Skip to content

Latest commit

 

History

History
87 lines (71 loc) · 3.67 KB

README.md

File metadata and controls

87 lines (71 loc) · 3.67 KB

Voron.Graph

A lightweight persisted graph library, based on Voron - a new transactional key value store developed from scratch by Hibernating Rhinos.
It is a pet project - a sandbox to play with graph theory related stuff, and perhaps also make something usefull out of it.
By something useful I mean easy to use rich graph library functionality, with persisted graph data, and without the need to load entire graph into memory.
(with emphasis on easy to use :) )

Note : this project is work in progress and is far from finished

Compiling and doing something useful with this project

Voron.Graph depends on Voron storage, and if you choose to download and compile, you need to:

  • Get Voron sources
  • Update references to Voron in the project
  • Compile!

Show me the code

And how do I use it?
Usage of the library is simple. The following code creates graph, creates hierarchy of objects and then queries for adjacent nodes of a certain node.
All code snippets presented here are taken from unit tests with minor adaptation

using (var storage = new StorageEnvironment(StorageEnvironmentOptions.CreateMemoryOnly()))
{
  var graph = new GraphStorage("TestGraph", storage);
  Node node1, node2, node3;

  using (var tx = graph.NewTransaction(TransactionFlags.ReadWrite))
  {
    node1 = graph.Commands.CreateNode(tx, JsonFromValue("test1"));
    node2 = graph.Commands.CreateNode(tx, JsonFromValue("test2"));
    node3 = graph.Commands.CreateNode(tx, JsonFromValue("test3"));

    graph.Commands.CreateEdgeBetween(tx, node3, node1);
    graph.Commands.CreateEdgeBetween(tx, node3, node2);

    //looping edge also ok!
    //adding multiple loops will overwrite each other
    graph.Commands.CreateEdgeBetween(tx, node2, node2);
    
    tx.Commit();
  }

  using (var tx = graph.NewTransaction(TransactionFlags.Read))
  {
    var adjacentNodes = graph.Queries.GetAdjacentOf(tx, node3).ToList();
    adjacentNodes.Select(x => x.Key).Should().Contain(new[] { node1.Key, node2.Key });
  }
}

#### Algorithms Using algorithm implementations in Voron.Graph is also simple.
In this code snippet a graph with hierarchy is created, and then with shortest & cheapest path is found between node1 and node4
*Assume that Env is StorageEnvironment of the Voron that was initialized earlier.*
```C# var graph = new GraphStorage("TestGraph", Env);

Node node1, node2, node3, node4; using (var tx = graph.NewTransaction(TransactionFlags.ReadWrite)) { node1 = graph.Commands.CreateNode(tx, JsonFromValue(1)); node2 = graph.Commands.CreateNode(tx, JsonFromValue(2)); node3 = graph.Commands.CreateNode(tx, JsonFromValue(3)); node4 = graph.Commands.CreateNode(tx, JsonFromValue(4));

node1.ConnectWith(tx, node2, graph, 1); node2.ConnectWith(tx, node3, graph, 1); node3.ConnectWith(tx, node4, graph, 1);

node1.ConnectWith(tx, node4, graph, 10);

tx.Commit(); }

using (var tx = graph.NewTransaction(TransactionFlags.Read)) { var shortestPathAlgorithm = new DijkstraShortestPath(tx, graph, node1, cancelTokenSource.Token); var shortestPathsData = shortestPathAlgorithm.Execute();

var shortestNodePath = shortestPathsData.GetShortestPathToNode(node4); shortestNodePath.Should().ContainInOrder(node1.Key, node2.Key, node3.Key, node4.Key); }


####License
Apache v2 License for the code of this project. Probably having/not having a license doesn't matter at this point, but still, just in case, I've added one. <br/>
About licensing of the Voron project - you need to contact [Hibernating Rhinos](http://hibernatingrhinos.com/) to inquire more about it.