-
Geoff Cureton authoredLatest changes to griddingAndGranulation.c brought in from the CSPP VIIRS EDR project. From the same project we add LandWaterMask.py and PrecipWater.py as examples of the use of grid2gran_nearest() and grid2gran_bilinearInterp() respectively.
Geoff Cureton authoredLatest changes to griddingAndGranulation.c brought in from the CSPP VIIRS EDR project. From the same project we add LandWaterMask.py and PrecipWater.py as examples of the use of grid2gran_nearest() and grid2gran_bilinearInterp() respectively.
PrecipWater.py 16.76 KiB
#!/usr/bin/env python
# encoding: utf-8
"""
PrecipWater.py
* DESCRIPTION: Class to granulate the ViirsAncPrecipWater data product
Created by Geoff Cureton on 2013-02-25.
Copyright (c) 2011 University of Wisconsin SSEC. All rights reserved.
"""
file_Date = '$Date: 2014-09-19 15:13:47 -0500 (Fri, 19 Sep 2014) $'
file_Revision = '$Revision: 2235 $'
file_Author = '$Author: geoffc $'
file_HeadURL = '$HeadURL: https://svn.ssec.wisc.edu/repos/jpss_adl/trunk/scripts/edr/ANC/PrecipWater.py $'
file_Id = '$Id: PrecipWater.py 2235 2014-09-19 20:13:47Z geoffc $'
__author__ = 'G.P. Cureton <geoff.cureton@ssec.wisc.edu>'
__version__ = '$Id: PrecipWater.py 2235 2014-09-19 20:13:47Z geoffc $'
__docformat__ = 'Epytext'
import os, sys, logging, traceback
from os import path,uname,environ
import string
import re
import uuid
from glob import glob
from time import time
from datetime import datetime,timedelta
from scipy import round_
import numpy as np
from numpy import ma
import copy
from bisect import bisect_left,bisect_right
import ctypes
from numpy.ctypeslib import ndpointer
import ViirsData
from NCEPtoBlob import NCEPclass
# skim and convert routines for reading .asc metadata fields of interest
import adl_blob2 as adl_blob
import adl_asc
from adl_asc import skim_dir, contiguous_granule_groups, granule_groups_contain, \
effective_anc_contains,eliminate_duplicates,_is_contiguous, \
RDR_REQUIRED_KEYS, POLARWANDER_REQUIRED_KEYS
from adl_common import ADL_HOME, CSPP_RT_HOME, CSPP_RT_ANC_PATH, \
CSPP_RT_ANC_CACHE_DIR, COMMON_LOG_CHECK_TABLE
# every module should have a LOG object
try :
sourcename= file_Id.split(" ")
LOG = logging.getLogger(sourcename[1])
except :
LOG = logging.getLogger('PrecipWater')
from Utils import getURID, getAscLine, getAscStructs, findDatelineCrossings, \
shipOutToFile
class PrecipWater() :
def __init__(self,inDir=None, sdrEndian=None, ancEndian=None):
self.collectionShortName = 'VIIRS-ANC-Preci-Wtr-Mod-Gran'
self.xmlName = 'VIIRS_ANC_PRECI_WTR_MOD_GRAN.xml'
self.blobDatasetName = 'totalPrecipitableWater'
self.dataType = 'float32'
self.sourceType = 'NCEP_ANC_Int'
self.sourceList = ['']
self.trimObj = ViirsData.ViirsTrimTable()
if inDir is None :
self.inDir = path.abspath(path.curdir)
else :
self.inDir = inDir
if sdrEndian is None :
self.sdrEndian = adl_blob.LITTLE_ENDIAN
else :
self.sdrEndian = sdrEndian
if ancEndian is None :
self.ancEndian = adl_blob.LITTLE_ENDIAN
else :
self.ancEndian = ancEndian
def ingest(self,ancBlob=None):
'''
Ingest the ancillary dataset.
