__all__ = ['wind']
__doc__ = """
.. _Read
:mod:`Read` -- wind Read interface
============================================
.. module:: Read
:platform: Unix, Windows
:synopsis: Provides :ref:`PseudoNetCDF` random access read for CAMx
wind files. See PseudoNetCDF.sci_var.PseudoNetCDFFile
for interface details
.. moduleauthor:: Barron Henderson <barronh@unc.edu>
"""
# Distribution packages
import unittest
import struct
# Site-Packages
from numpy import zeros, array
# This Package modules
from PseudoNetCDF.camxfiles.timetuple import timediff, timeadd, timerange
from PseudoNetCDF.camxfiles.FortranFileUtil import OpenRecordFile, read_into
from PseudoNetCDF.sci_var import PseudoNetCDFFile, PseudoNetCDFVariables
# for use in identifying uncaught nan
listnan = struct.unpack('>f', b'\xff\xc0\x00\x00')[0]
checkarray = zeros((1,), 'f')
checkarray[0] = listnan
array_nan = checkarray[0]
[docs]
class wind(PseudoNetCDFFile):
"""
wind provides a PseudoNetCDF interface for CAMx
wind files. Where possible, the inteface follows
IOAPI conventions (see www.baronams.com).
ex:
>>> wind_path = 'camx_wind.bin'
>>> rows,cols = 65,83
>>> windfile = wind(wind_path,rows,cols)
>>> windfile.variables.keys()
['TFLAG', 'U', 'V']
>>> v = windfile.variables['V']
>>> tflag = windfile.variables['TFLAG']
>>> tflag.dimensions
('TSTEP', 'VAR', 'DATE-TIME')
>>> tflag[0,0,:]
array([2005185, 0])
>>> tflag[-1,0,:]
array([2005185, 240000])
>>> v.dimensions
('TSTEP', 'LAY', 'ROW', 'COL')
>>> v.shape
(25, 28, 65, 83)
>>> windfile.dimensions
{'TSTEP': 25, 'LAY': 28, 'ROW': 65, 'COL': 83}
"""
time_hdr_fmts = {12: "fii", 8: "fi"}
data_fmt = "f"
def __init__(self, rf, rows=None, cols=None):
"""
Initialization included reading the header and learning
about the format.
see __readheader and __gettimestep() for more info
"""
self.rffile = OpenRecordFile(rf)
self.time_hdr_fmt = self.time_hdr_fmts[self.rffile.record_size]
self.time_hdr_size = struct.calcsize(self.time_hdr_fmt)
self.padded_time_hdr_size = struct.calcsize("ii" + self.time_hdr_fmt)
self.__readheader()
self.__gettimestep()
if rows is None and cols is None:
rows = self.cell_count
cols = 1
elif rows is None:
rows = self.cell_count / cols
elif cols is None:
cols = self.cell_count / rows
else:
if cols * rows != self.cell_count:
raise ValueError(("The product of cols (%d) and rows (%d) " +
"must equal cells (%d)") %
(cols, rows, self.cell_count))
self.createDimension('TSTEP', self.time_step_count)
self.createDimension('COL', cols)
self.createDimension('ROW', rows)
self.createDimension('LAY', self.nlayers)
self.variables = PseudoNetCDFVariables(self.__var_get, ['U', 'V'])
def __var_get(self, key):
def constr(uv):
return self.getArray()[
:, {'U': 0, 'V': 1}[uv], :, :, :].copy()
def decor(uv):
return dict(units='m/s',
var_desc=uv.ljust(16),
long_name=uv.ljust(16))
values = constr(key)
var = self.createVariable(key, 'f', ('TSTEP', 'LAY', 'ROW', 'COL'))
var[:] = values
for k, v in decor(key).items():
setattr(var, k, v)
return var
def __readheader(self):
"""
__readheader reads the header section of the ipr file
it initializes each header field (see CAMx Users Manual for a list)
as properties of the ipr class
"""
self.data_start_byte = 0
if self.time_hdr_fmt == 'fii':
self.start_time, self.start_date, self.lstagger = self.rffile.read(
self.time_hdr_fmt)
elif self.time_hdr_fmt == 'fi':
self.start_time, self.start_date = self.rffile.read(
self.time_hdr_fmt)
self.lstagger = None
else:
raise NotImplementedError("Header format is unknown")
self.record_size = self.rffile.record_size
self.padded_size = self.record_size + 8
self.cell_count = self.record_size // struct.calcsize(self.data_fmt)
self.record_fmt = self.data_fmt * self.cell_count
def __gettimestep(self):
"""
Header information provides start and end date, but does not
indicate the increment between. This routine reads the first
and second date/time and initializes variables indicating the
timestep length and the anticipated number.
