MetPy 0.8 Released

MetPy 0.8.0 has been released. This release has a wide collection of new features as well as minor bug fixes, including several contributions from our community.

Highlights

• MetPy now requires XArray. Using XArray, MetPy provides some custom accessors under the .metpy attribute on Dataset and DataArray. This allows simplified access to the underlying grid projection as a CartoPy projection, as well as simplified access to data as an array with units attached. (See the XArray_Projections.py example) MetPy calculation functions will also now automatically convert an XArray DataArray with units metadata into a proper array with units. Future releases of MetPy will improve the integration with XArray.
• GINIFile now works directly as an XArray storage backend. See the updated GINI_Water_Vapor.py example for details. Support for Unidata NEXRAD composites in GINI format was also added.
• MetPy's colortables are now available from metpy.plots.colortables rather than needing to do metpy.plots.ctables.registry.
• Fixed a few corner cases with isentropic interpolation, including problems with columns of entirely missing data (frequently encountered with NARR data)
• lfc and el now work with arbitrary parcel profiles
• Gradient and related functions now properly propagate masked values
• Text and time formatting for add_timestamp() are now configurable
• New calculations
  • Absolute vorticity
  • Ageostrophic wind
  • Apparent temperature (combines wind chill and heat index)
  • Brunt-Väisälä frequency and period
  • Critical angle (severe weather parameter)
  • Dewpoint (from specific humidity)
  • Exner function
  • Inertial advective wind
  • Parse angle (converts strings like "ESE" to degree values)
  • Potential temperature to temperature
  • Potential vorticity
  • Q Vector
  • Saturation equivalent potential temperature
  • Static stability
  • Wet bulb temperature
• Other assorted bug fixes
• Reminder: MetPy plans to drop support for Python 2.7 in Fall 2019

API Changes

• lat_lon_grid_spacing has been deprecated in favor of lat_lon_grid_deltas. The latter gives properly signed values depending on the ordering of points in the lat/lon grid.
• The deprecated upper air data reading support has been removed. Use the upper air support in siphon instead.
• The function delete_masked_points has been renamed to _delete_masked_points, as it was not intended to be part of the supported API.
• GINIFile.to_dataset(), which gives a netCDF4-python-like interface to the information in the file, has been deprecated. Instead, now an instance of GINIFile can be passed directly to xarray.open_dataset() to provide a similar interface. to_dataset() (as well as the non-public cdm module) will be removed in 0.11.

Contributors

@ahuang11, @akrherz, @asselapathirana, @eliteuser26, @jthielen, @kishan3, @Meteodan, @mwilson14, @shofer16450, @tjwixtrom, @WEP11, @jrleeman, and @dopplershift all contributed commits to this release.

For full release notes see the GitHub Release Page.

MetPy packages are available for Conda on the conda-forge channel and for pip from the Python Package Index.

Let us know if you run into any problems, either at MetPy issue tracker, using support-python@unidata.ucar.edu, or on the python-users list. You can also ask questions using the "MetPy" tag on Stack Overflow.