Iterator workflow¶

Description¶

The SiestaIterator is a tool to facilitate the submission of several Siesta Calculations in an automatic way. It allows the iteration over Siesta parameters and, more in general, over inputs of a SiestaBaseWorkChain. An example on the use of the SiestaConverger is /aiida_siesta/examples/workflows/example_iterate.py.

Supported Siesta versions¶

At least 4.0.1 of the 4.0 series, 4.1-b3 of the 4.1 series and the MaX-1.0 release, which can be found in the development platform (https://gitlab.com/siesta-project/siesta). For more up to date info on compatibility, please check the wiki.

Inputs¶

All the SiestaBaseWorkChain inputs are as well inputs of the SiestaIterator, therefore the system and DFT specifications (structure, parameters, etc.) are inputted in the WorkChain using the same syntax explained in the SiestaBaseWorkChain documentation. The additional inputs are:

iterate_over, class Dict, Mandatory

A dictionary where each key is the name of a parameter we want to iterate over (str) and each value is a list with all the values to iterate over for the corresponding key. Accepted keys are:
- Name of the input ports of the SiestaBaseWorkChain. Meaning all the names listed here. In this case, the corresponding values list must contains the list of Data nodes (stored or unstored) accepted by the key. Examples are:
```
code1 = load_code("SiestaHere@localhost")
code2 = load_code("SiestaThere@remotemachine")
iterate_over = {"code" : [code1,code2]}

struct1 = StructureData(ase=ase_struct_1)
struct2 = StructureData(ase=ase_struct_2)
iterate_over = {"structure" : [struct1,struct2]}
```
- Name of accepted Siesta input keywords (for instance mesh-cutoff, pao-energy-shift, etc …). In this case, the corresponding values list must contains the list of values directly, meaning str, float, int or bool python types. Examples are:
```
iterate_over = {"spin" : ["polarized", "spin-orbit"]}
```
  Warning
  
  In order to guarantee full flexibility, no check on the Siesta parameters is performed. If you pass as key something not recognized by Siesta, the SiestaIterator will include it in the parameters input and run the calculation with no warning issued. Because Siesta will not understand the keyword, it will ignore it, resulting in a series of identical calculations.
The iterate_over is a dictionary because it is possible to iterate over several keywords at the same time. Something of this kind:
```
struct1 = StructureData(ase=ase_struct_1)
struct2 = StructureData(ase=ase_struct_2)
iterate_over = {"structure" : [struct1,struct2], "spin" : ["polarized", "spin-orbit"]}
```
is perfectly acceptable and the way the algorithm handle with these multiple iterations is decided by the SiestaIterator input explained next in this list.

iterate_mode, class Str, Optional

Indicates the way the parameters should be iterated. Currently allowed values are ‘zip’ (zips all the parameters together, this imposes that all keys should have the same number of values in the list!) and ‘product’ (performs a cartesian product of the parameters, meaning that all possible combinations of parameters and values are explored).

The option ‘zip’ is the default one.

batch_size, class Int, Optional

The maximum number of simulations that should run at the same time. You can set this to a very large number if you want that all simulations run in one single batch. As default, only one single calculation at the time is submitted.

Outputs¶

This WorkChain does not generate any output! It is, in fact, a tool to help the submission of multiple calculations and keep them all connected and easy accessible through the main workchain node, but it does not have any precise scope. AiiDA provides a powerful querying system to explore all the results of the submitted calculations and a tool to organize the data.

Protocol system¶

The protocol system is not directly available for this WorkChain. However inputs of the SiestaBaseWorkChain can be obtained in a dictionary in this way:

inp_gen = SiestaBaseWorkChain.inputs_generator()
inputs = inp_gen.get_inputs_dict(structure, calc_engines, protocols)

The inputs of get_inputs_dict are explained in the protocols documentation. Then the user must define at least the input iterate_over in order to be able to submit the SiestaIterator WorkChain.