SoFiA (inactive)

Results 30 issues of SoFiA (inactive)

In principal it should be straightforward to support data cubes with more than one Stokes parameter as long as the Stokes axis comes last after the sky coordinates and frequency...

Enhancement
4 Low

It would be desirable to have one or two more metrics that characterise the local noise in the vicinity of detections, in particular metrics that are sensitive to non-Gaussianity such...

Enhancement
3 Medium

It would be useful to provide more flexible flagging options by allowing different flagging shapes, e.g. pixels, channels, circles on the sky, etc. This could be implemented through a more...

Enhancement
4 Low

Noise scaling is currently restricted to spectral or local, but it might be desirable to add further modes. Examples include the 1D-2D mode from SoFiA 1 where the noise is...

Enhancement
4 Low

An option should be implemented for dilating groups of pixels flagged by the auto-flagger (SoFiA 1 parameter `filterArtefacts.dilation`) by a user-specified amount. As the auto-flagger in SoFiA 2 is far...

Enhancement
3 Medium

An option should be created to allow users to import a flagging cube (SoFiA 1 parameter `flag.file`). All pixels flagged in that flagging cube should then also be flagged in...

Enhancement
4 Low

When loading data, it should be possible to specify a region not just in pixel coordinates, but also in WCS coordinates (SoFiA 1 parameter `import.subcubeMode`). WCS regions should be specified...

Enhancement
3 Medium

Implement option to specify list of source IDs to be processed when a mask cube is loaded (SoFiA 1 parameter `import.sources`). All sources not on that list should be discarded...

Enhancement
4 Low

The purpose of this issue is to capture the idea of having an algorithm that automatically shifts source masks into areas of the cube that don’t contain sources. This would...

Enhancement
4 Low

The linker is the only time-critical algorithm left that has not yet been parallelised. Parallelisation is hampered by the recursive nature of the algorithm, but a possible solution might be...

Enhancement
3 Medium