"""
# This is a bit of a hack, but should work:
# Search for the next record that is the same
# length as self.padded_time_hdr_size
#
# This should be the next date record
# 1) date - startdate = timestep
# 2) (record_start - self.padded_time_hdr_size)/self.padded_size
# = klayers
self.rffile._newrecord(0)
self.rffile.next()
nlayers = 0
while not self.rffile.record_size == self.time_hdr_size:
self.rffile.next()
nlayers += 1
self.nlayers = (nlayers - 1) // 2
if self.time_hdr_fmt == "fi":
time, date = self.rffile.read(self.time_hdr_fmt)
elif self.time_hdr_fmt == "fii":
time, date, lstagger = self.rffile.read(self.time_hdr_fmt)
self.end_time, self.end_date = time, date
self.time_step = timediff(
(self.start_date, self.start_time), (date, time))
while True:
try:
for i in range(self.nlayers * 2 + 1):
self.rffile.next()
if self.rffile.record_size == 8:
self.end_time, self.end_date = self.rffile.read("fi")
elif self.rffile.record_size == 12:
self.end_time, self.end_date, lstagger = self.rffile.read(
"fii")
else:
raise KeyError()
except Exception:
break
self.time_step_count = int(timediff((self.start_date, self.start_time),
(self.end_date, self.end_time)) /
self.time_step) + 1
def __layerrecords(self, k):
return k - 1
def __timerecords(self, dt):
"""
routine returns the number of records to increment from the
data start byte to find the first time
"""
d, t = dt
nsteps = int(
timediff((self.start_date, self.start_time),
(d, t)) / self.time_step)
nlays = self.__layerrecords(self.nlayers + 1)
return nsteps * nlays
def __recordposition(self, date, time, k, duv):
"""
routine uses pagridrecords, timerecords,irecords,
jrecords, and krecords multiplied by the fortran padded size
to return the byte position of the specified record
pagrid - integer
date - integer
time - float
duv - integer (0=date,1=uwind,2=vwind)
"""
bytes = self.data_start_byte
nsteps = self.__timerecords((date, time))
bytes += int(nsteps / self.nlayers) * self.padded_time_hdr_size
bytes += int(nsteps / self.nlayers) * 12
bytes += nsteps * self.padded_size * 2
if not duv == 0:
bytes += self.padded_time_hdr_size
bytes += self.__layerrecords(k) * 2 * self.padded_size
if duv == 2:
bytes += self.padded_size
return bytes
[docs]
def seek(self, date=None, time=None, k=1, uv=1):
"""
Move file cursor to beginning of specified record
see __recordposition for a definition of variables
"""
if date is None:
date = self.start_date
if time is None:
time = self.start_time
chkvar = True
if (
chkvar and
timediff((self.end_date, self.end_time), (date, time)) > 0 or
timediff((self.start_date, self.start_time), (date, time)) < 0
):
raise KeyError(("Wind file includes (%i,%6.1f) thru (%i,%6.1f); " +
"you requested (%i,%6.1f)") %
(self.start_date, self.start_time,
self.end_date, self.end_time, date, time))
if chkvar and uv < 0 or uv > 2:
raise KeyError("Wind file includes Date (uv: 0), u velocity " +
"(uv: 1) and v velocity (uv: 2); you requested %i" %
(uv))
self.rffile._newrecord(self.__recordposition(date, time, k, uv))
[docs]
def read(self):
"""
provide direct access to the underlying RecordFile read
method
"""
return self.rffile.read(self.record_fmt)
[docs]
def read_into(self, dest):
"""
put values from rffile read into dest
dest - numpy or numeric array
"""
return read_into(self.rffile, dest, "", self.data_fmt)
[docs]
def seekandreadinto(self, dest, date=None, time=None, k=1, duv=1):
"""
see seek and read_into
"""
self.seek(date, time, k, duv)
self.read_into(dest)
[docs]
def seekandread(self, date=None, time=None, k=1, duv=1):
"""
see seek and read
"""
self.seek(date, time, k, duv)
return self.read()
