History¶
0.3.0 (2021-09-09)¶
- Deprecate Python 3.6 support, add Python 3.9 support (GH#248)
- Using contextlib.suppress instead of deprecated dask.util.ignoring in EstimatingProgressBar (GH#256)
- Disallow numba 0.54.0 (GH#254)
- Upgrade to CuPy 9.0 and fix template encoding (GH#251)
- Parse and zero spectral models containing ‘nan’ and ‘inf’ in wsclean model files (GH#250)
- Clarify _wrapper names (GH#247)
- Baseline-Dependent Time-and-Channel Averaging (GH#173, GH#243)
0.2.10 (2021-02-09)¶
- Don’t let dof go to zero during spi fitting (GH#240)
- Add support for Shapelets and Zernike Polynomials (GH#231)
- Add beam model during SPI fitting (GH#238)
- Add double accumulation option and Hessian function to wgridder (GH#237)
- Upgrade ducc0 to version 0.8.0 (GH#236)
- Add mindet to avoid div0 errors in spi fitter and fix alpha and I0 variance estimates (GH#234)
0.2.9 (2020-12-15)¶
0.2.8 (2020-10-08)¶
0.2.7 (2020-09-23)¶
0.2.6 (2020-08-07)¶
0.2.5 (2020-07-01)¶
0.2.4 (2020-05-29)¶
- Support overriding the l and m axis sign in beam_grids (GH#199)
- Upgrade to python-casacore 3.3.1 (GH#197)
- Upgrade to jax 0.1.68 and jaxlib 0.1.47 (GH#197)
- Upgrade to scipy 1.4.0 (GH#197)
- Use github workflows (GH#196, GH#197, GH#198, GH#199)
- Make CASA parallactic angles thread-safe (GH#195)
- Fix spectral model documentation (GH#190), to match changes in (GH#189)
0.2.3 (2020-05-14)¶
- Fix incorrect SPI calculation and make predict defaults MeqTree equivalent (GH#189)
- Depend on pytest-flake8 >= 1.0.6 (GH#187, GH#188)
- MeqTrees Comparison Script Updates (GH#160)
- Improve requirements handling (GH#187)
- Use python-casacore wheels for travis testing, instead of kernsuite packages (GH#185)
0.2.1 (2020-04-03)¶
- Update trove to latest master commit (GH#178)
- Added Cubic Spline support (GH#174)
- Depend on python-casacore >= 3.2.0 (GH#172)
- Drop Python 3.5 support and test Python 3.7 (GH#168)
- Implement optimised WSClean predict (GH#166, GH#167, GH#177, GH#179, GH#180, GH#181)
- Simplify dask predict_vis code (GH#164, GH#165)
- Document and check weight shapes in simple gridder and degridder (GH#162, GH#163)
- Restructuring calibration module (GH#127)
- Upgrade to numba 0.46.0, using new inlining functionality in the RIME and averaging code.
- Modified predict to be compatible with eidos fits headers (GH#158)
0.2.0 (2019-09-30)¶
- Added standalone SPI fitter (GH#153)
- Fail earlier and explain duplicate averaging rows (GH#155)
- CUDA Beam Implementation (GH#152)
- Fix documentation package versions (GH#151)
- Deprecate experimental w-stacking gridder in favour of nifty gridder (GH#148)
- Expand travis build matrix (GH#147)
- Drop Python 2 support (GH#146, GH#149, GH#150)
- Support the beam in the predict example (GH#145)
- Fix weight indexing in averaging (GH#144)
- Support EFFECTIVE_BW and RESOLUTION in averaging (GH#144)
- Optimise predict_vis jones coherency summation (GH#143)
- Remove use of @wraps (GH#141, GH#142)
- Set row chunks to nan in dask averaging code. (GH#139)
- predict_vis documentation improvements (GH#135, GH#140)
- Upgrade to dask-ms in the examples (GH#134, GH#138)
- Explain how to obtain predict_vis time_index argument (GH#130)
- Update RIME predict example to support Tigger LSM’s and Gaussians (GH#129)
- Add dask wrappers for the nifty gridder (GH#116, GH#136, GH#146)
- Testing and requirement updates. (GH#124)
- Upgraded DFT kernels to have a correlation axis and added flags for vis_to_im. Added predict_from_fits example. (GH#122)
- Fixed segfault when using _unique_internal on empty ndarrays (GH#123)
- Removed apply_gains. Use africanus.calibration.utils.correct_vis instead (GH#118)
- Add streams parameter to dask predict_vis (GH#118)
- Implement the beam in numba (GH#112)
- Add residual_vis, correct_vis, phase_only_GN (GH#113)
0.1.8 (2019-05-28)¶
0.1.7 (2019-05-09)¶
- Even more support for automated travis deploys.
