What is ANTS?

ANTS is a Python library that provides interfaces and routines for carrying out common operations needed to produce ancillary files. It is built on top of Iris and uses Mule to write Unified Model (UM) ancillary files.

In addition, some standard applications are provided to support:

• regridding of data

• filling data

• merging datasets

• converting files to ancillary file format

• generation of polygon files that may be needed for ancillary generation

What is ANTS not?

While ANTS provides a library for carrying out common operations needed for ancillary generation, it is:

• not where ancillary science routines live

• not where ancillary generation workflows live

• not the library for generation of Unstructured Grid ancillaries

Related areas

When carrying out work generating or writing new ancillary generation code, a number of areas beyond the ANTS library itself will likely be encountered.

Ancillary-file-science

The ancillary-file-science repository is a central location for the lodging of ANTS based ancillary science routines. The codes in there are owned by scientists, with technical advice and assurance provided by the ANTS developers. A rose stem workflow exists in ancillary-file-science to enable routine testing of the ancillary science routines lodged there - known as “Apps” - as updates are made to the ANTS library.

Datasets

In order to generate an ancillary file, you will normally need some input data which is then passed to an ancillary science routine to produce the required outputs. Such datasets - commonly refered to as “master files” - may come in a variety of formats, from a range of different sources, with differing structures even for the same sort of input domain. As such, some “pre-processing” will typically be required in order to get the source dataset in an appropriate state for use by the ancillary science routines it is to be processed by. It is likely that when using a new dataset you will need to write a new “pre-processor” for it. You will want to refer to the relevant ancillary-file-science App to understand what is needed and there may be related pre-processors available that you could use as starting points.

Ideally, files input into ancillary science routines should be in NetCDF format as ANTS (and Iris) is well optimised for working with this format. It also has the added benefit of allowing metadata to be recorded with the data itself - allowing capture of provenance, licensing, and distribution information.

Ancillary Workflows

Ancillary Workflows are cylc based workflows (suites) that are owned, maintained, and developed by science teams and are not maintained by the ANTS developers. These workflows will often include one or more jobs running ancillary-file-science Apps to process “Master Files” into ancillary files, and are the typical entry point for someone wanting to produce a new ancillary file rather than the ANTS library itself. Examples of these are the Global Atmosphere (GA), Regional Atmosphere (RA), and JULES suites.

The CAP

The CAP was a Fortran based libary previously used to generate Unified Model ancillary files. It is, however, long since unmaintained and unsupported. Much work has been put into providing ancillary-file-science implementation of equivalent, or newer, science to enable migration away from the CAP to ANTS based ancillary-file-science implementations.

UG-ANTS / UG-Ancillary-file-science

With the move to LFRic, our models move to using a cubesphere grid and it becomes necessary to have tooling and write new science routines to support the new grid and model requirements. ANTS does not support generation of the unstructured grid ancillary files needed for this. Instead, a separate UG-ANTS and associated ug-ancillary-file-science are being developed. Due to the various complexities both in terms of implementations needed, along with the structure of the existing ANTS and ancillary-file-science codes, an active decision was made to develop these as distinct, separate things. These libraries follow the same principles as regular ANTS and ancillary-file-science - UG-ANTS is the underpinning Python library, and ug-ancillary-file-science is where ancillary science routines for generating UGrid ancillary files live.

Contacting Us

If you want support with ANTS, want to report a bug, or are interested in contributing new functionality feel free to contact us.