A set of Python 3 tools for interfacing with the GRASP retrieval code
The core of the code is in runGRASP.py which contains several classes:
graspDB – This takes many pixel or graspRun objects and runs them all though a series of individual calls to GRASP. The class achieves two ends: (1) distribute large batches of pixels into chunks that are manageable from a memory perspective and (2) allow the individual calls to the GRASP binary to occur asynchronously in accordance with user specified parameters (e.g., max number of CPUs to use at a given time).
graspRun – This represents a collection of pixels, generally designed to be called in a single run. This class handles input/output to GRASP through the SDATA and output text files, respectively. The output data is stored in a list of rslt dicts, which are defined below.
pixel – This is class holds most of the information that would be in a typical row (entry for a single pixel) of an SDATA file.
graspYAML – This class has a variety of methods for changing settings in the YAML files.
for key,val in simBase.rsltFwd[0].items():
print("%23s: %27s %11s" % (("rslt['%s']" % key), str(type(val)), (str(val.shape) if 'numpy' in str(type(val)) else '()')))
rslt['datetime']: <class 'datetime.datetime'> ()
rslt['longitude']: <class 'numpy.float64'> ()
rslt['latitude']: <class 'numpy.float64'> ()
rslt['r']: <class 'numpy.ndarray'> (N_modes, N_radii)
rslt['dVdlnr']: <class 'numpy.ndarray'> (N_modes, N_radii)
rslt['rv']: <class 'numpy.ndarray'> (N_modes,)
rslt['sigma']: <class 'numpy.ndarray'> (N_modes,)
rslt['vol']: <class 'numpy.ndarray'> (N_modes,)
rslt['sph']: <class 'numpy.ndarray'> (N_modes,)
rslt['height']: <class 'numpy.ndarray'> (N_modes,)
rslt['heightStd']: <class 'numpy.ndarray'> (N_modes,)
rslt['lambda']: <class 'numpy.ndarray'> (N_lambda,)
rslt['aod']: <class 'numpy.ndarray'> (N_lambda,)
rslt['aodMode']: <class 'numpy.ndarray'> (N_modes, N_lambda)
rslt['ssa']: <class 'numpy.ndarray'> (N_lambda,)
rslt['ssaMode']: <class 'numpy.ndarray'> (N_modes, N_lambda)
rslt['n']: <class 'numpy.ndarray'> (N_modes, N_lambda)
rslt['k']: <class 'numpy.ndarray'> (N_modes, N_lambda)
rslt['rEff']: <class 'numpy.float64'> ()
rslt['rEffCalc']: <class 'numpy.float64'> ()
rslt['albedo']: <class 'numpy.ndarray'> (N_lambda,)
rslt['wtrSurf']: <class 'numpy.ndarray'> (3, N_lambda)
rslt['brdf']: <class 'numpy.ndarray'> (3, N_lambda)
rslt['bpdf']: <class 'numpy.ndarray'> (N_lambda,)
rslt['costVal']: <class 'float'> ()
rslt['sza']: <class 'numpy.ndarray'> (N_views, N_lambda)
rslt['vis']: <class 'numpy.ndarray'> (N_views, N_lambda)
rslt['fis']: <class 'numpy.ndarray'> (N_views, N_lambda)
rslt['sca_ang']: <class 'numpy.ndarray'> (N_views, N_lambda)
rslt['meas_I']: <class 'numpy.ndarray'> (N_views, N_lambda)
rslt['fit_I']: <class 'numpy.ndarray'> (N_views, N_lambda)
rslt['meas_Q']: <class 'numpy.ndarray'> (N_views, N_lambda)
rslt['fit_Q']: <class 'numpy.ndarray'> (N_views, N_lambda)
rslt['meas_U']: <class 'numpy.ndarray'> (N_views, N_lambda)
rslt['fit_U']: <class 'numpy.ndarray'> (N_views, N_lambda)
rslt['angle']: <class 'numpy.ndarray'> (N_ang, N_modes, N_lambda)
rslt['p11']: <class 'numpy.ndarray'> (N_ang, N_modes, N_lambda)
rslt['p12']: <class 'numpy.ndarray'> (N_ang, N_modes, N_lambda)
rslt['p22']: <class 'numpy.ndarray'> (N_ang, N_modes, N_lambda)
rslt['p33']: <class 'numpy.ndarray'> (N_ang, N_modes, N_lambda)
rslt['p34']: <class 'numpy.ndarray'> (N_ang, N_modes, N_lambda)
rslt['p44']: <class 'numpy.ndarray'> (N_ang, N_modes, N_lambda)
rslt['g']: <class 'numpy.ndarray'> (N_lambda,)
rslt['LidarRatio']: <class 'numpy.ndarray'> (N_lambda,)
rslt['LidarDepol']: <class 'numpy.ndarray'> (N_lambda,)
rslt['gMode']: <class 'numpy.ndarray'> (N_modes, N_lambda)
rslt['LidarRatioMode']: <class 'numpy.ndarray'> (N_modes, N_lambda)
rslt['LidarDepolMode']: <class 'numpy.ndarray'> (N_modes, N_lambda)