Update to Forio Epicenter with autoRestore

#1

Hello Everybody,

A frequent question we hear from customers concerns how to handle timeouts in simulations. By default, simulations “time out” after 30 minutes of non-use. The simulation pages will continue to load, but the model is removed from memory and data cannot be changed or directly accessed with the current run. This can result in an eerie situation in which a page half loads and half doesn’t, with it being difficult to figure out why.

I’m pleased to note that we now can banish these irritating poltergeists with a “autoRestore” feature for Epicenter applications. If a model times out, setting a variable or calling an operation will automatically reload the model into memory. The user may see a brief pause while the current run is restored, but following that the run will be exactly where it had left off before the timeout. Often users will not even notice there’s been any change.

Take a look at the documentation for details on how this feature works. It works automatically for setting variable and calling operations. (You may want your application to do the same when retrieving data, but this requires some extra configuration.)

As always, Epicenter release notes this and other releases are available at:
https://forio.com/epicenter/docs/public/releases/

I’ve included the release notes for v1.34 below.

Happy Halloween! :ghost:

WILL

##v1.34 Release Notes

Release Date: 10/27/2015

###Updates and Improvements

  • Automatically bring runs from the database into memory.

    Runs must be in memory in order for you to update variables or call operations on them. (See more information on run persistence.)

    In previous releases, a run that had been removed from the current session had to be explicitly replayed in order to bring it from the database back into memory.

    New in this release, runs are automatically replayed whenever you attempt to update variables (PATCH) or call operations (POST).

    Additionally, you can optionally replay a run, bringing it into memory, whenever you look it up (GET). To replay on GET requests, add a header to your request.

    You can control the behavior for PATCH, POST, and GET in your model.cfg file by including your autoRestore preference. (See more on configuration files for SimLang, Julia, Python, R, or Vensim.)

  • Copy decisions from one run to another when cloning runs.

    You can now use the Model State API to clone a run (that is, make a new run based on an existing run) AND specify that only the final values of the decisions and variables are copied to the new run.

    The typical use case is an end user interacting with a project and advancing a run to completion — for example, changing interest rates repeatedly and watching the total investment values change over time — and then wanting to “play again” starting with the final decisions from the previous run. This is most common for SimLang and Vensim models.

    To clone a run and have only the final values of decisions copied to the new run, specify in the clone request that you want to exclude the operations that advance the model. See more information and details of the API call.

###Notable Bug Fixes

  • The Support Forum now correctly detects whether or not you are already logged into the Epicenter authoring interface.
  • Improved error messages for some of the REST APIs.
  • Several minor fixes for performance and stability.
1 Like