Sdss-mangadap

-------------------

- Put in default paths when environmental variables not defined and
warn user, both when running setup.py and when importing mangadap
- Start testing suite (still very limited)
- Fix error in constructing parameter tying vector passed to ppxf
- Allow change of templates used during stellar-continuum and
emission-ine fit.
- Fixed bug in post-fit chi-square calculation in PPXFFit
- Updated requirements
- Fixed bug in `passband_integral` error calculation, but MC tests show
that it's a poor substitute for an MC simulation.
- Significant changes to the resampling code; replaced `resample1d`
with `Resample` class.
- Added test data to `data/tests/`
- Added tests for more core routines, like PPXFFit.fit() and
Sasuke.fit(). Tests take about 1 min to execute.
- Fix FORESTAR bug in spectral indices.
- Fix some bugs in the bandpass integral functions introduced during
recent testing.
- Add spectral indices test
- Change model-cube data model to use `STCNT` and `STCNT_MASK`
extensions and remove the `EMLINE_BASE` and `EMLINE_MASK` extensions.
- Allow the emission-line fitter to use the bin ID numbers directly
instead of matching the spaxels to bins by coordinate proximity
- Construction of the parameter tying object in the emission-line
fitter is now done just before each spectrum is fit by ppxf (not
globally in Sasuke) to better handle when components are omitted.
- When deconstructing bins into spaxels, the second fit iteration
*only* fits spaxels that are components of binned spectra and does
not refit spectra that constitute an entire bin themselves. I.e.
this removes some largely redundant fitting.
- PPXFFit and Sasuke include calculations of the chi-square growth; and
changed the names of the growth columns in the reference files.
- Changed definitions of A to be the model amplitude; A/N is the model
amplitude divided by the median noise in the two sidebands defined
for the emission-line EW calculation.
- Consolidated figures-of-merit for the stellar-continuum fitting into
a single MAPS extension and added additional quantities.
- Added MAPS with the amplitude, A/N, and local reduced chi-square for
each line fit in the emission-line modeling.
- Added MAPS with the full-spectrum figures-of-merit for the
emission-line module, similar to the stellar-continuum output from
ppxf.
- Major changes to survey-level code: drpcomplete.py and rundap.py.
- Fixed minor issue in EmissionLineMoments.single_band_moments that
caused many "UserWarning: Warning: converting a masked element to
nan." messages.
- Added MODEL_MASK to model cube file to differentiate between model
mask and binned spectra mask
- Added fit QA plots for stellar and emission line modules for each
PLTIFU, and a QA plot combining all observations on a given plate.
- Added calls to additional QA plots in rundap.py
- Changed daptype to binning-stellartemplates-emissionlinetemplates
using default_dap_method. StellarContinuumModel reference file moved
to the common directory.
- Add databases and configuration files that include the HeI 3889 line
and the NI ~5200 doublet.
- Added post-processing qa plots
- Fixed minor issue in propagating masks from the reference files to
the maps files; primarily an issue for the hybrid binning scheme.
- Ignore minimum S/N limitation in emission-line moments and spectral
indices for hybrid scheme as a stop-gap to minimize differences in
moments, models, and indices BINID.

--------------------
- Code migrated from SVN to GitHub

---------------------------------------------
- Add DAPFRMT and DAPTYPE header keywords to MAPS and LOGCUBE files.

-----------------------------------------
- Fixed NaN problems in DAPall file
- Emission-lines moments are remeasured after the emission-line model
fitting to account for the adjustments made to the continuum during
that process.
- During the emission-line modeling, fit metrics are now computed in
15-pixel wide windows around each fitted line. The metrics are
amplitude, amplitude-to-noise, chi-square, root-mean-square
difference with the model, and fractional RMS wrt the model. These
data are currently only in the reference files, not the output maps.
- Begun coding for switching the templates between what is used for the
stellar kinematics fitting and what's used for the emission-line
modeling. Coding incomplete, but does not effect runs that use the
same template sets for both analysis steps (as done in DR15).
- MILES-HC templates changed to reflect re-run of the analysis by M.
Cappellari so that we have documented which MILES template is
incorporated into each MILES-HC template. The total number of
templates was again 49; however, 7 of these have been removed because
they had artifacts and/or emission-lines in their spectra. See
`./data/spectral_templates/miles_cluster/README` .
- Included redshift fix file that can be used to replace the redshift
used for any PLATEIFU either because it's not provided by the
relevant ancillary programs or the NSA redshift is incorrect.
- Corrected error in the emission-line fluxes for the fact that the
input template flux increases with the fitted redshift. So: `F_true
= (1+z)*F_MPL6`.
- Corrected the emission-line equivalent width measurements from
observed to rest frame.
- Corrected the absorption-line spectral-index equivalent widths
(angstrom units) from observed to rest frame.
- Emission-line moments and spectral-indices (anything depending on the
integration over a set of bandpasses) are now flagged as UNRELIABLE
in the output maps if *any* pixels are masked in *any* their
respective band passes. This is true for both the main passband and
any sidebands.
- Spectral index measurements are now provided for the HYB binning
scheme.
- Spectral-index velocity-dispersion corrections are done adopting the
NSA redshift for all spectra. This means that there will be a
*velocity* dependence of the correction that can be more significant
that the correction for the velocity dispersion itself. The approach
to the velocity dispersion corrections should be improved in future
releases.

-------------------
- Minor bug fix

-------------------
- Calculation of binned flux statistics now uses response function
provided to the ReductionAssessments class
- Binned flux statistics now determined before applying the reddening
correction; now matches individual spaxel measurements.
- Fixed some roundoff errors in the `SPX_*` extensions due to
conversion from double to single-precision floats in output MAPS
- Removed resolution-vector-based stellar velocity dispersion
corrections from output MAPS files. Only empirical (more robust)
correction now provided, but resolution-vector calculation still in
reference files.
- Propagated this change though to the ppxf and spotcheck qa plots
- Added `EMLINE_TPLSIGMA` to output MAPS files, providing the line
width of the templates used during Sasuke. Also added this into the
measurements returned by pPXF such that `EMLINE_GSIGMA` and
`EMLINE_INSTSIGMA` have the same meaning as in MPL-5: `EMLINE_GSIGMA`
is the effective observed sigma of the line and `EMLINE_INSTSIGMA` is
the instrumental dispersion at the line centroid.