'''
dates = []
ncepBlobFiles = []
for gridBlobStruct in ancBlob:
timeObj = gridBlobStruct[0]
ncepBlobFile = gridBlobStruct[1]
LOG.debug("VIIRS-ANC-Preci-Wtr-Mod-Gran %s --> %s" % \
(ncepBlobFile,timeObj.strftime("%Y-%m-%d %H:%M:%S:%f")))
dates.append(timeObj)
ncepBlobFiles.append(ncepBlobFile)
self.date_0 = dates[0]
self.date_1 = dates[1]
LOG.debug("Minimum NCEP date is: %s" %(self.date_0.strftime("%Y-%m-%d %H:%M:%S:%f")))
LOG.debug("Maximum NCEP date is: %s" %(self.date_1.strftime("%Y-%m-%d %H:%M:%S:%f")))
ncepBlobFile_0 = ncepBlobFiles[0]
ncepBlobFile_1 = ncepBlobFiles[1]
self.gridData_0 = getattr(ncepBlobFile_0,self.blobDatasetName).astype(self.dataType)
self.gridData_1 = getattr(ncepBlobFile_1,self.blobDatasetName).astype(self.dataType)
def setGeolocationInfo(self,dicts):
'''
Populate this class instance with the geolocation data for a single granule
'''
# Set some environment variables and paths
ANC_SCRIPTS_PATH = path.join(CSPP_RT_HOME,'viirs')
ADL_ASC_TEMPLATES = path.join(ANC_SCRIPTS_PATH,'asc_templates')
# Collect some data from the geolocation dictionary
self.geoDict = dicts
URID = dicts['URID']
geo_Collection_ShortName = dicts['N_Collection_Short_Name']
N_Granule_ID = dicts['N_Granule_ID']
ObservedStartTimeObj = dicts['ObservedStartTime']
geoFiles = glob('%s/%s*' % (self.inDir,URID))
geoFiles.sort()
LOG.debug("###########################")
LOG.debug(" Geolocation Information ")
LOG.debug("###########################")
LOG.debug("N_Granule_ID : %r" % (N_Granule_ID))
LOG.debug("ObservedStartTime : %s" % (ObservedStartTimeObj.__str__()))
LOG.debug("N_Collection_Short_Name : %s" %(geo_Collection_ShortName))
LOG.debug("URID : %r" % (URID))
LOG.debug("geoFiles : %r" % (geoFiles))
LOG.debug("###########################")
timeDelta = (self.date_1 - self.date_0).total_seconds()
LOG.debug("timeDelta is %r seconds" %(timeDelta))
timePrime = (ObservedStartTimeObj - self.date_0).total_seconds()
LOG.debug("timePrime is %r seconds (%f percent along time interval)" % \
(timePrime,(timePrime/timeDelta)*100.))
delta_gridData = self.gridData_1 - self.gridData_0
self.gridData = (delta_gridData/timeDelta) * timePrime + self.gridData_0
gridData_0_avg = np.average(self.gridData_0)
gridData_1_avg = np.average(self.gridData_1)
gridData_avg = np.average(self.gridData)
LOG.debug("average(gridData_0) = %f" %(np.average(self.gridData_0)))
LOG.debug("average(gridData_1) = %f" %(np.average(self.gridData_1)))
LOG.debug("average(gridData) = %f" %(np.average(self.gridData)))
# Do we have terrain corrected geolocation?
terrainCorrectedGeo = True if 'GEO-TC' in geo_Collection_ShortName else False
# Do we have long or short style geolocation field names?
if (geo_Collection_ShortName=='VIIRS-MOD-GEO-TC' or geo_Collection_ShortName=='VIIRS-MOD-RGEO') :
longFormGeoNames = True
LOG.debug("We have long form geolocation names")
elif (geo_Collection_ShortName=='VIIRS-MOD-GEO' or geo_Collection_ShortName=='VIIRS-MOD-RGEO-TC') :
LOG.debug("We have short form geolocation names")
longFormGeoNames = False
else :
LOG.error("Invalid geolocation shortname: %s" %(geo_Collection_ShortName))
return -1
# Get the geolocation xml file
geoXmlFile = "%s.xml" % (string.replace(geo_Collection_ShortName,'-','_'))
geoXmlFile = path.join(ADL_HOME,'xml/VIIRS',geoXmlFile)
if path.exists(geoXmlFile):
LOG.debug("We are using for %s: %s,%s" %(geo_Collection_ShortName,geoXmlFile,geoFiles[0]))
# Open the geolocation blob and get the latitude and longitude
endian = self.sdrEndian
geoBlobObj = adl_blob.map(geoXmlFile,geoFiles[0], endian=endian)
# Get scan_mode to find any bad scans
scanMode = geoBlobObj.scan_mode[:]
badScanIdx = np.where(scanMode==254)[0]
LOG.debug("Bad Scans: %r" % (badScanIdx))
# Detemine the min, max and range of the latitude and longitude,
# taking care to exclude any fill values.
if longFormGeoNames :
if endian==adl_blob.BIG_ENDIAN:
latitude = getattr(geoBlobObj,'latitude').byteswap()
longitude = getattr(geoBlobObj,'longitude').byteswap()
latitude = latitude.astype('float')
longitude = longitude.astype('float')
else:
latitude = getattr(geoBlobObj,'latitude').astype('float')
longitude = getattr(geoBlobObj,'longitude').astype('float')
else :
latitude = getattr(geoBlobObj,'lat').astype('float')
longitude = getattr(geoBlobObj,'lon').astype('float')
latitude = ma.masked_less(latitude,-800.)
