def _load_omn_data(self):
"""
Load omn data
"""
# get the time range from onset data
omnStartDate = self.onsetDF.index.min() - datetime.timedelta(\
minutes=(self.omnHistory+self.omn_time_delay+10))
omnEndDate = self.onsetDF.index.max() + datetime.timedelta(\
minutes=10)
# create the obj and load data
omnObj = omn_utils.OmnData(omnStartDate, omnEndDate, self.omnDBDir,\
self.omnDbName, self.omnTabName,\
self.omnTrain, self.omnNormParamFile,\
imf_normalize=self.imfNormalize,\
db_time_resolution=self.omnDBRes,\
omn_train_params = self.omnTrainParams,\
sml_train=self.sml_train,\
sml_norm_file=self.sml_norm_file,\
smlDbName=self.smlDbName,\
sml_normalize=self.sml_normalize,\
smlTabName=self.smlTabName,\
include_omn=self.include_omn,\
include_sml=self.include_sml,\
sml_train_params=self.sml_train_params,\
omn_time_delay=self.omn_time_delay)
# set the datetime as index since we are working off of it
omnObj.omnDF = omnObj.omnDF.set_index("datetime")
omnObj.omnDF = omnObj.omnDF[self.omnTrainParams +
self.sml_train_params]
return omnObj.omnDF
def _load_omn_data(self, omnDbName, omnTabName):
"""
Load omn data
"""
# get the time range from onset data
omnStartDate = self.paramTimeRange[0] - datetime.timedelta(\
minutes=self.plotTimeHist)
omnEndDate = self.paramTimeRange[1]
# create the obj and load data
omnObj = omn_utils.OmnData(omnStartDate, omnEndDate, self.paramDBDir,\
omnDbName, omnTabName,\
True, None,\
imf_normalize=False,\
db_time_resolution=1,\
omn_train_params = self.omnParams)
return omnObj.omnDF
def load_omn_data(omnStartDate, omnEndDate, omnDBDir,
omnDbName, omnTabName,
omnTrain, omnNormParamFile,
imf_normalize=True,
db_time_resolution=1,
omn_train_params=["By", "Bz", "Bx", "Vx", "Np"]):
# create the obj and load data
omnObj = omn_utils.OmnData(omnStartDate, omnEndDate, omnDBDir,
omnDbName, omnTabName,
omnTrain, omnNormParamFile,
imf_normalize=imfNormalize,
db_time_resolution=omnDBRes,
omn_train_params = omnTrainParams)
# Set the datetime as index since we are working off of it
omnObj.omnDF = omnObj.omnDF.set_index(omnObj.omnDF["datetime"])
omnObj.omnDF = omnObj.omnDF[omnTrainParams]
return omnObj.omnDF
# load onset data
ssOnsetDF = pandas.read_csv(smlFname,\
parse_dates=["Date_UTC"])
ssOnsetDF.columns = ["date", "mlat", "mlt"]
print("loaded onset data")
# get the time range and load OMNI data
if predDateRange is not None:
omnStartDate = predDateRange[0] - datetime.timedelta(minutes=60)
omnEndDate = predDateRange[1]
else:
omnStartDate = ssOnsetDF["date"].min() - datetime.timedelta(minutes=60)
omnEndDate = ssOnsetDF["date"].max()
# create the obj and load data
omnObj = omn_utils.OmnData(omnStartDate, omnEndDate, omnDBDir,\
omnDbName, omnTabName,\
True, None,\
imf_normalize=False,\
db_time_resolution=1,\
omn_train_params = omnParams)
# add a 10 minute time delay to the omni dataset
omnObj.omnDF["datetime"] = omnObj.omnDF["datetime"] + datetime.timedelta(
minutes=omnTimeDelay)
omnObj.omnDF.set_index(omnObj.omnDF["datetime"], inplace=True)
# loop through and make the predictions
# using the conditions
predDict = {}
predDict["date"] = []
predDict["prediction"] = []
predDict["actual"] = []
predDict["onsetDate"] = []
predDict["triggerTime"] = []