def opendss(filename, server='local'):
"""
opendss( filename, server='local') :
opens a dss file with the given filename and returns it in a group object
The group object behaves like an array of data references. Each such data
reference contains the data set.
"""
return DSSUtil.createGroup(server,filename)
def make_flag_value(flag_val):
flag_vals = string.split(flag_val,"|")
if len(flag_vals) !=2 : raise "Invalid flag: %s"%flag_val
flag_type = FlagUtils.getQualityFlagId(flag_vals[0])
user_id = DSSUtil.getUserId(string.lower(flag_vals[1]))
_dummy_dse.setFlag(0)
if flag_type == 0:
FlagUtils.clearAllFlags(_dummy_dse,user_id)
else:
FlagUtils.setQualityFlag(_dummy_dse,flag_type,user_id)
return _dummy_dse.getFlag()
def writedss(filename, pathname, ds) :
"""
writedss(filename,pathname,ds):
writes the given data set ds to the filename and the pathname.
"""
DSSUtil.writeData(filename,pathname,ds)
import re
import string
from vista.set import Constants, DefaultReference,\
DataReference, DataSetAttr, DataType, \
FlagUtils, FlaggedDataSetElement, \
Pathname, \
RegularTimeSeries, IrregularTimeSeries, \
SetUtils, PathnamePredicate, \
PartNamePredicate, PathPartPredicate, \
SortMechanism
from vista.db.dss import DSSUtil
from vista.time import TimeFactory
from vista.app import MainProperties
DSSUtil.setAccessProperties(MainProperties.getProperties())
def wrap_data(ds, filename='', server='',pathname=''):
"""
wrap_data(dataset, filename='', server='', pathname=''
wraps data set in a filename, servername and pathname
"""
return gen_ref(ds,filename,server,pathname)
#
def gen_ref(ds, filename='', server='', pathname=''):
"generates a DataReference given a dataset"
if isinstance(ds, DataReference): return ds
if (filename == '' and server == '' and pathname == ''):
return DefaultReference(ds)
else :
return DefaultReference(server, filename, pathname, ds)
#
def flag_data(ftype, dataset, valArray, log = 'flag.log', \
Special = False, ResetExist = False):
"""
flag_data(ftype, dataset, valArray, log = 'flag.log', Special = False):
Flags a datastream's values for bad data:
ftype='R': datavalue not within valArray[0]:valArray[1] marked as reject.
='D': datavalue difference from previous value not within
valArray[0] range marked as reject. If len(valArray) > 1,
valArray[0] is an exceedance factor for values, and
valArray[1] is an exceedance factor for 1st diff of values,
both compared to Trailing Moving Average
='M': datavalue equals or very close to val[0:], an array of Missing Value markers;
DSS flag set to missing. Optional Special means marker value
is within normal data range, check more carefully.
='+|*': datavalue within range valArray[0]:valArray[1] scaled by valArray[2]
amount, that is, datavalue is added/multiplied to/by valArray[2]. Used mainly
for CDEC EC data that switches between milli- and micro-mhos/cm and
to add 100 to USGS stage data.
All values marked are written to log file.
Special=True: Special treatment for Missing values that are within the
normal operating range of the parameter
ResetExist=True: reset any existing flag values to Unscreened
Flags added to timeseries if needed.
"""
from jarray import zeros
def nearVal (val, target, tol=.001):
# return True if relative error is less than tol
if abs(val-target) < tol:
return True
else:
return False
if ftype == 'R':
if len(valArray) != 2:
# assume hi and lo range will be calculated from percentiles
yVals = sorted(SetUtils.createYArray(dataset))
# lo and hi: 1st and 99th percentiles
valArray = [yVals[int(len(yVals)*0.01)], yVals[int(len(yVals)*0.99)]]
rej_head = 'Check range ' + str(valArray[0]) + ' - ' + str(valArray[1])
rej_note = 'Range reject @ '
elif ftype == 'D':
nMoveAve = 5 # number of good elements in moving average
moveAve = 0.0
moveAveDiff = 0.0
xTA = zeros(nMoveAve,'d')
diffFactor = False
rej_head = 'Check diff w/ moving average ' + str(valArray[0])
rej_note = 'Diff reject @ '
if len(valArray) > 1: # factor
diffFactor = True
rej_head = 'Check vals, 1st diffs w/ MA ' + str(valArray[0])+ ', ' + str(valArray[1]) + ' factors'
# yVals = sorted(SetUtils.createYArray(dataset))
# # difference between 10th and 90th percentiles
# pdiff = abs(yVals[int(len(yVals)*0.10)] - yVals[int(len(yVals)*0.90)])
rej_note = 'Diff reject % @ '
elif ftype == 'M':
if len(valArray) < 1: raise 'At least one value must be given for Missing check.'
