Pypeit

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

- Introduces quick look scripts for MOS and NIRES
- Bumps dependencies including Python 3.7
- Modest refactoring of reduce/extraction/skysub codes
- Refactor of ScienceImage Par into pieces
- Finally dealt with 'random' windowing of Shane_kast_red
- Dynamic namp setting for LRISr when instantiating Spectrograph

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

- Hotfixes: np.histogram error in core/coadd1d.py, np.linspace using
float number of steps in core/wave.py, and sets numpy version to 1.16

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

- Implemented MOSFIRE and further implemented NIRSPEC for Y-band
spectroscopy.
- Fixed bug in coadd2d.
- Add VLT/FORS filters to our database
- Improved DEIMOS frame typing
- Brings Gemini/GMOS into the suite (R400)
- Also an important change for autoid.full_template()
- Fixed trace extrapolation, to fix bugs in object finding. Tweaks to
object finding algorithm.
- Major improvements to echelle object finding.
- Improved outlier rejection and coefficient fitting in pca_trace
- Major improvements to coadd routines in coadd1d
- Introduced telluric module and telluric correction routines
- Implemented tilt image type which is now a required frame type
- Streamlined and abstracted echelle properties and echelle routine in
spectrograph classes.
- Revamped 2-d coadding routines and introduced 2-d coadding of
MultiSlit data
- Improved ginga plotting routines.
- Fixed bug associated with astropy.stats.sigma_clipped_stats when
astropy.stats.mad_std is used.
- Refactor BPM generation
- Merge raw_image loading with datasec_img and oscansec_img generation
- Sync datasec_img to image in ProcessRawImage
- Started (barely) on a path to having calibration images in counts and
not ADU
- Refactors GMOS for get_rawimage method
- Enables GMOS overscan subtraction
- Adds R400 wavelength solution for old E2V chip
- Revises simple_calib() method for quick and dirty wavelength
calibration
- Adds a related show_wvcalib script
- Changes to ech_combspec to better treat filenames
- Fixed bug when bias was set to 'force' which was not bias subtracting
- Implemented changes to vlt_xshooter_nir to now require darks taken
between flats
- Made flat fielding code a bit more robust against hot pixels at edge
of orders
- Added pypeit_chk_flat script to view flat images
- Refactored image objects into RawImage, ProcessRawImage, PypeItImage,
BuildImage
- Moved load() and save() methods from MasterFrame to the individual
calibration objects
- Converted ArcImage and FlatImages into counts
- Added code to allow for IVAR and RN2 image generation for calibs
- Added several from_master_file() instantiation methods
- Use coadd2d.weighted_combine() to stack calibration images
- Major refactor of slit edge tracing
- Added 'Identify' tool to allow manual identification and calibration
of an arc spectrum
- Added support for WHT/ISIS
- Added 'Object Tracing' tool to allow interactive object tracing
- Added code of conduct
- Deprecated previous tracing code: `pypeit.traceslits` and
`pypeit.core.trace_slits`, as well as some functions in
`pypeit.core.extract` that were replaced by
`pypeit.core.moment.moment1d` and functions in `pypeit.core.trace`.
- PCA now saved to MasterEdges file; added I/O methods
- Improved CuAr linelists and archives for Gemini wavelength solutions
- New data model for specobj and specobsj objects (spec1d)
- Started some improvements to Coadd2D, TBC
- Allow for the continuum of the arc image to be modeled and subtracted
when tracing the line-centroid tilts
- Include a mask in the line detection in extracted central arc spectrum
of each slit/order. For VLT XShooter NIR, this was needed to ensure
the sigma calculation didn't include the off-order spectral positions.
- Added a staticmethed to :class:`pypeit.edgetrace.EdgeTraceSet` that
constructs a ``tslits_dict`` object directly from the Master file.

---------

- Add DOI

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

- Add magellan_mage, including a new ThAr linelist and an archived
solution
- Polish several key echelle methods
- Modify create_linelist to default to vacuum
- Update Xshooter, NIRES, and GNIRS
- Refactor ProcessImages into ProcessRawImage, PypeItImage,
CalibrationImage, ScienceImage, and ImageMask
- Refactor ScienceImage into SciImgStack
- Fix arc tilts bug
- Started an X-Shooter doc and introduced a [process][bias] parameter
- Modified processing steps for bias + overscan subtraction
- Started notes on how to generate a new spectrograph in PypeIt
- Refactoring of reduce to take a ScienceImage object for the images and
the mask
- Updates to many spectrograph files to put datasec, oscansec in the raw
frame
- Add find_trim_edge and std_prof_nsigma parameters
- A bit of tuning for MagE
- Fixes for Echelle in fluxspec
- Writes a chosen set of header cards to the spec1D and coadd files
- Updates for FORS2
- Introduced new coadd1d module and some new coadd functinality.
- modified interface to robust_polyfit_djs, robust_optimize, and
djs_reject.
- Added utility routine cap_ivar for capping the noise level.
- Fixed a bug in optimal extraction which was causing hot pixels when a
large fraction of the pixels on the object profile were masked.
- Major bug fixes and improvements to echelle object finding. Orders
which did not cover the entire detector were not being treated
properly.

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

- Minor bug fix to allow for `None` exposure times when typing frames.