DP3 version 5.0

This release adds preliminary support for Baseline Dependent Averaging (BDA), support for multiple prediction sources in DDECal, and support for a-terms when using IDG predictions. IDG prediction makes it computationally feasible to predict large source models.

Changes since v4.2:

• Add MSBDAReader, BDAAverager and MSBDAWriter steps that support BDA data.
• Support BDA data in the ScaleData step.
• DP3 now uses EveryBeam instead of LOFARBeam. This dependency is now required.
• DP3 now includes aocommon as an external git submodule.
• All code is now formatted using clang-format according to the Google C++ Style.
• CI improvements: Enforce code formatting, use SKA-compatible pipeline, add build without IDG, report test coverage.
• Enable various integration tests that were disabled.
• Convert FacetPredict into the IDGPredict step.
• Use new IDG bulk degridder in the IDGPredict step.
• Support a-terms in the IDGPredict step.
• Add ColumnReader step, for reading a specific column from a Measurement Set.
• Support multiple different prediction sources in DDECal.

DP3 version 4.2

This release contains a year worth of incremental improvements compared to the v4.1 branch. It is planned to be the last version that uses the LOFARBeam library. Instead, future releases will make use of the EveryBeam library, which contains the same functionality and adds support for other telescopes.

A summary of the changes:

• DDECal prediction using IDG, with spectral multi-term support
• Fixes to support latest casacore version
• Fixes to support g++ 10
• A 'setbeam' step to change metadata
• Improved error handling / reporting
• Better documentation
• More and improved unit tests
• Switched to Python 3
• Improved selection of reference antenna
• Bugfix: incorrect weighting of auto-correlations causes Dysco-compression to malfunction
• Add option msin.starttimeslot

DP3 version 4.1

This is an intermediate release.

DP3 version 4.0

This is the first release of the standalone version of DP3. We believe this version is stable and ready for production.

DP3 was extracted from and branched off LOFAR Release 3.2 (Sept 2018), with a minimum set of dependencies. Since then, it has undergone quite some changes both in its architecture as well as by adding some functionality. Some of the new functionality includes:

• More options & saver handling of beam (#68, #60, #80)
• Direct subtraction of model in DDECal (#25)
• Support for fulljones mode in H5ParmPredict (#9, #54)
• No more runtime linking issues thanks to direct internal linking of previously external steps (#52)
• A smoothness constraint

In total, 29 tickets were closed and 40 pull requests were merged.

To use the LOFAR beam in DP3, you will have to specify the location of the station response library during compilation. You can either install the station response library by compiling the full LOFAR repository, or by installing the standalone LOFARBeam package.