How to run an App on Brainlife

Once your App is registered on Brainlife, the next step is to run it on Brainlife. As specified in How to register an App on Brainlife, at first the App is private (i.e. linked to a private project, only members of this project can view and execute this App).

The steps to run an App are presented more precisely in Brainlife documentation: data processing and processes.

There are two ways to run an App depending on if you run only one App or if you run small pipelines (e.g. a series of Apps):

1. If you want to run only one App:

   1. go directly to the App and then execute it by clicking on “Execute”

      

   2. BL creates a new process in the selected project, stages all of the input files you selected, and submits your App in a single step

2. If you want to run a small pipeline:

   1. go to the page of the project linked to the App and then go the tab “Processes”

      

      A Process is where you can submit one or a group of Apps that can share input/output datasets.

   2. create a new process

   3. stage any dataset from our archive to your process (each process can only process data that is either staged or generated by other Apps)

   4. select the App you want to run

   5. click submit

Results of the App

Once resources have been assigned to your App, you can run it. Your task will go through different statuses:

1. Requested: the task waits to be assigned to a resource (if there are a lot of tasks running on Brainlife at the same time, it may take several minutes)

2. Running: once a resource is free, your task will enter a queue for the resource and then start Running

3. Finished: the task has been completed, we can see the outputs files

4. Failed: an error occurred, see the error.log in outputs for details

If you have created a product.json, messages will be displayed directly in the Brainlife UI. For instance in the screenshot below, the success message “MaxFilter was applied successfully” was displayed. Error messages can be displayed as well if they are caught in the product.json (see product.json).