rej_head = 'Check Missing value marker ' + str(valArray)
rej_note = 'Missing @ '
elif ftype == '+' or ftype == '*':
if len(valArray) != 3: raise 'Three values must be given for Shift.'
rej_head = 'Check scale ' + str(valArray[0]) + ' - ' + str(valArray[1])
rej_note = 'Value scaled @ '
else: raise 'First arg must be a single character R, D or M.'
# nominal time interval in minutes to detect gaps in data
intvls = {'IR-DAY': 15, 'IR-MON': 60, 'IR-YEAR': 1440, \
'2MIN': 2, '10MIN': 10, '15MIN': 15, '1HOUR': 60, \
'1DAY': 1440, '1MON': 43920}
ePart = dataset.getName().split('/')[5]
nomIntvl = intvls[ePart]
# a flag to check if any flag was changed
changedFlag = False
# get the filter for missing values
filter = Constants.DEFAULT_FLAG_FILTER
# check if ds already has flags, if not, make them
# open log file
logfile = open(log, 'a')
logfile.write('\n\n' + 'Name: ' + dataset.getName())
logfile.write('\n' + 'Units: ' + dataset.getAttributes().getYUnits())
logfile.write('\n' + rej_head)
if dataset.isFlagged(): ds = dataset
else: ds = ds_add_flags(dataset)
# get user id for setting flags
uId = DSSUtil.getUserId('datachecker')
# create a missing data element
ex = dataset.getElementAt(0)
ex.setY(Constants.MISSING_VALUE)
ex.setFlag(FlagUtils.MISSING_FLAG)
eBad = None
intvl = None
nGood = 0
for i in range(dataset.size()):
changedEl = False
e1 = dataset.getElementAt(i)
if ResetExist:
FlagUtils.clearAllFlags(e1, uId)
dataset.putElementAt(i, e1) # put the element back into the data set
if not filter.isAcceptable(e1) or \
FlagUtils.getQualityFlag(e1) == FlagUtils.MISSING_FLAG: continue
# get the data elements at the i, i-1, and i+1 positions
if i > 0:
e0 = dataset.getElementAt(i - 1)
intvl = int(e1.getX() - e0.getX() + .01)
else:
e0 = ex
if i < len(dataset) - 1:
e2 = dataset.getElementAt(i + 1)
else:
e2 = ex
if ftype == 'R': # Range
if e1.y < valArray[0] or e1.y > valArray[1] :
FlagUtils.setQualityFlag(e1, FlagUtils.REJECT_FLAG, uId)
changedEl = True
elif ftype == 'D': # Difference (abs or %) between this and moving-ave value
# if large data time gap noted, flag element as questionable and reset moving average
if intvl > nomIntvl*6:
FlagUtils.setQualityFlag(e1, FlagUtils.QUESTIONABLE_FLAG, uId)
changedEl = True
moveAve = 0
xTA = zeros(nMoveAve,'d')
nGood = 0
if not diffFactor:
diff1 = abs(e1.y - moveAve)
else:
if moveAve <> 0:
diff1 = abs(e1.y / moveAve)
else:
diff1 = 1
if moveAveDiff <> 0:
diff2 = abs((e1.y - e0.y) / moveAveDiff)
else:
diff2 = 0
if diff2 == 0:
diff2 = .1
if diff2 < 1:
diff2 = 1 / diff2
# check for spikes and longer plateaus near a bad value
if diff1 == 0:
diff1 = .1
if diff1 < 1:
diff1 = 1 / diff1
if (nGood >= nMoveAve) and \
( (diff1 > valArray[0] and (not diffFactor or (diffFactor and diff2 > valArray[1]))) or \
(eBad and nearVal(e1.y, eBad.y, abs(eBad.y*.001)))):
FlagUtils.setQualityFlag(e1, FlagUtils.REJECT_FLAG, uId)
changedEl = True
if not eBad:
eBad = e1
else:
eBad = None
if filter.isAcceptable(e1) and \
FlagUtils.getQualityFlag(e1) != FlagUtils.MISSING_FLAG:
nGood += 1
xTA.pop(0)
xTA.append(e1.y)
# moving average of data values
moveAve = sum(xTA) / nMoveAve
