Getting / Setting data on graph elements¶
Tulip allows to associate different kind of data to the graph elements. These data are stored
in a structure called “property” which can be seen as an array indexed by the nodes and edges of
a graph. The class tlp.PropertyInterface
is the base interface for each graph property object.
For each data type that can be attached to graph elements, there is a specific property class.
Below is an exhaustive list of these classes:
tlp.BooleanProperty
: associate boolean values to nodes and edgestlp.ColorProperty
: associate colors (r,g,b,a) to nodes and edgestlp.DoubleProperty
: associate floating point values to nodes and edgestlp.IntegerProperty
: associate integer values to nodes and edgestlp.LayoutProperty
: associate positions (x,y,z) to nodes and bends/control points to edgestlp.SizeProperty
: associate sizes (width,height,depth) to nodes and edgestlp.StringProperty
: associate string values to nodes and edges
Lists of values can also be associated to graph elements through the use of the vector property classes listed below :
tlp.BooleanVectorProperty
: associate lists of boolean values to nodes and edgestlp.ColorVectorProperty
: associate lists of colors (r,g,b,a) to nodes and edgestlp.DoubleVectorProperty
: associate lists of floating point values to nodes and edgestlp.IntegerVectorProperty
: associate lists of integer values to nodes and edgestlp.CoordVectorProperty
: associate lists of positions (x,y,z) to nodes and to edgestlp.SizeVectorProperty
: associate lists of sizes (width,height,depth) to nodes and edgestlp.StringVectorProperty
: associate lists of string values to nodes and edges
In Tulip, there is two different kinds of graph properties :
- Local property : means that the property belongs to the graph. Its ancestors in the subgraph hierarchy do not have access to it but its descendants do.
- Inherited property : means that the property is inherited from an ancestor in the subgraph hierarchy. Only the graph whose the property is local can delete it.
Creating / Accessing / Deleting a graph property¶
Before using a graph property, you have to create it. To do so, use specific methods from the tlp.Graph
class like tlp.Graph.getDoubleProperty()
,
tlp.Graph.getLayoutProperty()
, tlp.Graph.getIntegerVectorProperty()
, ... (see Access and compute graph properties).
All these methods take a string identifier as parameter to name the property. The created property can then be accessed / deleted through its identifier.
It is also possible to create anonymous properties that won’t be managed by a graph. Below is
some code samples that illustrate these features:
# create a DoubleProperty called "myMetric" if it does not exist
# otherwise return the already created property
myMetric = graph.getDoubleProperty("myMetric")
# Once created, a property can also be accessed through this syntax
myMetric = graph["myMetric"]
# The property can be deleted by the graph that has created it
graph.delLocalProperty("myMetric")
# Anonymous properties can also be instantiated the following way
myOtherMetric = tlp.DoubleProperty(graph)
Working with graph properties¶
The sample code below illustrate how to set / get data on graph elements through the use of a graph property:
# creating an integer property called "degree"
degree = graph.getIntegerProperty("degree")
# creating a boolean property called "loop"
loop = graph.getBooleanProperty("loop")
# filling the properties
for n in graph.getNodes():
degree.setNodeValue(n, graph.deg(n))
# the syntax below can also be used
# degree[n] = graph.deg(n)
for e in graph.getEdges():
loop.setEdgeValue(e, graph.source(e) == graph.target(e))
# the syntax below can also be used
# loop[e] = (graph.source(e) == graph.target(e))
# reading the properties
for n in graph.getNodes():
d = degree.getNodeValue(n)
# the syntax below can also be used
# d = degree[n]
print "degree of ", n, " = ", d
for e in graph.getEdges():
isLoop = loop.getEdgeValue(e)
# the syntax below can also be used
# isLoop = loop[e]
if isLoop:
print e, " is a loop"