
And you can do a lot more cool stuff from there, such as measuring
centrality in the results, laying out the corresponding subgraph etc. etc.
I hope this can help you!
If you have some feature request, sure you will, or other issues, please
try to post them on our github, as we have moved to this platform.
Thank you very much,
Benjamin
On Fri, 28 Jun 2019 at 16:09, Pascal elastika@users.sourceforge.net wrote:
I really appreciate your time Benjamin, your example is very helpfull, many thanks. You are right, Tulip approach is different but the library seems very powerfull and efficient. I will follow Tulip on Github!
Thanks a lot Pascal for the kind words. I'm very happy we could be of help.
If you have any further issue, please do not hesitate to contact us
(through Github, easier to track and does not mess with my email fonts :).
I hope you will enjoy all the fun of the tulip API!
--
Benjamin
On Fri, 28 Jun 2019 at 22:47, Pascal elastika@users.sourceforge.net wrote: