Extracting vertices with a given property

graph-tool
1

Hi all,
I was wondering if there is a clever way to extract the list of vertices
with a given value of a property or if you have to run a loop on the
vertices on build the list by hand (I searched on the documentation but did
not find it...).
Thanks,
F.

attachment.html (322 Bytes)

2

Hi Flavien,
you may use filters as explained here
http://graph-tool.skewed.de/static/doc/quickstart.html#graph-filtering

Best,
Giuseppe

2014-07-07 8:06 GMT+02:00 Flavien Lambert <petit.lepton(a)gmail.com>:

attachment.html (1.24 KB)

3

Thanks a lot! To be sure to understand, in the XML file, a property must be
defined as a boolean on the nodes and set as either 0 or 1, right?

attachment.html (2.11 KB)

4

With filters, you can either directly select a boolean property or apply a
function. For example, if you have a property representing the temperature
of nodes, you can select all nodes above a specific value and so on.

Giuseppe

2014-07-07 9:32 GMT+02:00 Flavien Lambert <petit.lepton(a)gmail.com>:

attachment.html (3.06 KB)

5

You can create a property map after the file is loaded, as explained here:

    http://graph-tool.skewed.de/static/doc/quickstart.html#property-maps

For your purposes, also consider the find_vertex() function:

    http://graph-tool.skewed.de/static/doc/util.html#graph_tool.util.find_vertex

Best,
Tiago

6

Something like this may work for you:

# Setup your boolean property map, which should give you an empty array of zeros
v_mask = g.new_vertex_property('bool')

# Assign values to the mask property map array from another property map array
# In this case where a certain property map is <= to a value
# Note use of the .a to directly access the arrays -> shortcut for get_array() method
v_mask.a = v_propMap.a <= val

# Then activate your filter
g.set_vertex_filter(v_mask)

# Remember to de-activate prior to running other operations on the whole graph
g.set_vertex_filter(None)

# Remember to clear or recreate your mask prior to new assignments
# This is what I use -> requires 'import numpy'
v_mask.a = numpy.zeros_like(v_mask.a)

There may be other methods that work better, but this has so far worked nicely for me.
Gareth

attachment.html (4.75 KB)

7

Thanks for all these answers! The find_vertex is nice.

Is there a way to use this sample of nodes directly with in_hist or should
I do a loop?

attachment.html (5.58 KB)

8

Hi! Gareth, I just realized that you gave me the answer, sorry for the slow
understanding!
Best,
F.

attachment.html (6.12 KB)