Version based prediction of LUIS is not accessible

Question

I have a LUIS application which has multiple versions named, v1, v2, and v3.

v1 is used in Production slot; v2 is used in Staging Slot; v3 is still in development state.

        LuisRecognizerOptionsV3 recognizerOptions = new LuisRecognizerOptionsV3(luisApplication)
        {
            TelemetryClient = telemetryClient,
            IncludeAPIResults = true,
            PredictionOptions = new Microsoft.Bot.Builder.AI.LuisV3.LuisPredictionOptions()
            {
                Version = "v3", // Just not seem to work. 404 exception.
                IncludeInstanceData = true,
                IncludeAllIntents = true,
                PreferExternalEntities = true,
            }
        };

1. Is it possible to use the luis model version v3 without publishing to any slot?
2. Is it mandatory to use define both Slot and Version attributes?
3. I get 404 error incase if I use only version without publishing it to any slot.
4. Already the endpoints are made public in all versions but still error.

How to overcome these challenges?

Thanks, pdeepa

Answer 1

Iterative app design for LUIS: https://learn.microsoft.com/en-us/azure/cognitive-services/LUIS/luis-concept-app-iteration

Trained versions aren't automatically available at your LUIS app's endpoint. You must publish or republish a version in order for it to be available at your LUIS app endpoint.

If more versions of the app need to be available at an endpoint, you should export the version and reimport it to a new app. The new app has a different app ID.