Syntax

Graphs are defined via facts using predicates node, edge, graph and attr.

Note

A custom prefix for the predicate names can be defined to avoid clashes with other existing predicates.

Nodes

A node is defined using predicate node where the argument is an identifier.

Example 1

node(john).
node(jane).

Edges

An edge from A to B is defined using predicate edge where the argument is a tuple (A,B). The edges will be directed when creating a digraph and not directed one created a graph.

Example 1 (continuation)

edge((john,jane)).

Graphs

A graph is defined using predicate graph/1.

Example

graph(family).

Multiple graphs

Multiple graphs can defined within the same set of facts. Each of them will generate a different graph. In order to assign a node or edge to an specific graph we add the identifier of the graph in the second argument.

Warning

All nodes and edges defined with a single argument are assigned to the default graph called default. A different graph can be defined as the default as an option.

Example 2

graph(toms_family).
graph(bills_family).
node(tom, toms_family).
node(max, toms_family).
edge((tom, max), toms_family).
node(bill, bills_family).
node(jen, bills_family).
edge((bill, jen), bills_family).

Graph toms_family

Graph bills_family

Note

The creation of multiple graphs is specially useful for representing a dynamic domain where we need a different visualization for each time step

Subgraphs

Subgraphs can be defined by adding a parent graph as a second argument to the graph predicate. Then, nodes and edges can be assigned to a subgraph or a graph in the same way.

Example 3

graph(house).
graph(bathroom, house).
graph(bedroom, house).
node(toilet,bathroom).
node(bed,bedroom).
node(desk,bedroom).
edge((toilet,bed),house).

Attributes

The attributes of an element (graph, node or edge) are defined using predicate attr. This predicate will add any attributes accepted by graphviz (see here) specified by name-value pairs to the element.

• attr(ELEMENT_TYPE, ELEMENT_ID, ATTR_NAME, ATTR_VALUE).

  • ELEMENT_TYPE: The type of element: can be graph, node or edge. Additionally, the element types graph_nodes and graph_edges assign an attribute to all nodes/edges of a graph identified with ELEMENT_ID.

  • ELEMENT_ID: The identifier of the element

  • ATTR_NAME: The name of the graphviz attribute

  • ATTR_VALUE: The value of the graphviz attribute

Example 3 (continuation)

attr(graph, house, label, "Tom's House").
attr(graph, bathroom, style, dotted).
attr(graph, bathroom, label, "Bathroom").
attr(graph, bedroom, style, dotted).
attr(graph, bedroom, label, "Bedroom").
attr(graph_nodes, house, style, filled).
attr(graph_nodes, house, color, cyan).
attr(node, toilet, shape, circle).
attr(node, bed, shape, square).
attr(node, desk, shape, square).
attr(edge, (toilet,bed), color, red).

Note

To edit the attributes of the default graph one must use the default name as element id.

Example 1 (continuation)

attr(graph, default, label, "Does' family").
attr(graph_nodes, default, style, filled).
attr(node, john, label, "John Doe").
attr(node, jane, label, "Jane Doe").

Multi-attribute

If multiple occurrences of the same attribute name for a specific element are present, then the values will be concatenated in no specific order. To define an order the attribute name can be a tuple where the second value is the position where the value will be considered. Furthermore, if the position is sep the value will be considered as the separator of the strings in the concatenation.

Example 4

node(mike).
attr(node, mike, (label,0), "Michel").
attr(node, mike, (label,1), "Scott").
node(toby).
edge((mike,toby)).
attr(edge, (mike,toby)), label, "enemy").
attr(edge, (mike,toby)), label, "hate").
attr(edge, (mike,toby)), label, "boss").
attr(edge, (mike,toby)), (label,sep), "-").