Dependency Edge Coverage example

This is an example of how to use GraphWalker dependency edge coverage stop condition.

Edge Dependency Definition

In real product usage, not every edge from the designed (anticipated) model has the same importance. Some edges are more used than others, and therefore they have a higher business value. One way to depict the importance of a certain edge is by using the dependency value. The dependency of a certain edge shows how much the target vertex depends on the source vertex. The dependency values of the edges can be obtained by applying process mining techniques over production log files. With process mining (the heuristic miner) one can obtain a model of actual usage of a certain product, including the edge dependency. With the dependency edge coverage stop condition one can test only the "important" business edges, i.e. only the ones with high dependency. More info can be found in this presentation.

Example Model

JSON File Format

{
  "models": [
    {
      "name": "Dependency model",
      "generator": "random(dependency_edge_coverage(80))",
      "startElementId": "e0",
      "vertices": [
        {
          "name": "v_VerifySomeAction",
          "id": "n0"
        },
        {
          "name": "v_VerifySomeOtherAction",
          "id": "n1"
        }
      ],
      "edges": [
        {
          "name": "e_FirstAction",
          "id": "e0",
          "targetVertexId": "n0",
          "dependency": "100"
        },
        {
          "name": "e_AnotherAction",
          "id": "e1",
          "sourceVertexId": "n0",
          "targetVertexId": "n1",
          "dependency": "100"
        },
        {
          "name": "e_SomeOtherAction",
          "id": "e2",
          "sourceVertexId": "n1",
          "targetVertexId": "n1",
          "dependency": "85"
        },
        {
          "name": "e_SomeOtherAction",
          "id": "e3",
          "sourceVertexId": "n1",
          "targetVertexId": "n0",
          "dependency": "15"
        }
      ]
    }
  ]
}

Get the C# example source code

git clone https://github.com/GraphWalker/graphwalker-example.git
cd graphwalker-example/c-sharp-websocket/SmallModel

Run the test

In a terminal, start GraphWalker as a REST service:

java -jar graphwalker-cli-4.3.1.jar -d all online --service RESTFUL

Upload the model DependencyModel.json to the service using curl (or any tool that makes a http post request to the REST service):

curl -vH "Content-Type: text/plain" --data @DependencyModel.json http://localhost:8887/graphwalker/load
{"result":"ok"}

In the folder of graphwalker-example/c-sharp-websocket/SmallModel build the C# project, on Linux:

xbuild

Then run the test, on Linux:

mono SmallModel/bin/Debug/SmallModel.exe

What happened

After running the test we expect that graphwalker using this example will stop running after visiting the edges that have a dependency higher, or equal, to the one set in the model: "random(dependency_edge_coverage(80))". The results below show that only the edges with dependency higher, or equal, to 80%, such as e_FirstAction, e_AnotherAction, and e_SomeOtherAction with 85% (e2) have been visited.

SmallModel.e_FirstAction
{}
SmallModel.v_VerifySomeAction
{}
SmallModel.e_AnotherAction
{}
SmallModel.v_VerifySomeOtherAction
{}
SmallModel.e_SomeOtherAction
{}
SmallModel.v_VerifySomeOtherAction
{}
{
  "totalFailedNumberOfModels": 0,
  "totalNotExecutedNumberOfModels": 0,
  "totalNumberOfUnvisitedVertices": 0,
  "verticesNotVisited": [],
  "totalNumberOfModels": 1,
  "totalCompletedNumberOfModels": 1,
  "totalNumberOfVisitedEdges": 3,
  "totalIncompleteNumberOfModels": 0,
  "edgesNotVisited": [
    {
      "modelName": "Dependency model",
      "edgeId": "e3",
      "edgeName": "e_SomeOtherAction"
    }
  ],
  "result": "ok",
  "vertexCoverage": 100,
  "totalNumberOfEdges": 4,
  "totalNumberOfVisitedVertices": 2,
  "edgeCoverage": 75,
  "totalNumberOfVertices": 2,
  "totalNumberOfUnvisitedEdges": 1
}

Moreover, in the summary part next to the detailed results, it is concluded that the total number of completed models is 1, which is expected, the total number of visited edges is 3, and the total number of unvisited edges is 1, the e_SomeOtherAction that has a dependency of 15% (e3):

"modelName": "Dependency model",
"edgeId": "e3",
"edgeName": "e_SomeOtherAction"