Frequently Asked Questions (FAQ)

How to obtain forces and stress of each ionic step?

By default, only the forces, stress and energies of the last ionic step are stored in the misc output. If you want those for each ionic step, you can modify the parser to enable the output trajectory node:

from aiida.orm import WorkflowFactory
VaspWorkChain = WorkflowFactory('vasp.vasp')
builder = VaspWorkChain.get_builder()
settings ={'parser_settings': {'include_node': ['trajectory']}}
builder.settings = settings

Why the parser system is so complex?

Unfortunately a single DFT calculation generates lots of data, only some of which are useful in most cases (i.e. energy, forces, stress) and it depends on the type of calculations run.

A further complication is that VASP generate multiple output files and some quantities are repeated in different files.

The role of the parser is to parse quantities from the files and organize them into different output nodes in a meaningful way. In addition, to avoid overfloing the database, some quantities / output nodes are excluded by default, and the user can choose to include them by setting 'include_quantity' or 'include_node' in the 'parser_settings' dictionary inside the settings input node of the VaspCalculation/VaspWorkChain.

The current logic of the parser system works like this:

• Instantiate all content parser for each kind of output file, which essentially parse everything and store them as their own attributes

• Collect quantities into a nested dictionary. All quantities declared as available by the content parser are collected unless they are explicitly excluded.

  • There is a default list of excluded quantities, and the user can include them by setting include_quantity: ['<quantitity'] in parser settings

  • The main reason for having a list of default excluded quantities is because some quantities are not needed in most cases, but the node containing them should still be created (e.g. the misc output).

• Try to compose all nodes except those are excluded by default (again, can be overridden by include_node: ['<node>']).

  • If a node cannot be composed due to lack of quantity, we simply skip it, as it is the responsibility of the CalcJob and the higher-level workflows to check for required output.

  • Again the reason for having excluded nodes is that some nodes are only needed in specific cases, but the underlying quantities are always available. For example, the 'eigenvalues' are avaliable in every single calculation, but they are mostly only needed for constructing the bands structure.

• Finally, we collect the compose nodes and store them under the outputs attribute which is a dictionary.