def loadDBstation(dbfile,stationID,varname,timeinfo=None,filttype=None,cutoff=3600.0,output_meta=False):
"""
Load station data from a database file
Inputs:
dbfile - location of database file
stationID - Station ID in database
varname - variable name e.g. 'waterlevel', 'discharge', 'salinity'
timeinfo (optional) - tuple with (starttime,endtime,dt). Format 'yyyymmdd.HHMMSS'
Use this to interpolate onto a constant time vector
filttype (optional) - 'low' or 'high'
Set this to filter data
Returns:
timeseries object
-1 on error
"""
from netcdfio import queryNC
outvar = ['NetCDF_Filename','NetCDF_GroupID','StationName']
tablename = 'observations'
#condition = 'Variable_Name = "%s" and StationID = "%s"' % (varname,stationID)
condition = 'Variable_Name = "%s" and StationID LIKE "%%%s"' % (varname,stationID)
print 'Querying database...'
print condition
data, query = queryNC(dbfile,outvar,tablename,condition)
if len(data)==0:
print '!!! Warning - Did not find any stations matching query. Returning -1 !!!'
return -1
else:
ts = timeseries(data[0]['time'],data[0][varname].squeeze())
if not timeinfo==None:
print 'Interpolating station data between %s and %s\n'%(timeinfo[0],timeinfo[1])
tnew,ynew = ts.interp((timeinfo[0],timeinfo[1],timeinfo[2]))
ts = timeseries(tnew,ynew)
ts.dt = timeinfo[2] # This needs updating
if not filttype==None:
print '%s-pass filtering output data. Cutoff period = %f [s].'%(filttype,cutoff)
yfilt = ts.filt(cutoff,btype=filttype,axis=-1)
ts.y = yfilt.copy()
if output_meta:
if data[0].has_key('elevation'):
ele = data[0]['elevation']
else:
ele = 0.0
meta = {'longitude':data[0]['longitude'],'latitude':data[0]['latitude'],'elevation':ele,'StationName':query['StationName'][0]}
return ts, meta
else:
return ts
def interpWeatherStations(latlon,tstart,tend,dt,utmzone,dbfile, maxgap=40, showplot=False):
""" Temporally interpolate weather station data onto a specified time grid"""
###
# Convert to datetime format
timestart = datetime.strptime(tstart,'%Y%m%d')
timeend = datetime.strptime(tend,'%Y%m%d')
# Create the time variable
timeList = []
tnow=timestart
while tnow<timeend:
timeList.append(tnow)
tnow += timedelta(hours=dt)
nctime = convertTime(timeList)
ntime = len(timeList)
varnames = ['Tair','Pair','Uwind','Vwind','RH','rain','cloud']
coords={}
output = {}
# Read in the semi-processed data
for vv in varnames:
print 'Interpolating variable %s...'%vv
outvar = ['NetCDF_Filename','NetCDF_GroupID','StationName']
tablename = 'observations'
condition = 'Variable_Name = "%s"' % vv + \
'and time_start <= "%s"'% datetime.strftime(timestart,'%Y-%m-%d %H:%M:%S') + \
'and time_end >= "%s"'% datetime.strftime(timeend,'%Y-%m-%d %H:%M:%S') + \
'and lon_start >= %3.6f '%latlon[0] + 'and lon_end <= %3.6f '%latlon[1] + \
'and lat_start >= %3.6f '%latlon[2] + 'and lat_end <= %3.6f '%latlon[3]
data, query = netcdfio.queryNC(dbfile,outvar,tablename,condition)
#print data[0].keys()
ii=0
for dd in data:
ind = np.isfinite(np.ravel(dd[vv]))
timenow = convertTime(dd['time'])
timegood = timenow[ind]
if nctime[0] <= timegood[0] or nctime[-1] >= timegood[-1]:
data.pop(ii)
else:
ii+=1
# Remove points that have large gaps
ii=0
for dd in data:
ind = np.isfinite(dd[vv])
timenow = convertTime(dd['time'])
i=-1
for t in timenow:
i+=1
if t < nctime[0]:
