How to make a PMML persistent model using openscoring?

Question

I am using openscoring web service to score PMML models. However, I am not getting how to store those model. I am using this curl -X PUT --data-binary @DecisionTreeIris.pmml -H "Content-type: text/xml" http://localhost:8080/openscoring/model/DecisionTreeIris to put a model on server before model querying. My concern is when I stop the webserver and starts again, the model no longer exists. Is there any way to store these model so that I can Query using POST anytime after restarting a server?

Answer 1

Starting from Openscoring version 1.1.3, you can use the model auto-deployment directory option. Use the command-line argument --deploy-dir to tell the Openscoring server to start watching the specified directory. Then, you can 1) deploy a model by copying its PMML file to the auto-deploy directory and 2) undeploy the previously auto-deployed model by deleting its PMML file from the auto-deploy directory. There is no need to deal with cURL PUT and DELETE commands anymore.

However, please notice that the --deploy-dir command-line argument has been renamed to --model-dir in the GitHub trunk.