if diffFactor:
# moving average of 1st difference of data values
xTA1 = xTA[1:nMoveAve]
xTA0 = xTA[0:nMoveAve-1]
moveAveDiff = sum([abs(xTA1[j]-xTA0[j]) for j in range(nMoveAve-1)]) / (nMoveAve -1)
elif ftype == 'M': # Missing values
for vA in valArray:
if nearVal(vA, e1.y):
if not Special or (Special and \
# Special treatment for Missing values that are within the
# normal operating range of the parameter; check that the value
# before or after is also Missing or not acceptable before
# marking this value as Missing
(not filter.isAcceptable(e0) or not filter.isAcceptable(e2)) or \
(nearVal(vA, e0.y) or nearVal(vA, e2.y))):
FlagUtils.setQualityFlag(e1, FlagUtils.MISSING_FLAG, uId)
#e1.y=Constants.MISSING_VALUE
changedEl = True
elif ftype == '+': # re-scale
if not filter.isAcceptable(e1): continue
if e1.y >= valArray[0] and e1.y <= valArray[1] :
e1.y += valArray[2]
changedEl = True
elif ftype == '*': # re-scale
if not filter.isAcceptable(e1): continue
if e1.y >= valArray[0] and e1.y <= valArray[1] :
e1.y *= valArray[2]
changedEl = True
if changedEl:
changedFlag = True
dataset.putElementAt(i, e1) # put the element back into the data set
logfile.write('\n' + rej_note + e1.getXString() + " : " + e1.getYString())
# end the for loop
logfile.close()
if changedFlag or ResetExist:
return ds
else:
return None
System.loadLibrary("errno")
System.loadLibrary("posix")
except:
pass
# check for display
from java.awt import Toolkit
display = 1
try :
tk = Toolkit.getDefaultToolkit()
except:
print 'Problem with display'
display = 0
#
from vista.db.dss import DSSUtil
from vista.app import MainProperties
DSSUtil.setAccessProperties(MainProperties.getProperties())
#
from vdss import *
from vtimeseries import *
from vdisplay import *
#
def exit():
sys.exit()
#
def flag_data(ftype, dataset, valArray, log = 'flag.log', \
Special = False, ResetExist = False):
"""
flag_data(ftype, dataset, valArray, log = 'flag.log', Special = False):
Flags a datastream's values for bad data:
ftype='R': datavalue not within valArray[0]:valArray[1] marked as reject.
='D': datavalue difference from previous value not within
def flag_data(ftype, dataset, valArray, log = 'flag.log', \
Special = False, ResetExist = False):
"""
flag_data(ftype, dataset, valArray, log = 'flag.log', Special = False):
Flags a datastream's values for bad data:
ftype='R': datavalue not within valArray[0]:valArray[1] marked as reject.
='D': datavalue difference from previous value not within
valArray[0] range marked as reject. If len(valArray) > 1,
valArray[0] is an exceedance factor for values, and
valArray[1] is an exceedance factor for 1st diff of values,
both compared to Trailing Moving Average
='M': datavalue equals or very close to val[0:], an array of Missing Value markers;
DSS flag set to missing. Optional Special means marker value
is within normal data range, check more carefully.
='+|*': datavalue within range valArray[0]:valArray[1] scaled by valArray[2]
amount, that is, datavalue is added/multiplied to/by valArray[2]. Used mainly
for CDEC EC data that switches between milli- and micro-mhos/cm and
to add 100 to USGS stage data.
