Fleur Workchains

General design

All of the WorkChains have a similar interface and they share several common input nodes. First, all of the workflows accept FLEUR and inpgen code nodes as an input. inpgen node is not required since sometimes there is no need to generate a new inp.xml file.

There is always an optional wf_parameters: Dict node for controlling the workflow which has some reasonable defaults. It contains all the parameters related to physical aspects of the workchain. That is why the content of wf_parameters vary between different workchains.


There is always an inpxml_changes nested list that can be specified in the wf_parameters. This list can be used to apply any supported changes into inp.xml that do not have a shortcut in the workchain.

The other common input is an options: Dict node where the technical parameters (AiiDA options) are specified i.e resources, queue name and so on.

Regarding an input crystal structure, it can be set in two ways in the most of the workflows:

  1. Provide a structure: StructureData node and an optional calc_parameters: Dict. It this case an inpgen code node is required. The workflow will call inpgen calculation and create a new FleurinpData that will be used in the workchain. You can find more information in the inpgen calculation section.
  2. Provide a fleurinp: FleurinpData node which contains a complete input for a FLEUR calculation.

Next, a remote_data: RemoteData can be optionally given to use the last charge density or other files from this previous calculation. For example, it is used when one wants to start a SCF calculation from a given charge density to speed up the calculation.

Most of the workchains return a workflow specific ParameterData (Dict) node named output_name_wc_para which contains the main results and some information about the workchain.

There are additional workflow specific input and output nodes, please read the documentation of a particular workchain that you are interested in.

All of the workchains are split into the groups. First, we separate technical and scientific workflows. This separation is purely subjective: technical workchains tend to be less complex and represent basic routine tasks that people usually encounter. Scientific workflows make a particular physical task which is more specific to the particular project.

There are the sub-group of the force theorem calculations and their self-consistent analogs in the scientific workchains group.


The plot_fleur function provides a quick visualization for every workflow node or node list. Inputs are uuid, pk, workchain nodes or ParameterData (workchain output) nodes.