Log in to your Fractal server (e.g. https://fractal-beta.mls.uzh.ch/)

Create a new Project on the Projects page

Open the project & create a new dataset

1. The output of your workflow will be stored in the Zarr dir. By default, the Zarr dir will be set to your project directory.

   

Create a new workflow or import an existing workflow

1. Creating a new workflow will create an empty workflow for you.

2. Choose from available existing workflows you have or get one from Example workflows

   

On the workflow page:

1. On the top, you see which dataset is currently selected for processing (⇒ where the output of the workflow will be put)

2. You see the list of task in the sequence they will be run on the left

   

3. By clicking on a task, you see all its parameters and can change what parameters the task will run with. The (i) icons provide help text for each parameter

   

Click “Run workflow” to submit the workflow to the cluster

1. This runs the workflow for the selected dataset
2. You can decide which task to start from (especially useful if you return to a workflow that you have partially run before) & until which task the workflow should run (by default, the whole workflow will run at once. But you may want to run it in smaller subsets to check the outcomes first)
3. Click Run, then Confirm

You can add additional tasks at the end: Click on the plus next to workflow sequence

1. These tasks get added at the end of your workflow. You can set their parameters and then continue the workflow

Continue a workflow:

1. You can use the “Continue workflow” button to continue with an existing workflow, including all the filters that were set on the datasets and available images on the dataset
2. You can use the “Restart workflow” button if you want to start from the beginning. This resets your dataset, such that it “forgets” about existing OME-Zarrs, resets all filters etc.