[docs]
def keys(self):
for d, t in timerange((self.start_date, self.start_time),
timeadd((self.end_date, self.end_time),
(0, self.time_step)),
self.time_step):
for k in range(1, self.nlayers + 1):
yield d, t, k
[docs]
def values(self):
for d, t, k in self.keys():
yield self.seekandread(d, t, k, 1), self.seekandread(d, t, k, 2)
[docs]
def items(self):
for d, t, k in self.keys():
y1, y2 = self.seekandread(d, t, k, 1), self.seekandread(d, t, k, 2)
yield d, t, k, y1, y2
__iter__ = keys
[docs]
def getArray(self, krange=slice(1, None)):
if not isinstance(krange, slice):
if isinstance(krange, tuple):
krange = slice(*krange)
if isinstance(krange, int):
krange = slice(krange, krange + 1)
a = zeros(
(
self.time_step_count,
2,
len(range(*krange.indices(self.nlayers + 1))),
len(self.dimensions['ROW']),
len(self.dimensions['COL']),
), 'f')
nlay = self.nlayers
for i, (d, t) in enumerate(self.timerange()):
for uv in range(1, 3):
for ki, k in enumerate(range(*krange.indices(nlay + 1))):
uvi = uv - 1
ki = k - 1
self.seekandreadinto(a[i, uvi, ki, :, :], d, t, k, uv)
return a
[docs]
def timerange(self):
return timerange((self.start_date, self.start_time),
timeadd((self.end_date, self.end_time),
(0, self.time_step)),
self.time_step)
class TestRead(unittest.TestCase):
def runTest(self):
pass
def setUp(self):
pass
def testWD(self):
import PseudoNetCDF.testcase
wdfile = wind(PseudoNetCDF.testcase.camxfiles_paths['wind'], 4, 5)
checkv = array([
-1.73236704e+00, -1.99612117e+00, -3.00912833e+00, -3.92667437e+00,
-3.49521232e+00, 2.04542422e+00, 8.57666790e-01, -1.71201074e+00,
-4.24386787e+00, -5.37704515e+00, 1.85697508e+00, 6.34313405e-01,
-1.21529281e+00, -3.03180861e+00, -4.36278439e+00, -1.90753967e-01,
-1.08261776e+00, -1.73634803e+00, -2.10829663e+00, -2.28424144e+00,
-1.88443780e+00, -2.02582169e+00, -3.09955978e+00, -4.14587784e+00,
-3.72402787e+00, 2.16277528e+00, 8.94082963e-01, -1.86343944e+00,
-4.58147812e+00, -5.81837606e+00, 1.97949493e+00, 6.12511635e-01,
-1.35096896e+00, -3.25313163e+00, -4.67790413e+00, -1.89851984e-01,
-1.16381800e+00, -1.84269297e+00, -2.21348834e+00, -2.40952253e+00,
-2.04972148e+00, -2.11795568e+00, -3.06094027e+00, -4.11207581e+00,
-3.74964952e+00, 2.09780049e+00, 8.01259458e-01, -1.90404522e+00,
-4.59170580e+00, -5.83114100e+00, 1.97475648e+00, 5.54396451e-01,
-1.41695607e+00, -3.28227353e+00, -4.67724609e+00, -1.94723800e-01,
-1.18353117e+00, -1.86556363e+00, -2.22842574e+00, -2.42080784e+00,
-1.65720737e+00, -1.58054411e+00, -2.25336742e+00, -3.06462526e+00,
-2.47261453e+00, 1.37642264e+00, 1.16142654e+00, -6.82058990e-01,
-2.68112469e+00, -3.38680530e+00, 1.80796599e+00, 1.48641026e+00,
-1.71508826e-02, -1.68607295e+00, -2.89399385e+00, 3.40398103e-01,
3.25049832e-02, -5.91206312e-01, -1.19038010e+00, -1.52301860e+00,
-1.83006203e+00, -1.74505961e+00, -2.50190806e+00, -3.29507184e+00,
-2.65367699e+00, 1.55719578e+00, 1.25234461e+00, -9.14537191e-01,
-3.16307521e+00, -4.00584650e+00, 2.07018161e+00, 1.60957754e+00,
-1.46312386e-01, -2.04018188e+00, -3.40377665e+00, 4.11731720e-01,
-2.29119677e-02, -7.27373540e-01, -1.35116744e+00, -1.70711970e+00,
-1.72859466e+00, -1.73683071e+00, -2.65253377e+00, -3.43689489e+00,
-2.75470304e+00, 1.58366191e+00, 1.19656324e+00, -1.09935236e+00,
-3.38544369e+00, -4.26436615e+00, 2.04826832e+00, 1.53576791e+00,
-2.60809243e-01, -2.18679833e+00, -3.59082842e+00, 3.77060443e-01,
-1.05680525e-01, -8.10511589e-01, -1.40993130e+00,
-1.76300752e+00], dtype='f').reshape(2, 3, 4, 5)
self.assertTrue((wdfile.variables['V'][:] == checkv).all())
if __name__ == '__main__':
unittest.main()