Configuration: Filtering

Introduction

This page will explain to you the steps you will have to take in order to use the filtering system.
We will only cover features configurable pre-runtime via XML. In order to understand how to use dynamic filtering, look here.

Configuring Pipelines

Your XML should contain a <filter-pipeline/> top-level node. Inside of that node, you should find two subnodes:

1. <per-instance>: Corresponds to the Filter Pipeline applied on each individual MD instance after each coupling step.
2. <post-multi-instance>: Corresponds to the Filter Pipeline applied on averaged data acquired after multi-instance sampling.
   Both of the above may contain an optional attribute output, which is set to "md" by default, implying the unfiltered dataset from MD is sent to the CS.

Configuring Sequences

For each of the two Pipelines currently in use, you can add an arbitrary amount of Filter Sequences. For that, simply add a subnode to the corresponding Filter Pipeline. The following arguments are supported:

• filtered-values: Quantities of the cells that are filtered. Simply pass a string containing all of the below that you wish to have filtered.
  • micro-mass: microscopic mass
  • micro-momentum: microscopic momentum
  • macro-mass: macroscopic mass
  • macro-momentum: macroscopic momentum
  • potential-energy: potential energy
  • velocity: velocity
  • temperature: temperature
  • all: all of the above (default)
• domain-start and domain-end: Subselection of the MD-To-Macro-Domain. Pass XML-Style vector indices using mamico indexing.
  (cf. Cell Indexing Semantics) Default: Entire MD-To-Macro domain.
• input: String containing name(s) of input sequence(s). For normal Filter Sequences, this must be a singular input name. For use cases utilizing two or more inputs, cf. Filter Junctions and Asymmetrical Filter Junctions.
  Using "md" will set the unfiltered output of the MD solver as input. Default: "md"

The name of a sequence is the identifier of its XML tag.
The following names are reserved:

• "md"
• "outer"

The following name should be used with caution:

• SYNTHETHIC-MD (cf. Synthetic MD)

Configuring Filters

You can add a filter to a Filter Sequence by adding a subnode to the sequence's node. This subnode must have a name matching the identifier of a filter known to the filtering system. Filter variables are then set via XML attributes.
For an overview of filters, their identifiers and their parametrization options, have a look at this wiki's List of Filters.