jaycedowell commented 3 years ago

This PR updates the lsl.imaging.util.ImgWPlus class and the lsl.imaging._gridder module to address a dependence of source position on frequency found by @nsbruce. This problem was fixed by (1) switching over to the same sinc-windowed Kaiser-Bessel function used by wsclean in _gridder and (2) applying a correction for the gridding kernel to the final image in ImgWPlus.

jaycedowell commented 3 years ago

Why is the test suite failing? It looks like there are two failures, both in test_imaging tests. One is test_clean_sources and the other is test_clean_leastsq.

jaycedowell commented 3 years ago

Ah, the failures seem to be related to a failure trying to fit a Gaussian to the point spread function in the lsl.imaging.deconv module.

coveralls commented 3 years ago

Coverage increased (+0.0002%) to 82.611% when pulling e335dd2c5484bccde67dabda334fb49f8e7fec88 on gridder-fix into 67bb4c2fe2f64c1268288052d6e95fe3cdf3987a on master.

codecov[bot] commented 3 years ago

Codecov Report

Merging #20 (e335dd2) into master (67bb4c2) will decrease coverage by 0.20%. The diff coverage is 86.66%.

@@            Coverage Diff             @@
##           master      #20      +/-   ##
==========================================
- Coverage   83.13%   82.92%   -0.21%     
==========================================
  Files          66       66              
  Lines       21962    21968       +6     
==========================================
- Hits        18258    18218      -40     
- Misses       3704     3750      +46

Impacted Files	Coverage Δ
lsl/imaging/deconv.py	`87.63% <84.61%> (-0.05%)`	:arrow_down:
lsl/imaging/utils.py	`75.34% <100.00%> (ø)`
lsl/misc/ionosphere.py	`69.52% <0.00%> (-6.51%)`	:arrow_down:
lsl/common/progress.py	`87.43% <0.00%> (-2.10%)`	:arrow_down:

lwa-project / lsl

Gridder Fix #20

Codecov Report