Interface Graph<N>

  • Type Parameters:
    N - Node parameter type
    All Superinterfaces:
    PredecessorsFunction<N>, SuccessorsFunction<N>
    All Known Subinterfaces:
    MutableGraph<N>
    All Known Implementing Classes:
    AbstractGraph, ImmutableGraph

    @Beta
    @DoNotMock("Use GraphBuilder to create a real instance")
    public interface Graph<N>
    An interface for graph-structured data, whose edges are anonymous entities with no identity or information of their own.

    A graph is composed of a set of nodes and a set of edges connecting pairs of nodes.

    There are three primary interfaces provided to represent graphs. In order of increasing complexity they are: Graph, ValueGraph, and Network. You should generally prefer the simplest interface that satisfies your use case. See the "Choosing the right graph type" section of the Guava User Guide for more details.

    Capabilities

    Graph supports the following use cases (definitions of terms):

    • directed graphs
    • undirected graphs
    • graphs that do/don't allow self-loops
    • graphs whose nodes/edges are insertion-ordered, sorted, or unordered

    Graph explicitly does not support parallel edges, and forbids implementations or extensions with parallel edges. If you need parallel edges, use Network.

    Building a Graph

    The implementation classes that common.graph provides are not public, by design. To create an instance of one of the built-in implementations of Graph, use the GraphBuilder class:

    
     MutableGraph<Integer> graph = GraphBuilder.undirected().build();
     

    GraphBuilder.build() returns an instance of MutableGraph, which is a subtype of Graph that provides methods for adding and removing nodes and edges. If you do not need to mutate a graph (e.g. if you write a method than runs a read-only algorithm on the graph), you should use the non-mutating Graph interface, or an ImmutableGraph.

    You can create an immutable copy of an existing Graph using ImmutableGraph.copyOf(Graph):

    
     ImmutableGraph<Integer> immutableGraph = ImmutableGraph.copyOf(graph);
     

    Instances of ImmutableGraph do not implement MutableGraph (obviously!) and are contractually guaranteed to be unmodifiable and thread-safe.

    The Guava User Guide has more information on (and examples of) building graphs.

    Additional documentation

    See the Guava User Guide for the common.graph package ("Graphs Explained") for additional documentation, including:

    Since:
    20.0
    Author:
    James Sexton, Joshua O'Madadhain
    • Method Summary

      All Methods Instance Methods Abstract Methods 
      Modifier and Type Method Description
      Set<N> adjacentNodes​(N node)
      Returns the nodes which have an incident edge in common with node in this graph.
      boolean allowsSelfLoops()
      Returns true if this graph allows self-loops (edges that connect a node to itself).
      int degree​(N node)
      Returns the count of node's incident edges, counting self-loops twice (equivalently, the number of times an edge touches node).
      Set<EndpointPair<N>> edges()
      Returns all edges in this graph.
      boolean equals​(@Nullable Object object)
      Returns true iff object is a Graph that has the same elements and the same structural relationships as those in this graph.
      boolean hasEdgeConnecting​(EndpointPair<N> endpoints)
      Returns true if there is an edge that directly connects endpoints (in the order, if any, specified by endpoints).
      boolean hasEdgeConnecting​(N nodeU, N nodeV)
      Returns true if there is an edge that directly connects nodeU to nodeV.
      int hashCode()
      Returns the hash code for this graph.
      ElementOrder<N> incidentEdgeOrder()
      Returns an ElementOrder that specifies the order of iteration for the elements of edges(), adjacentNodes(Object), predecessors(Object), successors(Object) and incidentEdges(Object).
      Set<EndpointPair<N>> incidentEdges​(N node)
      Returns the edges in this graph whose endpoints include node.
      int inDegree​(N node)
      Returns the count of node's incoming edges (equal to predecessors(node).size()) in a directed graph.
      boolean isDirected()
      Returns true if the edges in this graph are directed.
      ElementOrder<N> nodeOrder()
      Returns the order of iteration for the elements of nodes().
      Set<N> nodes()
      Returns all nodes in this graph, in the order specified by nodeOrder().
      int outDegree​(N node)
      Returns the count of node's outgoing edges (equal to successors(node).size()) in a directed graph.
      Set<N> predecessors​(N node)
      Returns all nodes in this graph adjacent to node which can be reached by traversing node's incoming edges against the direction (if any) of the edge.
      Set<N> successors​(N node)
      Returns all nodes in this graph adjacent to node which can be reached by traversing node's outgoing edges in the direction (if any) of the edge.
    • Method Detail

