Discover Erdos-Graph-Framework/Erdos Open Source project

Erdos is a very light, modular and super easy to use modern Graph theoretic algorithms framework for Java. It contains graph algorithms that you can apply swiftly with one line of code and was primarily developed to back a worker manager tasks for various Java projects including one in Android.

Erdos was born because other frameworks in Java were very hard to get started with or just plain heavy (overhead wise) or just too opinionated. Erdos let the developer design it's own graph engines to optimise the runtime and comes with all of the graph algorithms you can expect.

How to use

Option 1: Jitpack

Add Jitpack in your root build.gradle at the end of repositories:

allprojects {
    repositories {
        ...
        maven { url "https://jitpack.io" }
    }
}

Add to your dependencies:

dependencies {
    compile 'com.github.Erdos-Graph-Framework:Erdos:v1.0'
    // or the following for the current master snapshot
    // compile 'com.github.Erdos-Graph-Framework:Erdos:master-SNAPSHOT'
}

Option 2: Simply fork or download the project, you can also download and create .jar file yourself, with

git clone https://github.com/Erdos-Graph-Framework/Erdos.git erdos
cd erdos
./gradlew build
ls -l build/libs

Option 3: grab the latest released jar

https://github.com/Erdos-Graph-Framework/Erdos/releases

Notable technical features

compatible with Java 7
compose your graph by features and engine. modular design.
production proved code. Used in a commercial project.

Supported graphs

simple graph, directed and undirected
multi edge graph, directed and undirected
pseudo graph
custom graph, configure by self-loops, multi-edges and a graph engine.

builtin algorithms

Erdos is a framework to easily help you design graph algorithms with the correct abstractions and utilities. The builtin algorithms are:

search algorithms
- BFS - Breadth First Search
- DFS - Depth First Search
sorting
- Topological sorting
structure
- Strongly Connected Components
minimum spanning tree
- Prim's algorithm
- Kruskal's algorithm
single source shortest path
- Bellman–Ford algorithm
- Dijkstra's algorithm
- Directed acyclic graph algorithm
all pairs shortest path
- Floyd–Warshall algorithm
- Johnson's algorithm
minimum spanning tree
- Prim's algorithm
- Kruskal's algorithm
transformations
- Square of a graph
- transitive closure of a graph
- transpose of a graph

builtin graph engines

Adjacency and Incidence list based graph engine
designed for optimal complexity for algorithms that require more than a moderate edge queries.
in the future, a adjacency matrix engine will be added. That will be good to certain types of graphs, where queries are small, and memory should be kept as small as possible.
you can add your own graph engine by implementing AbstractGraphEngine.

Instructions, code by examples

1. creating a very simple graph

SimpleDirectedGraph graph_triangle = new SimpleDirectedGraph();

Vertex v0 = new Vertex();
Vertex v1 = new Vertex();
Vertex v2 = new Vertex("tag_v2");

graph_triangle.addVertex(v0);
graph_triangle.addVertex(v1);
graph_triangle.addVertex(v2);

Edge e_0 = graph_triangle.addEdge(v0, v1);
graph_triangle.addEdge(v1, v2);
graph_triangle.addEdge(v2, v3);

graph_triangle.print();

// iterate the graph vertices directly
for (IVertex vertex : graph_triangle) {
    System.out.println(vertex.toString());
}

// iterate the edges of the graph
for (Edge edge : graph_triangle.edges()) {
    System.out.println(edge.toString());
}

// removing a vertex in any of the following ways will remove it's connected edges as well,
// also removing any edge in similar fashion will update the graph :)
graph_triangle.removeVertex(v0);
graph_triangle.vertices().remove(v1);
graph_triangle.vertices().iterator().remove();

2. use a factory for custom graph

you can define your graph in terms of self loops, multi edges (per vertex) and a custom implementation of a graph engine.

boolean allow_self_loops = true;
boolean allow_multi_edges = true;

UndirectedGraph graph_undirected = Erdos.newUndirectedGraphWithEngine(new AdjIncidenceGraphEngine(), 
                                                                      allow_self_loops, allow_multi_edges);

DirectedGraph graph = Erdos.newGraphWithEngine(new AdjIncidenceGraphEngine(), 
                                               Edge.EDGE_DIRECTION.DIRECTED,
                                               allow_self_loops, allow_multi_edges);

3. algorithms introduction

every algorithm extends AbstractGraphAlgorithm<T, E extends IGraph>, which is generically typed E for input graph and T for output and must implement

