Add some support for X-Array in a staged manner.

[scollis]

Creating a Checklist so we can keep track of this:

• [x] Add an xradar branch to develop/test on
• [x] Add helper functions to convert to an xradar object (#1251 )
• [ ] Move aux_io readers to xradar
• [x] Implement a radar.to_xradar() method
• [x] Add a separate xradar object, similar to the radar object, that we can feed metadata + values directly into

[tjlang]

Did someone say cartopy support for grids? Y'all already got that. Here's a plot with multiple NEXRADs mosaicked via Py-ART, visualizing an ISS overpass of thunderstorms near San Antonio, Austin, and College Station (red stars) on 1/2/2017. Base map provided by cartopy and its StamenTerrain tiling. Was like two lines to get the radar on there, using Grid's method to get lat/lons, then a pcolormesh call.

[scollis]

Yep.. But it is one line if you have it in an xarray container :) Sweet plot.

[zssherman]

Xarray was added to grid routine, but not the radardisplay. Xarry was also used in the spectra object. I will leave this open for now .

[mgrover1]

Coming back to this thread - we should like this to Discussion #1047

[mgrover1]

Requirements for Integration

As mentioned at previous meetings/discussions, we are moving forward with an xradar accessor, enabling us to use the full suite of functionality of Py-ART with xradar.

Overall Goal

The idea here is add an accessor

xradar_tree.pyart.to_radar()

Which adds a collection of additional methods (ex. .get_gate_x_y_z) to the xradar object that mimic the existing Py-ART radar object.

Required Methods

The existing pyart.core.Radar() object has the following methods:

• [x] __getstate__(self)
  • Returns the state of the object, including the data
• [x] __setstate__(self, state)
  • Restore unpickleable entries from the pickled object
• [ ] init_rays_per_sweep(self)
  • Initialize or reset the rays_per_sweep attribute
  • [ ] init_gate_x_y_z(self)
  • Initialize or reset the gate_{x, y, z} attributes.
  • [ ] init_gate_longitude_latitude(self)
  • Initialize or reset the gate_longitude and gate_latitude attributes.
  • [ ] init_gate_altitude(self):
  • Initialize the gate_altitude attribute.
  • [ ] _check_sweep_in_range(self, sweep)
  • Check that a sweep number is in range.
  • [ ] check_field_exists(self, field_name)
  • Check that a field exists in the fields dictionary. field_name : str, Name of field to check.
  • [ ] iter_start(self)
  • """Return an iterator over the sweep start indices."""
  • [ ] iter_end(self)
  • Return an iterator over the sweep end indices
  • [ ] iter_start_end(self)
  • Return an iterator over the sweep start and end indices
  • [ ] iter_slice(self)
  • Return an iterator over the sweep slice objects
  • [ ] iter_field(self, field_name)
  • Return an iterator which returns sweep field data
  • [ ] iter_azimuth(self)
  • Return an iterator which returns sweep azimuth data
  • [ ] iter_elevation
  • Return an iterator which returns sweep elevation data
  • [ ] get_field(self, sweep, field_name, copy=False)
  • Return the field data for a given sweep. Returns the numpy data array.
  • [ ] get_azimuth(self, sweep, copy=False)
  • Return an array of azimuth angles for a given sweep. Returns a numpy array of azimuth angles.
  • [ ] ` get_elevation(self, sweep, copy=False)
  • Return an array of elevation angles for a given sweep. Returns a numpy array of elevation angles.
  • [ ] get_gate_x_y_z(self, sweep, edges=False, filter_transitions=False)
  • Return the x, y and z gate locations in meters for a given sweep. With the default parameter this method returns the same data as contained in the gate_x, gate_y and gate_z attributes but this method performs the gate location calculations only for the specified sweep and therefore is more efficient than accessing this data through these attribute.
  • [ ] get_gate_area(self, sweep)
  • Return the area of each gate in a sweep. Units of area will be the same as those of the range variable, squared.
  • [ ] get_gate_lat_lon_alt(self, sweep, reset_gate_coords=False, filter_transitions=False ):
  • Return the longitude, latitude and altitude gate locations. Longitude and latitude are in degrees and altitude in meters. With the default parameter this method returns the same data as contained in the gate_latitude, gate_longitude and gate_altitude attributes but this method performs the gate location calculations only for the specified sweep and therefore is more efficient than accessing this data through these attribute. If coordinates have at all, please use the reset_gate_coords parameter.
  • [ ] get_nyquist_vel(self, sweep, check_uniform=True)
  • Return the Nyquist velocity in meters per second for a given sweep.
  • [ ] info(self, level="standard", out=sys.stdout)
  • Print information on radar.
  • [ ] _dic_info(self, attr, level, out, dic=None, ident_level=0)
  • Print information on a dictionary attribute
  • [ ] add_field(self, field_name, dic, replace_existing=False)
  • Add a field to the object.
  • [ ] add_filter(self, gatefilter, replace_existing=False, include_fields=None)
  • Updates the radar object with an applied gatefilter provided by the user that masks values in fields within the radar object.
  • [ ] add_field_like( self, existing_field_name, field_name, data, replace_existing=False ):
  • Add a field to the object with metadata from a existing field.
  • [ ] extract_sweeps(self, sweeps)
  • Create a new radar contains only the data from select sweeps.
  • [ ] _gate_lon_lat_data_factory(radar, coordinate)
  • Return a function which returns the geographic locations of gates
  • [ ] _gate_altitude_data_factory(radar)
  • Return a function which returns the gate altitudes

[zssherman]

@mgrover1 How were we in terms of this list?