0.1.6 (2019-05-09)¶
- Support automated travis deploys.
0.1.5 (2019-05-09)¶
- Predict script enhancements (GH#103) and dask channel chunking fix (GH#104).
- Directly jit DFT functions (GH#100, GH#101)
- Spectral Models (GH#86)
- Fix radec sign conversion in wsclean sky model (GH#96)
- Full Time and Channel Averaging Implementation (GH#80, GH#97, GH#98)
- Support integer seconds in wsclean ra and dec columns (GH#91, GH#93)
- Fix ratio computation in Gaussian Shape (GH#89, GH#90)
0.1.4 (2019-03-11)¶
- Support complete and complete-cuda to support non-GPU installs (GH#87)
- Gaussian Shape Parameter Implementation (GH#82, GH#83)
- WSClean Spectral Model (GH#81)
- Compare predict versus MeqTrees (GH#79)
- Time and channel averaging (GH#75)
- cupy implementation of predict_vis (GH#73)
- Introduce transpose in second antenna term of predict (GH#72)
- cupy implementation of feed_rotation (GH#67)
- cupy implementation of stokes_convert kernel (GH#65)
- Add a basic RIME example (GH#64)
- requires_optional accepts ImportError’s for a better debugging experience (GH#62, GH#63)
- Added fit_component_spi function (GH#61)
- cupy implementation of the phase_delay kernel (GH#59)
- Correct phase_delay argument ordering (GH#57)
- Support dask for radec_to_lmn and lmn_to_radec. Also add support for radec_to_lm and lm_to_radec (GH#56)
- Added test for dft to test if image space covariance is symmetric(GH#55)
- Correct Parallactic Angle Computation (GH#49)
- Enhance visibility predict (GH#50)
- Fix Kaiser Bessel filter and taper (GH#48)
- Stokes/Correlation conversion (GH#41)
- Fix gridding examples (GH#43)
- Add simple dask gridder example (GH#42)
- Implement Kaiser Bessel filter (GH#38)
- Implement W-stacking gridder/degridder (GH#38)
- Use 2D filters by default (GH#37)
- Fixed bug in im_to_vis. Added more tests for im_to_vis. Removed division by \(n\) since it is trivial to reinstate after the fact. (GH#34)
- Move numba implementations out of API functions. (GH#33)
- Zernike Polynomial Direction Dependent Effects (GH#18, GH#30)
- Added division by \(n\) to DFT. Fixed dask chunking issue. Updated test_vis_to_im_dask (GH#29).
- Implement RIME visibility predict (GH#24, GH#25)
- Direct Fourier Transform (GH#19)
- Parallactic Angle computation (GH#15)
- Implement Feed Rotation term (GH#14)
- Swap gridding correlation dimensions (GH#13)
- Implement Direction Dependent Effect beam cubes (GH#12)
- Implement Brightness Matrix Calculation (GH#9)
- Implement RIME Phase Delay term (GH#8)
- Support user supplied grids (GH#7)
- Add dask wrappers to the gridder and degridder (GH#4)
- Add weights to gridder/degridder and remove PSF function (GH#2)
0.1.2 (2018-03-28)¶
- First release on PyPI.