All values marked are written to log file.
Special=True: Special treatment for Missing values that are within the
normal operating range of the parameter
ResetExist=True: reset any existing flag values to Unscreened
Flags added to timeseries if needed.
"""
from jarray import zeros
def nearVal(val, target, tol=.001):
# return True if relative error is less than tol
if abs(val - target) < tol:
return True
else:
return False
if ftype == 'R':
if len(valArray) != 2:
# assume hi and lo range will be calculated from percentiles
yVals = sorted(SetUtils.createYArray(dataset))
# lo and hi: 1st and 99th percentiles
valArray = [
yVals[int(len(yVals) * 0.01)], yVals[int(len(yVals) * 0.99)]
]
rej_head = 'Check range ' + str(valArray[0]) + ' - ' + str(valArray[1])
rej_note = 'Range reject @ '
elif ftype == 'D':
nMoveAve = 5 # number of good elements in moving average
moveAve = 0.0
moveAveDiff = 0.0
xTA = zeros(nMoveAve, 'd')
diffFactor = False
rej_head = 'Check diff w/ moving average ' + str(valArray[0])
rej_note = 'Diff reject @ '
if len(valArray) > 1: # factor
diffFactor = True
rej_head = 'Check vals, 1st diffs w/ MA ' + str(
valArray[0]) + ', ' + str(valArray[1]) + ' factors'
# yVals = sorted(SetUtils.createYArray(dataset))
# # difference between 10th and 90th percentiles
# pdiff = abs(yVals[int(len(yVals)*0.10)] - yVals[int(len(yVals)*0.90)])
rej_note = 'Diff reject % @ '
elif ftype == 'M':
if len(valArray) < 1:
raise 'At least one value must be given for Missing check.'
rej_head = 'Check Missing value marker ' + str(valArray)
rej_note = 'Missing @ '
elif ftype == '+' or ftype == '*':
if len(valArray) != 3: raise 'Three values must be given for Shift.'
rej_head = 'Check scale ' + str(valArray[0]) + ' - ' + str(valArray[1])
rej_note = 'Value scaled @ '
else:
raise 'First arg must be a single character R, D or M.'
# nominal time interval in minutes to detect gaps in data
intvls = {'IR-DAY': 15, 'IR-MON': 60, 'IR-YEAR': 1440, \
'2MIN': 2, '10MIN': 10, '15MIN': 15, '1HOUR': 60, \
'1DAY': 1440, '1MON': 43920}
ePart = dataset.getName().split('/')[5]
nomIntvl = intvls[ePart]
# a flag to check if any flag was changed
changedFlag = False
# get the filter for missing values
filter = Constants.DEFAULT_FLAG_FILTER
# check if ds already has flags, if not, make them
# open log file
logfile = open(log, 'a')
logfile.write('\n\n' + 'Name: ' + dataset.getName())
logfile.write('\n' + 'Units: ' + dataset.getAttributes().getYUnits())
logfile.write('\n' + rej_head)
if dataset.isFlagged(): ds = dataset
else: ds = ds_add_flags(dataset)
# get user id for setting flags
uId = DSSUtil.getUserId('datachecker')
# create a missing data element
ex = dataset.getElementAt(0)
ex.setY(Constants.MISSING_VALUE)
ex.setFlag(FlagUtils.MISSING_FLAG)
eBad = None
intvl = None
nGood = 0
for i in range(dataset.size()):
changedEl = False
e1 = dataset.getElementAt(i)
if ResetExist:
FlagUtils.clearAllFlags(e1, uId)
dataset.putElementAt(i,
e1) # put the element back into the data set
if not filter.isAcceptable(e1) or \
FlagUtils.getQualityFlag(e1) == FlagUtils.MISSING_FLAG:
continue
# get the data elements at the i, i-1, and i+1 positions
if i > 0:
e0 = dataset.getElementAt(i - 1)
intvl = int(e1.getX() - e0.getX() + .01)