latMin,latMax = np.min(latitude),np.max(latitude)
latRange = latMax-latMin
longitude = ma.masked_less(longitude,-800.)
lonMin,lonMax = np.min(longitude),np.max(longitude)
lonRange = lonMax-lonMin
LOG.debug("min,max,range of latitide: %f %f %f" % (latMin,latMax,latRange))
LOG.debug("min,max,range of longitude: %f %f %f" % (lonMin,lonMax,lonRange))
# Determine the latitude and longitude fill masks, so we can restore the
# fill values after we have scaled...
latMask = latitude.mask
lonMask = longitude.mask
# Check if the geolocation is in radians, convert to degrees
if 'RGEO' in geo_Collection_ShortName :
LOG.debug("Geolocation is in radians, convert to degrees...")
latitude = np.degrees(latitude)
longitude = np.degrees(longitude)
latMin,latMax = np.min(latitude),np.max(latitude)
latRange = latMax-latMin
lonMin,lonMax = np.min(longitude),np.max(longitude)
lonRange = lonMax-lonMin
LOG.debug("New min,max,range of latitude: %f %f %f" % (latMin,latMax,latRange))
LOG.debug("New min,max,range of longitude: %f %f %f" % (lonMin,lonMax,lonRange))
# Restore fill values to masked pixels in geolocation
geoFillValue = self.trimObj.sdrTypeFill['VDNE_FLOAT64_FILL'][latitude.dtype.name]
latitude = ma.array(latitude,mask=latMask,fill_value=geoFillValue)
self.latitude = latitude.filled()
geoFillValue = self.trimObj.sdrTypeFill['VDNE_FLOAT64_FILL'][longitude.dtype.name]
longitude = ma.array(longitude,mask=lonMask,fill_value=geoFillValue)
self.longitude = longitude.filled()
# Record the corners, taking care to exclude any bad scans...
nDetectors = 16
firstGoodScan = np.where(scanMode<=2)[0][0]
lastGoodScan = np.where(scanMode<=2)[0][-1]
firstGoodRow = firstGoodScan * nDetectors
lastGoodRow = lastGoodScan * nDetectors + nDetectors - 1
latCrnList = [latitude[firstGoodRow,0],latitude[firstGoodRow,-1],latitude[lastGoodRow,0],latitude[lastGoodRow,-1]]
lonCrnList = [longitude[firstGoodRow,0],longitude[firstGoodRow,-1],longitude[lastGoodRow,0],longitude[lastGoodRow,-1]]
# Check for dateline/pole crossings
num180Crossings = findDatelineCrossings(latCrnList,lonCrnList)
LOG.debug("We have %d dateline crossings."%(num180Crossings))
# Copy the geolocation information to the class object
self.latMin = latMin
self.latMax = latMax
self.latRange = latRange
self.lonMin = lonMin
self.lonMax = lonMax
self.lonRange = lonRange
self.scanMode = scanMode
self.latitude = latitude
self.longitude = longitude
self.latCrnList = latCrnList
self.lonCrnList = lonCrnList
self.num180Crossings = num180Crossings
# Parse the geolocation asc file to get struct information which will be
# written to the ancillary asc files
geoAscFileName = path.join(self.inDir,URID+".asc")
LOG.debug("\nOpening %s..." % (geoAscFileName))
geoAscFile = open(geoAscFileName,'rt')
self.RangeDateTimeStr = getAscLine(geoAscFile,"ObservedDateTime")
self.RangeDateTimeStr = string.replace(self.RangeDateTimeStr,"ObservedDateTime","RangeDateTime")
self.GRingLatitudeStr = getAscStructs(geoAscFile,"GRingLatitude",12)
self.GRingLongitudeStr = getAscStructs(geoAscFile,"GRingLongitude",12)
geoAscFile.close()
def _grid2Gran_bilinearInterp(self,dataLat, dataLon, gridData, gridLat, gridLon):
'''Granulates a gridded dataset using an input geolocation'''
nData = np.int64(dataLat.size)
gridRows = np.int32(gridLat.shape[0])
gridCols = np.int32(gridLat.shape[1])
data = np.ones(np.shape(dataLat),dtype=np.float64)* -999.9
dataIdx = np.ones(np.shape(dataLat),dtype=np.int64) * -99999
ANC_SCRIPTS_PATH = path.join(CSPP_RT_HOME,'viirs')
libFile = path.join(ANC_SCRIPTS_PATH,'libgriddingAndGranulation.so.1.0.1')
