How to list all paths from a node to another in a indirect graph?

Question

I have a graph where its nodes are in the following structure:

Class Node:
   int source;
   int next;

So, I have the following nodes: [(1,2), (2,3), (3,4), (1,4)]

I wanted to list all possible paths from 1 to 3 it would list: [[(1,2),(2,3)],[(1,4),(4,3)].

I'm trying with this code, but I'm missing something:

public List<Node> getNodeNeighbors(Node n) {
    List<Node> filteredNodes = new ArrayList<Node>();
    List<Node> allNodes = (List<Node>) nodesRepository.findAll();
    for (Node node: allNodes) {
        if (node.source == n.next) {
            filteredNodes.add(node);
        }
    }
    return filteredNodes;
}

public List<Node> bfs(Node n, String destinationNodeNumber, List<Node> path) {
        visitedX.add(n); //visitedX is a global List to control visited nodes
        path.add(n); //local path to be listed
        List<Node> neighbors = getNodeNeighbors(n); //function to get node neighbors
        if (n.next.equals(destinationNodeNumber)) {
            allPaths.add(paths); //all paths to be listed
            path.remove(n);
        }
        for (Node nNode: neighbors) {
            if(!visitedX.contains(nNode)) {
                bfs(nNode, destinationNodeNumber, path);
            }
        }
        return null;
    }

Answer 1

There are many flaws in your code:

• the name of your class Node is misleading: Edge would be a better name,
• method getNodeNeighbors only considers one direction of each edge
• what are the fields aCompany and anotherCompany? i assume you meant source and next?
• what is class Contract?
• destinationNodeNumber is a String; it should be an int.
• the visitedX set prevents two paths from using a same edge; you just need to ensure that an edge doesn't appear more that once in a single path.
• you actually implemented a DFS, not a BFS
• you always add the same path to allPaths; you should make a copy instead.

Here is a class Edge:

public class Edge {
    final int source;
    final int next;

    Edge(int source, int next) {
        this.source = source;
        this.next = next;
    }

    @Override
    public String toString() {
        return "(" + source + "," + next + ')';
    }
}

Then the class Graph that contains the search algorithm:

public class Graph {
    private final Iterable<Edge> allNodes;

    public Graph(Iterable<Edge> allNodes) {
        this.allNodes = allNodes;
    }

    public List<Edge> edgesFrom(int vertex) {
        List<Edge> filteredNodes = new ArrayList<Edge>();
        for (Edge node : allNodes) {
            if (node.source == vertex || node.next == vertex) {
                filteredNodes.add(node);
            }
        }
        return filteredNodes;
    }

    public List<List<Edge>> allPaths(int source, int dest) {
        List<Edge> path = new ArrayList<>();
        List<List<Edge>> allPaths = new ArrayList<>();
        for (Edge n: edgesFrom(source)) {
            searchPaths(n, source, dest, path, allPaths);
        }            
        return allPaths;
    }

    private void searchPaths(Edge n, int source, int dest, List<Edge> path,
            List<List<Edge>> allPaths) {
        path.add(n); //local path to be listed
        int next = n.source == source ? n.next : n.source;
        List<Edge> neighbors = edgesFrom(next); //function to get node neighbors
        if (next == dest) {
            allPaths.add(new ArrayList<>(path)); //all paths to be listed
        }
        for (Edge nNode : neighbors) {
            if (!path.contains(nNode)) {
                searchPaths(nNode, next, dest, path, allPaths);
            }
        }
        path.remove(n);
    }
}

And here is the example that uses these classes:

Graph graph = new Graph(Arrays.asList(
        new Edge(1,2), new Edge(2,3), new Edge(3,4), new Edge(1,4)));
List<List<Edge>> allPaths = graph.allPaths(1,3);
for (List<Edge> path: allPaths) {
    System.out.println(path);
}