t1=i
i=-1
for t in timenow:
i+=1
if t < nctime[-1]:
t2=i
# Find the maximum gap size between the two time limits
gapsize = 0
gap=0
for gg in ind[t1:t2]:
if ~gg:
gap+=1
if gap > gapsize:
gapsize=gap
else:
gap = 0
#print t1,t2,len(timenow),gapsize
if gapsize > maxgap:
print 'Removing data point - gap size %d is > %d'%(gapsize,maxgap)
data.pop(ii)
else:
ii+=1
#print gapsize,percgood
# Work out the number of spatial points of each variable based on quality control
coords['x_'+vv] = []
coords['y_'+vv] = []
coords['z_'+vv] = []
for dd in data:
# Convert to utm
ll = np.hstack((dd['longitude'],dd['latitude']))
xy = ll2utm(ll,utmzone)
coords['x_'+vv].append(xy[0][0])
coords['y_'+vv].append(xy[0][1])
#coords['x_'+vv].append(dd['longitude'])
#coords['y_'+vv].append(dd['latitude'])
coords['z_'+vv].append(dd['elevation'])
varlen = len(data)
# Initialize the output arrays
output[vv] = {'Data':np.zeros((ntime,varlen))}
# Loop trough and interpolate each variables onto the time array
ctr=0
for dd in data:
# Interpolate the data
tmp = np.array(np.ravel(dd[vv]))
timenow = convertTime(dd['time'])
ind = np.isfinite(tmp)
F = interpolate.interp1d(timenow[ind],tmp[ind],kind='linear')
varinterp = F(nctime)
#F = interpolate.InterpolatedUnivariateSpline(timenow[ind],tmp[ind])
#varinterp = F(nctime)
#F = interpolate.splrep(timenow[ind],tmp[ind],s=0)
#varinterp = interpolate.splev(nctime,F,der=0)
#F = interpolate.Rbf(timenow[ind],tmp[ind])
#varinterp = F(nctime)
output[vv]['Data'][:,ctr]=varinterp
ctr+=1
# Add the other info
#output[vv].update({'long_name':dd[vv]['Longname'],'units':dd[vv]['Units']})
if showplot:
plt.figure()
plt.hold('on')
plt.plot(timenow,tmp)
plt.plot(nctime,varinterp,'r')
plt.title(dd['StationName']+' - '+vv)
plt.show()
# Return the data
return coords, output, nctime
def QueryNC(dbfile,staname=None,yearrange=None,cons=None):
"""
Query the tidal station data
"""
outvar = ['NetCDF_Filename','NetCDF_GroupID','StationName','StationID']
tablename = 'observations'
#condition = "Variable_Name = '%s' and (StationName = '%s' or StationName = '%s' or StationName = '%s')" % (varname,staname1,staname2,staname3 )
# Create the query
varname1 = 'ssh_amp'
varname2 = 'ssh_phs'
if staname == None and yearrange == None:
condition = "(Variable_Name = '%s' or Variable_Name = '%s')"%(varname1,varname2)
ydim = None
elif not staname == None and yearrange == None:
condition = "(Variable_Name = '%s' or Variable_Name = '%s') and StationName = '%s'"%(varname1,varname2, staname)
ydim = 'year'
elif not staname == None and not yearrange == None:
t1 = '%s-01-01 00:00:00'%yearrange[0]
t2 = '%4d-01-01 00:00:00'%yearrange[1]
condition = "(Variable_Name = '%s' or Variable_Name = '%s') and StationName = '%s' and time_start > %s and time_start < %s"%(varname1,varname2,staname,t1,t2)
ydim = None
elif staname == None and not yearrange == None:
t1 = '%s-01-01 00:00:00'%yearrange[0]
t2 = '%4d-01-01 00:00:00'%yearrange[1]
condition = "(Variable_Name = '%s' or Variable_Name = '%s') and time_start > '%s' and time_start < '%s'"%(varname1,varname2,t1,t2)
ydim = 'station'
data, query = netcdfio.queryNC(dbfile,outvar,tablename,condition,fastmode=True)
# Read the constituents from the netcdf file
ncfile = query['NetCDF_Filename'][0]
nc = Dataset(ncfile,'r')