LOG.debug("Gridding and granulation library file: %s" % (libFile))
lib = ctypes.cdll.LoadLibrary(libFile)
grid2gran_bilinearInterp = lib.grid2gran_bilinearInterp
grid2gran_bilinearInterp.restype = None
grid2gran_bilinearInterp.argtypes = [
ndpointer(ctypes.c_double,ndim=1,shape=(nData),flags='C_CONTIGUOUS'),
ndpointer(ctypes.c_double,ndim=1,shape=(nData),flags='C_CONTIGUOUS'),
ndpointer(ctypes.c_double,ndim=1,shape=(nData),flags='C_CONTIGUOUS'),
ctypes.c_int64,
ndpointer(ctypes.c_double,ndim=2,shape=(gridRows,gridCols),flags='C_CONTIGUOUS'),
ndpointer(ctypes.c_double,ndim=2,shape=(gridRows,gridCols),flags='C_CONTIGUOUS'),
ndpointer(ctypes.c_double,ndim=2,shape=(gridRows,gridCols),flags='C_CONTIGUOUS'),
ndpointer(ctypes.c_int64,ndim=1,shape=(nData),flags='C_CONTIGUOUS'),
ctypes.c_int32,
ctypes.c_int32
]
'''
int snapGrid_ctypes(double *lat,
double *lon,
double *data,
long nData,
double *gridLat,
double *gridLon,
double *gridData,
long *gridDataIdx,
int nGridRows,
int nGridCols
)
'''
LOG.debug("Calling C routine grid2gran_bilinearInterp()...")
retVal = grid2gran_bilinearInterp(dataLat,
dataLon,
data,
nData,
gridLat,
gridLon,
gridData,
dataIdx,
gridRows,
gridCols)
LOG.debug("Returning from C routine grid2gran_bilinearInterp()")
return data,dataIdx
def granulate(self,ANC_objects):
'''
Granulate the ancillary dataset.
'''
LOG.info("Granulating %s ..." % (self.collectionShortName))
degInc = 0.5
lats = np.arange(361.)*degInc - 90.
lons = np.arange(720.)*degInc - 180.
latitude = self.latitude
longitude = self.longitude
gridData = self.gridData[:,:]
if self.num180Crossings != 2 :
#gridData = np.roll(gridData,360) # old
gridData = np.roll(gridData,360,axis=1) # new
gridLon,gridLat = np.meshgrid(lons,lats)
LOG.debug("start,end NCEP Grid Latitude values : %f,%f"%(gridLat[0,0],gridLat[-1,0]))
LOG.debug("start,end NCEP Grid Longitude values : %f,%f"%(gridLon[0,0],gridLon[0,-1]))
else :
negLonIdx = np.where(lons<0)
lons[negLonIdx] += 360.
lons = np.roll(lons,360)
gridLon,gridLat = np.meshgrid(lons,lats)
longitudeNegIdx = np.where(longitude < 0.)
longitude[longitudeNegIdx] += 360.
LOG.debug("start,end NCEP Grid Latitude values : %f,%f"%(gridLat[0,0],gridLat[-1,0]))
LOG.debug("start,end NCEP Grid Longitude values : %f,%f"%(gridLon[0,0],gridLon[0,-1]))
LOG.debug("min of gridData = %r"%(np.min(gridData)))
LOG.debug("max of gridData = %r"%(np.max(gridData)))
t1 = time()
data,dataIdx = self._grid2Gran_bilinearInterp(np.ravel(latitude),
np.ravel(longitude),
gridData.astype(np.float64),
gridLat,
gridLon)
t2 = time()
elapsedTime = t2-t1
LOG.info("Granulation took %f seconds for %d points" % (elapsedTime,latitude.size))
data = data.reshape(latitude.shape)
dataIdx = dataIdx.reshape(latitude.shape)
LOG.debug("Shape of granulated %s data is %s" % (self.collectionShortName,np.shape(data)))
LOG.debug("Shape of granulated %s dataIdx is %s" % (self.collectionShortName,np.shape(dataIdx)))
# Moderate resolution trim table arrays. These are
# bool arrays, and the trim pixels are set to True.
modTrimMask = self.trimObj.createModTrimArray(nscans=48,trimType=bool)
# Fill the required pixel trim rows in the granulated NCEP data with
# the ONBOARD_PT_FILL value for the correct data type
fillValue = self.trimObj.sdrTypeFill['ONBOARD_PT_FILL'][self.dataType]
data = ma.array(data,mask=modTrimMask,fill_value=fillValue)
self.data = data.filled()
def shipOutToFile(self):
''' Pass the current class instance to this Utils method to generate
a blob/asc file pair from the input ancillary data object.'''
return shipOutToFile(self)