      • isDirected

        boolean isDirected()
        Returns true if the edges in this graph are directed. Directed edges connect a source node to a target node, while undirected edges connect a pair of nodes to each other.
      • allowsSelfLoops

        boolean allowsSelfLoops()
        Returns true if this graph allows self-loops (edges that connect a node to itself). Attempting to add a self-loop to a graph that does not allow them will throw an IllegalArgumentException.
      • incidentEdges

        Set<EndpointPair<N>> incidentEdges​(N node)
        Returns the edges in this graph whose endpoints include node.

        This is equal to the union of incoming and outgoing edges.

        Throws:
        IllegalArgumentException - if node is not an element of this graph
        Since:
        24.0
      • degree

        int degree​(N node)
        Returns the count of node's incident edges, counting self-loops twice (equivalently, the number of times an edge touches node).

        For directed graphs, this is equal to inDegree(node) + outDegree(node).

        For undirected graphs, this is equal to incidentEdges(node).size() + (number of self-loops incident to node).

        If the count is greater than Integer.MAX_VALUE, returns Integer.MAX_VALUE.

        Throws:
        IllegalArgumentException - if node is not an element of this graph
      • inDegree

        int inDegree​(N node)
        Returns the count of node's incoming edges (equal to predecessors(node).size()) in a directed graph. In an undirected graph, returns the degree(Object).

        If the count is greater than Integer.MAX_VALUE, returns Integer.MAX_VALUE.

        Throws:
        IllegalArgumentException - if node is not an element of this graph
      • outDegree

        int outDegree​(N node)
        Returns the count of node's outgoing edges (equal to successors(node).size()) in a directed graph. In an undirected graph, returns the degree(Object).

        If the count is greater than Integer.MAX_VALUE, returns Integer.MAX_VALUE.

        Throws:
        IllegalArgumentException - if node is not an element of this graph
      • hasEdgeConnecting

        boolean hasEdgeConnecting​(N nodeU,
                                  N nodeV)
        Returns true if there is an edge that directly connects nodeU to nodeV. This is equivalent to nodes().contains(nodeU) && successors(nodeU).contains(nodeV).

        In an undirected graph, this is equal to hasEdgeConnecting(nodeV, nodeU).

        Since:
        23.0
      • hasEdgeConnecting

        boolean hasEdgeConnecting​(EndpointPair<N> endpoints)
        Returns true if there is an edge that directly connects endpoints (in the order, if any, specified by endpoints). This is equivalent to edges().contains(endpoints).

        Unlike the other EndpointPair-accepting methods, this method does not throw if the endpoints are unordered and the graph is directed; it simply returns false. This is for consistency with the behavior of Collection.contains(Object) (which does not generally throw if the object cannot be present in the collection), and the desire to have this method's behavior be compatible with edges().contains(endpoints).

        Since:
        27.1
      • equals

        boolean equals​(@Nullable Object object)
        Returns true iff object is a Graph that has the same elements and the same structural relationships as those in this graph.

        Thus, two graphs A and B are equal if all of the following are true:

        Graph properties besides directedness do not affect equality. For example, two graphs may be considered equal even if one allows self-loops and the other doesn't. Additionally, the order in which nodes or edges are added to the graph, and the order in which they are iterated over, are irrelevant.

        A reference implementation of this is provided by AbstractGraph.equals(Object).

        Overrides:
        equals in class Object
        Parameters:
        object - the reference object with which to compare.
        Returns:
        true if this object is the same as the obj argument; false otherwise.
        See Also:
        Object.hashCode(), HashMap