-------------------
- Included code for DAPall file
- Fixed error in bin area calculation in unbinned data.
- Added convenience functions to DAPFits for use with the MAPS files.
- Fixed problem with velocity offset between template and galaxy data
due to difference in pixel sampling.
- Updated mangampl.py to include matplotlib version; added MPL-6
versions
- Fixed how the stellar-continuum-fitting masks are saved; still need
to make sure users can reconstruct these exactly from the output
model cube files.
- Changed how spectral step is calculated to avoid numerical precision
issues.
- Incorporated M. Cappellari's `gaussian_filter1d` into the
`convolution_variable_sigma` function in
python/mangadap/util/instrument.py. It's 2 orders of magnitude
faster now!
- Include spectral resolution handling hooks into spatial binning and
stellar continuum fitting. The code is compliant with data that both
does and does not have the 'DISP' extension in the DRP cube file for
backwards compatibility.
- "Database" parameter classes (e.g., EmissionLineDB) have been moved
from python/mangadap/par to python/mangadap/proc/.
- Incorporated spaxel-by-spaxel instrumental dispersion corrections.
- Significantly restructuring of most of the core classes to minimize
the size of the reference files and to use common functions where
possible (e.g., new DAPFitsUtil class).
- Added equivalent width measurements for the Gaussian fitted fluxes;
core function identical for the summed and Gaussian-fit fluxes.
- Moved construction of main output files to dapfits.py
- Include a map mask in reference files
- Allow rundap to read the plate, ifu, and mode combinations to analyze
from a file
- Include filtering approach in stellar kinematics fitting
- Incorporate ppxf v6.0.6
- Reading/writing of fits data is now done using DAPFitsUtil. When
reading with this utility, the fits image data are automatically
restructured into a row-major ordering to ease the confusion between
the way astropy.io.fits provides the data and the ordering of the
NAXIS keywords. At the moment, this restructuring does not mess up
the WCS coordinates in the header; however, depending on any
development of astropy.io.fits in the future, this may become a
problem. This feels like a big mess at the moment, but there's no
easy way around it.
- Restructured how the covariance data are read and written to adhere
to the format already adopted by the DRP. The Covariance class
should now be able to seemlessly read the broad-band correlation
matrices provided in the LOGCUBE files.
- Added DefaultConfig class to help manage config files. Edited the
StellarContinuumModelDef parameters to accommodate the new filtering
approach. Removed all but one StellarContinuumModel method.
- Include an estimate of the stellar velocity dispersion correction
based on fitting the native resolution optimized template to the same
spectrum that has been rsolution matched to object spectrum being
fit.
- Changed velocity dispersion correction calculation for stellar
kinematics: Old version used mean resolution difference between
galaxy and template spectra. New version uses a fit of the
best-fitting model template to a resolution-matched version of
itself.
- MAPS and LOGCUBE files set to single-precision float for survey
output.
- BINID in MAPS and LOGCUBE files set to int32 type for survey output.
- Added NeIII and higher-order Balmer series lines (up to H10).
- General improvements to the documentation
- New emission-line module Sasuke that uses pPXF to fit the lines.
- Adjusted pPXF filtering approach: remove final, fixed-kinematics fit
and set polynomial order of initial fit to be the effective order of
the filter (full length/smoothing box-1)
- Calculate and include nominal errors in template weights in output
reference file for stellar continuum model.
- HARDCODED: Use first moment of H-alpha as the initial velocity guess
for the emission lines.
- Include pre-pixelized Gaussian extension from DRP.
- Use 'z' column in plateTargets files to construct input parameter
files instead of the `nsa_z` column; this should catch the
ancillaries not in the NSA but with redshifts in the targeting
catalog.
- In rundap, wait for cluster processes to finish then also submit a
script to construct the DAPall file
- Updated versions of ppxf and `voronoi_2d_binning` and moved these
files out of the mangadap package and into a new captools directory
- Edited Sasuke to use iterations tested by Xihan Ji and Michele
Cappellari.