#names = nc.__dict__.keys()
names = nc.Tidal_Constituents.split(', ')
#print nc.['Tidal Constituents']
nc.close()
# Find the constituent indices
if cons == None:
cons=names
ind = []
for nn in cons:
ind.append((i for i, j in enumerate(names) if j == nn).next())
# Output the query data into a nicer format
#amp = [if dd.has_key('ssh_amp'): dd['ssh_amp'][[1,3,8]].ravel() for dd in data]
amp = []
phs = []
time = []
lon = []
lat = []
StationName=[]
for ii,dd in enumerate(data):
if dd.has_key('ssh_amp'):
amp.append(dd['ssh_amp'][ind].ravel())
if ydim == 'year':
time.append(dd['time'][0])
elif ydim == 'station':
lon.append(dd['longitude'])
lat.append(dd['latitude'])
StationName.append(query['StationName'][ii])
else:
lon.append(dd['longitude'])
lat.append(dd['latitude'])
StationName.append(query['StationName'][ii])
time.append(dd['time'][0])
if dd.has_key('ssh_phs'):
phs.append(dd['ssh_phs'][ind].ravel())
amp = np.array(amp)
phs = np.array(phs)
# Get the time and station coordinates
if ydim == 'station':
time = data[0]['time'][0]
lon = np.array(lon)
lat = np.array(lat)
elif ydim == 'year':
lon = data[0]['longitude']
lat = data[0]['latitude']
StationName = query['StationName'][0]
time = np.array(time)
return amp, phs, time, lon, lat, StationName, cons
def interpWeatherStations(latlon,
tstart,
tend,
dt,
utmzone,
dbfile,
maxgap=40,
showplot=False):
""" Temporally interpolate weather station data onto a specified time grid"""
###
# Convert to datetime format
timestart = datetime.strptime(tstart, '%Y%m%d')
timeend = datetime.strptime(tend, '%Y%m%d')
# Create the time variable
timeList = []
tnow = timestart
while tnow < timeend:
timeList.append(tnow)
tnow += timedelta(hours=dt)
nctime = convertTime(timeList)
ntime = len(timeList)
varnames = ['Tair', 'Pair', 'Uwind', 'Vwind', 'RH', 'rain', 'cloud']
coords = {}
output = {}
# Read in the semi-processed data
for vv in varnames:
print('Interpolating variable %s...' % vv)
outvar = ['NetCDF_Filename', 'NetCDF_GroupID', 'StationName']
tablename = 'observations'
condition = 'Variable_Name = "%s"' % vv + \
'and time_start <= "%s"'% datetime.strftime(timestart,'%Y-%m-%d %H:%M:%S') + \
'and time_end >= "%s"'% datetime.strftime(timeend,'%Y-%m-%d %H:%M:%S') + \
'and lon_start >= %3.6f '%latlon[0] + 'and lon_end <= %3.6f '%latlon[1] + \
'and lat_start >= %3.6f '%latlon[2] + 'and lat_end <= %3.6f '%latlon[3]
data, query = netcdfio.queryNC(dbfile, outvar, tablename, condition)
#print data[0].keys()
ii = 0
for dd in data:
ind = np.isfinite(np.ravel(dd[vv]))
timenow = convertTime(dd['time'])
timegood = timenow[ind]
if nctime[0] <= timegood[0] or nctime[-1] >= timegood[-1]:
data.pop(ii)
else:
ii += 1
# Remove points that have large gaps
ii = 0
for dd in data:
ind = np.isfinite(dd[vv])
timenow = convertTime(dd['time'])
i = -1
for t in timenow:
i += 1
if t < nctime[0]:
t1 = i
i = -1
for t in timenow:
i += 1
if t < nctime[-1]:
t2 = i
# Find the maximum gap size between the two time limits
gapsize = 0
gap = 0
for gg in ind[t1:t2]:
if ~gg:
gap += 1
if gap > gapsize:
gapsize = gap
else:
gap = 0
#print t1,t2,len(timenow),gapsize
if gapsize > maxgap:
print('Removing data point - gap size %d is > %d' %
(gapsize, maxgap))
data.pop(ii)
else:
ii += 1
#print gapsize,percgood
