What's-new Section

[nstarman]

We (@mhvk and @nstarman) met to discuss our paper section. These are the categories we came up with, and some PRs that might make it into those sections.

✅ New (Sub)packages / features:

• [x] Time Series sub-package
  • V3.1: New Box Least Squares Periodogram
• [x] Uncertainties and Distributions
• [x] ERFA module is now it’s own thing
  • Obliquely tie in that units remains in astropy
• ~Pytest plugin ecosystem~
• [x] A new SpectralCoord class for representing and transforming spectral quantities
• ~Velocity in Coordinates~
• [x] WCS API : A shared Python interface for World Coordinate Systems
  • Include APE reference

✅ Interoperability (w/in and w/out):

• [x] Quantity + Numpy

• ? Added True Equator Mean Equinox (TEME) frame for satellite two-line ephemeris data

• [x] New Unified I/O architecture

• [x] Fits

  • Support for writing Dask arrays to FITS files
  • Fits Time Standards
    • GSOC interfacing with larger community

• [x] Table Column mixins

  • [HDF5] (https://docs.astropy.org/en/older-docs-archive/v3.0/whatsnew/3.0.html#whatsnew-3-0-hdf5-mixins)
  • Support use of SkyCoord in table vstack, dstack, and insert_row
  • Added support for dask arrays in tables
  • Added support for registering array-like objects as mixin columns

• [x] Table I/O

  • Support for reading and writing tables to QDP format
  • Append table to existing FITS file
  • MRT Support for reading and writing tables to MRT format
  • GSOC interfacing with larger community
  • Support for reading and writing tables to Parquet format
  • Converting SkyCoord to QTable
  • https://docs.astropy.org/en/v3.2/whatsnew/3.2.html#support-for-asdf-readers-writers-for-table-class

• [x] Cosmology:

  • Support for reading, writing, and converting Cosmology
  • Redshift units & equivalency : Cosmology units module

• [x] Modeling

  • Support for ASDF serialization of models
  • ? Units in models : Support for units on otherwise unitless models via the Model.coerce_units method.

• [x] ECSV

  • Support for multidimensional and object columns in ECSV

✅ Precision, Accuracy, & Reproducibility:

• [x] Constants and Units:
  • Improved Consistency of Physical Constants and Units,
  • SI/CODATA 2018 constants
• [x] Time: Leap-seconds
• [x] Coordinates:
  • Transformations to AltAz are now much more precise (and faster)
  • Ecliptic Transformations
• [x] Cosmology:
  • Planck15, later superseded.
  • Pre-Publication Planck 2018 Cosmological Parameters
  • Planck18 Accepted (replace above)
• [x] REPRODUCIBILITY NOTE: Can use old cosmologies and constants.
  • For true reproducibility, pin your whole environment.
  • WMAP1 & 3 Cosmologies

✅ Performance Enhancements:

• Faster cosmology calculations
• Quantity
• Coordinate matching
• Table , also Improvements to Performance for Tables,
• Time performance improvements,
• Transformations to AltAz are now much more precise (and faster)
• Improvements in making Astropy thread-safe
  • For Supercomputing?
• Statistics: Performance improvements to sigma clipping

✅ Future Highlights / What to watch for : (Exists and being developed)

• [x] General masked class for Quantity and other ndarray subclasses
  • Support for masked quantity columns, including masked FITS columns with units
• ? Uncertainties
• Distributions

[adrn]

This looks excellent and comprehensive! If this feels like it is too long, I think it would be fine to remove 1–3 points from interoperability, precision, performance.

[mhvk]

@nstarman - unfortunately, I cannot edit the list above. I think I dealt with new features (including planned ones), except that I did not discuss proper motions and radial velocities, since those are from astropy 2.0, and hence in Paper II. I also wrote a performance section - which I kept very short, since I don't really think any of the details matter to the typical reader.

[mhvk]

p.s. I also pushed a few more commits which mostly have a few comments on your sections. Not sure this is the best way to do these, but obviously feel free to remove the commits as you deal with them!