T applyAlgorithm() method

for example, the Bellman-Ford algorithm for single source shortest path, followed by the Floyd-Warshall algorithm for all pairs shortest paths.

private void BellmanFord()
{
    SimpleDirectedGraph graph = new SimpleDirectedGraph();

    Vertex s = new Vertex("s");
    Vertex t = new Vertex("t");
    Vertex x = new Vertex("x");
    Vertex y = new Vertex("y");
    Vertex z = new Vertex("z");

    graph.addVertex(s);
    graph.addVertex(t);
    graph.addVertex(x);
    graph.addVertex(y);
    graph.addVertex(z);

    graph.addEdge(s, t, 6);
    graph.addEdge(t, x, 5);
    graph.addEdge(x, t, -2);
    graph.addEdge(s, y, 7);
    graph.addEdge(y, z, 9);
    graph.addEdge(t, y, 8);
    graph.addEdge(z, x, 7);
    graph.addEdge(t, z, -4);
    graph.addEdge(y, x, -3);
    graph.addEdge(z, s, 2);

    graph.setTag("graph");
    graph.print();

    // apply the Bellman-Ford algorithm
    ShortestPathsTree res = new BellmanFordShortestPath(graph).setStartVertex(s).applyAlgorithm();
    // print it
    res.print();
    // apply the Floyd-Warshall algorithm
    AllPairsShortPathResult floyd_result = new FloydWarshall(graph).applyAlgorithm();
    // print the shortest paths tree of the vertex
    floyd_result.shortestPathsTreeOf(s).print();
    // print the shortest path between two nodes
    System.out.println(floyd_result.shortestPathBetween(s, z).toString());
}

4. algorithms, more examples

this example shows the simplicity of the framework (hopefully ;)) where we apply 5 different algorithms sequentally

// perform a breadth first search
BFS.BreadthFirstTree breadthFirstTree = new BFS(graph, s).applyAlgorithm();
// perform a depth first search
DFS.DepthFirstForest depthFirstForest = new DFS(graph).applyAlgorithm();
// extract the strongly connected components of the graph
ArrayList<HashSet<IVertex>> hashSets = new SCC(graph).applyAlgorithm();
// perform a topological sort on the graph
LinkedList<IVertex> res_sort = new TopologicalSort(graph).applyAlgorithm();
// compute all pairs shortest paths using the Floyd-Warshall algorithm
AllPairsShortPathResult floyd_result = new FloydWarshall(graph).applyAlgorithm();

5. algorithms factories

for major algorithms types, you can comfortably use the following algorithms factories

MinSpanTreeFactory - for Minimum Spanning Tree/Forest, for example:

AbstractGraphAlgorithm<UndirectedGraph, IUndirectedGraph> alg = MinSpanTreeFactory.newMST(graph, MstAlgorithm.KRUSKAL, start_vertex);
AbstractGraphAlgorithm<UndirectedGraph, IUndirectedGraph> alg2 = MinSpanTreeFactory.newMST(graph, MstAlgorithm.PRIM, start_vertex);

SingleSourceShortPathFactory - for single source shortest path, for example:

AbstractShortestPathAlgorithm alg = SingleSourceShortPathFactory.newSingleSourceShortPath(graph, SSSPAlgorithm.DIJKSTRA, start_vertex, end_vertex);

AllPairsShortPathFactory - for shortest paths between all pairs, for example:

AbstractGraphAlgorithm<AllPairsShortPathResult, IDirectedGraph> alg2 = AllPairsShortPathFactory.newAllPairsShortPath(graph, APSPAlgorithm.Johnson);```

6. utilities

a bunch of helper utilities can be found in the package com.hendrix.erdos.utils

SVertexUtils.java - query vertex order information inside a graph
SEdgeUtils.java - query edge order information inside a graph
SMatrixUtils.java - compute the adjacency and incidence matrix of a graph
SGraphUtils.java - get a sorted list of the weighted edges in a graph

used in

Android-Zorn - for constructing a worker manager based on topological sorting in a graph.

Contributions

contributions are most welcomed, please consult CONTRIBUTING.md

what is the Erdős ?

homage to the great hungarian mathmatician Paul Erdős who part of his research was rooted in combinatorial graph theory.

License

If you like it -> star or share it with others

MIT License

Copyright (c) 2017 Tomer Shalev and Erdos (https://github.com/HendrixString)

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

Erdos-Graph-Framework/Erdos

Erdos-Graph-Framework

Reviews

Repository Details