# Work out the number of spatial points of each variable based on quality control
coords['x_' + vv] = []
coords['y_' + vv] = []
coords['z_' + vv] = []
for dd in data:
# Convert to utm
ll = np.hstack((dd['longitude'], dd['latitude']))
xy = ll2utm(ll, utmzone)
coords['x_' + vv].append(xy[0][0])
coords['y_' + vv].append(xy[0][1])
#coords['x_'+vv].append(dd['longitude'])
#coords['y_'+vv].append(dd['latitude'])
coords['z_' + vv].append(dd['elevation'])
varlen = len(data)
# Initialize the output arrays
output[vv] = {'Data': np.zeros((ntime, varlen))}
# Loop trough and interpolate each variables onto the time array
ctr = 0
for dd in data:
# Interpolate the data
tmp = np.array(np.ravel(dd[vv]))
timenow = convertTime(dd['time'])
ind = np.isfinite(tmp)
F = interpolate.interp1d(timenow[ind], tmp[ind], kind='linear')
varinterp = F(nctime)
#F = interpolate.InterpolatedUnivariateSpline(timenow[ind],tmp[ind])
#varinterp = F(nctime)
#F = interpolate.splrep(timenow[ind],tmp[ind],s=0)
#varinterp = interpolate.splev(nctime,F,der=0)
#F = interpolate.Rbf(timenow[ind],tmp[ind])
#varinterp = F(nctime)
output[vv]['Data'][:, ctr] = varinterp
ctr += 1
# Add the other info
#output[vv].update({'long_name':dd[vv]['Longname'],'units':dd[vv]['Units']})
if showplot:
plt.figure()
plt.hold('on')
plt.plot(timenow, tmp)
plt.plot(nctime, varinterp, 'r')
plt.title(dd['StationName'] + ' - ' + vv)
plt.show()
# Return the data
return coords, output, nctime
def loadDBstation(dbfile,
stationID,
varname,
timeinfo=None,
filttype=None,
cutoff=3600.0,
output_meta=False,
method='linear'):
"""
Load station data from a database file
Inputs:
dbfile - location of database file
stationID - Station ID in database
varname - variable name e.g. 'waterlevel', 'discharge', 'salinity'
timeinfo (optional) - tuple with (starttime,endtime,dt). Format 'yyyymmdd.HHMMSS'
Use this to interpolate onto a constant time vector
filttype (optional) - 'low' or 'high'
Set this to filter data
Returns:
timeseries object
-1 on error
"""
from netcdfio import queryNC
outvar = ['NetCDF_Filename', 'NetCDF_GroupID', 'StationName']
tablename = 'observations'
#condition = 'Variable_Name = "%s" and StationID = "%s"' % (varname,stationID)
condition = 'Variable_Name = "%s" and StationID LIKE "%%%s"' % (varname,
stationID)
print 'Querying database...'
print condition
data, query = queryNC(dbfile, outvar, tablename, condition)
yout = data[0][varname].squeeze()
# Zero nan
yout[np.isnan(yout)] = 0.0
if len(data) == 0:
print '!!! Warning - Did not find any stations matching query. Returning -1 !!!'
return -1
else:
ts = timeseries(data[0]['time'], yout)
if not timeinfo == None:
print 'Interpolating station data between %s and %s\n' % (timeinfo[0],
timeinfo[1])
tnew,ynew =\
ts.interp((timeinfo[0],timeinfo[1],timeinfo[2]),method=method)
ts = timeseries(tnew, ynew)
ts.dt = timeinfo[2] # This needs updating
if not filttype == None:
print '%s-pass filtering output data. Cutoff period = %f [s].' % (
filttype, cutoff)
yfilt = ts.filt(cutoff, btype=filttype, axis=-1)
ts.y = yfilt.copy()
if output_meta:
if data[0].has_key('elevation'):
ele = data[0]['elevation']
else:
ele = np.array([0.0])
meta = {
'longitude': data[0]['longitude'],
'latitude': data[0]['latitude'],
'elevation': ele,
'StationName': query['StationName'][0]
}
return ts, meta
else:
return ts