else:
e0 = ex
if i < len(dataset) - 1:
e2 = dataset.getElementAt(i + 1)
else:
e2 = ex
if ftype == 'R': # Range
if e1.y < valArray[0] or e1.y > valArray[1]:
FlagUtils.setQualityFlag(e1, FlagUtils.REJECT_FLAG, uId)
changedEl = True
elif ftype == 'D': # Difference (abs or %) between this and moving-ave value
# if large data time gap noted, flag element as questionable and reset moving average
if intvl > nomIntvl * 6:
FlagUtils.setQualityFlag(e1, FlagUtils.QUESTIONABLE_FLAG, uId)
changedEl = True
moveAve = 0
xTA = zeros(nMoveAve, 'd')
nGood = 0
if not diffFactor:
diff1 = abs(e1.y - moveAve)
else:
if moveAve <> 0:
diff1 = abs(e1.y / moveAve)
else:
diff1 = 1
if moveAveDiff <> 0:
diff2 = abs((e1.y - e0.y) / moveAveDiff)
else:
diff2 = 0
if diff2 == 0:
diff2 = .1
if diff2 < 1:
diff2 = 1 / diff2
# check for spikes and longer plateaus near a bad value
if diff1 == 0:
diff1 = .1
if diff1 < 1:
diff1 = 1 / diff1
if (nGood >= nMoveAve) and \
( (diff1 > valArray[0] and (not diffFactor or (diffFactor and diff2 > valArray[1]))) or \
(eBad and nearVal(e1.y, eBad.y, abs(eBad.y*.001)))):
FlagUtils.setQualityFlag(e1, FlagUtils.REJECT_FLAG, uId)
changedEl = True
if not eBad:
eBad = e1
else:
eBad = None
if filter.isAcceptable(e1) and \
FlagUtils.getQualityFlag(e1) != FlagUtils.MISSING_FLAG:
nGood += 1
xTA.pop(0)
xTA.append(e1.y)
# moving average of data values
moveAve = sum(xTA) / nMoveAve
if diffFactor:
# moving average of 1st difference of data values
xTA1 = xTA[1:nMoveAve]
xTA0 = xTA[0:nMoveAve - 1]
moveAveDiff = sum(
[abs(xTA1[j] - xTA0[j])
for j in range(nMoveAve - 1)]) / (nMoveAve - 1)
elif ftype == 'M': # Missing values
for vA in valArray:
if nearVal(vA, e1.y):
if not Special or (Special and \
# Special treatment for Missing values that are within the
# normal operating range of the parameter; check that the value
# before or after is also Missing or not acceptable before
# marking this value as Missing
(not filter.isAcceptable(e0) or not filter.isAcceptable(e2)) or \
(nearVal(vA, e0.y) or nearVal(vA, e2.y))):
FlagUtils.setQualityFlag(e1, FlagUtils.MISSING_FLAG,
uId)
#e1.y=Constants.MISSING_VALUE
changedEl = True
elif ftype == '+': # re-scale
if not filter.isAcceptable(e1): continue
if e1.y >= valArray[0] and e1.y <= valArray[1]:
e1.y += valArray[2]
changedEl = True
elif ftype == '*': # re-scale
if not filter.isAcceptable(e1): continue
if e1.y >= valArray[0] and e1.y <= valArray[1]:
e1.y *= valArray[2]
changedEl = True
if changedEl:
changedFlag = True
dataset.putElementAt(i,
e1) # put the element back into the data set
logfile.write('\n' + rej_note + e1.getXString() + " : " +
e1.getYString())
# end the for loop
logfile.close()
if changedFlag or ResetExist:
return ds
else:
return None
except:
pass
# check for display
from java.awt import Toolkit
display = 1
try:
tk = Toolkit.getDefaultToolkit()
except:
print 'Problem with display'
display = 0
#
from vista.db.dss import DSSUtil
from vista.app import MainProperties
DSSUtil.setAccessProperties(MainProperties.getProperties())
#
from vdss import *
from vtimeseries import *
from vdisplay import *
#
def exit():
sys.exit()
#
def flag_data(ftype, dataset, valArray, log = 'flag.log', \
Special = False, ResetExist = False):
"""
