- Include
PowerExp
1D model - Allow the surface-brightness used during the model construction to be defined separately from the binned surface-brightness data. This is useful for modeling the stellar kinematics.
- Allow the surface-brightness masks to be patched by a Gaussian smoothing algorithm to fill surface-brightness holes.
AxisymmetriDisk
now usesConvolveFFTW
by default.- Included analytic calculations of derivatives for use with the
AxisymmetricDisk
fits. Analytic calculations are used by default, but thelsq_fit
function allows for a fall back to the finite-difference methods provided byscipy.optimize.least_squares
. - Included additional 1D functions for, e.g., rotation curves.
- Bug fixes when using covariance and fitting without dispersion in
nirvana_manga_axisym
script. Includes propagation of sigma covariance to correctedsigma**2
covariance. - Added the
nirvana_manga_axisym_collate
script to collate the metadata rows from allnirvana_manga_axisym
output files into a single fits table. - Minor adjustments to the datamodel for the
nirvana_manga_axisym
metadata table.
- Initial version
- Add beam-smearing
- Restructure into python package
- Added
Kinematics
class for presenting the data to be modeled - Added the MaNGA-specific I/O objects that subclass from
Kinematics
- Changes the geometry computation; see
barfit.models.geometry
- Speeds up the beam-smearing convolution, and checks that the convolution with the MaNGA PSF does not shift the model
- Dramatically speeds up the
Kinematics.bin
function usingscipy.sparse
matrices. - Adds a bunch of test code.
- Begins the implementation of a general
AxisymmetricDisk
class. - Adds fixing of center
- Adds mock data set
- Added in fitting of dispersion in radial bins
- Changed
barfit.barfit.unpack
to use a dictionary - Lots of changes to
barfit.barfit
code to utilize dictionary - Various changes to
FitArgs
and script to accommodate dictionary - Huge changes to
barfit.plotting
for dispersion and dictionary - Switched to using
MPL-10
by default - Tests and docs for
barfit.models.oned
. - Added
barfit.models.beam.ConvolveFFTW
class for doing convolutions with the FFTW library. - Added masking of edges of mocks to mitigate convolution edge effects
- Renamed to NIRVANA
- Removed MCMC and associated arguments
- Enable fit to velocity dispersion using
AxisymmetricDisk
. - Force calculation of
mom0
innirvana.model.beam.smear
even if the surface brightness is not provided, using a unity array. - Add continuous integration tests using GitHub Workflow
- Add initial docs
- Copy over the BitMask class from the sdss-mangadap
- Update masking and add method that determines the target sample given the targeting bits
- Added construction of MaNGA covariance matrices and incorporated them
into
Kinematics
. - Added handling of covariance matrices in
AxisymmetricDisk
. - Added
nirvana.data.scatter.IntrinsicScatter
used to both reject outlying residuals and determine the intrinsic scatter in the data. - Added galaxy pngs, drpall, and dapall to test data
- Added
nirvana.data.meta.GlobalPar
andnirvana.data.manga.MaNGAGlobalPar
classes to hold metadata - Added iterative axisymmetric disk fitting, assessment plots, and output file.
- Fix
nirvana.data.manga.MaNGAGlobalPar
to handle when photometry is unavailable. New header keywordPHOT_KEY
add to binary table output for axisymmetric fit giving which photometry was used. - Axisymmetric fit changes:
- Include flag to set minimum number of valid spaxels for axisymmetric fit to proceed.
- set lower bound on inclination to 1 degree to avoid div-by-zero errors.
- Added option to find the largest coherent region of valid spaxels for the fit.
- Added manga download scripts