Viola is a streaming conformance checking approach which allows the discovery and conformance of a process, where the data is generated from IoT devices.

• Automatically derive a EGSM process model starting from a labeled sensor log
• Process a stream of unlabeled data, in order to be provided to the EGSM engine
• Recognize activities starting from a stream of data
• Verify the conformance at runtime

VIOLA requires Python and Node.js to run.

Pre-process the labeled sensor log in order to obtain a format suitable for the classification algorithm.

For example:

run "Experiments - Linac/normal/LogPreprocessing - Windowing.ipynb"

Derive a Decision Tree and convert it to rules.

For example:

run "Experiments - Linac/normal/Dtree - Classifier.ipynb"

Discover a Directly-Follows Graph.

For example:

python "Experiments - Linac/discover-dfg.py"

Create the EGSM process model.

For example:

python "Experiments - Linac/build-egsm.py"

Run the EGSM engine:

• Download the SMARTifact E-GSM Engine
• Copy the files siena.xml and infoModel.xsd produced during the offline phase inside the data/dfg folder (create the folder if does not exist)
• Start the engine:

  node server.js 

Process the stream of data:

• Run the CEP system.

For example:

python "Experiments - Linac/Server-lerror1.py"

• Start streaming events. Make a REST GET call to the following endpoint (e.g., with a web browser, cURL or Postman): http:https://localhost:8080/api/start

And visualize the results!

For example:

python "Experiments - Linac/validator.py"