From d784bdf3750b7cf595aa230c5406efd42f2973c9 Mon Sep 17 00:00:00 2001
From: Felix Kleinert <f.kleinert@fz-juelich.de>
Date: Thu, 19 Aug 2021 09:17:42 +0200
Subject: [PATCH] include solar_time conv.
---
mlair/data_handler/data_handler_wrf_chem.py | 41 ++++++++++++++-------
1 file changed, 27 insertions(+), 14 deletions(-)
diff --git a/mlair/data_handler/data_handler_wrf_chem.py b/mlair/data_handler/data_handler_wrf_chem.py
index e117effa..4dbaf827 100644
--- a/mlair/data_handler/data_handler_wrf_chem.py
+++ b/mlair/data_handler/data_handler_wrf_chem.py
@@ -9,6 +9,7 @@ import glob
import matplotlib.pyplot as plt
from dask.diagnostics import ProgressBar
from tzwhere import tzwhere
+import datetime as dt
from toarstats import toarstats
import hashlib
@@ -413,8 +414,8 @@ class SingleGridColumnWrfChemDataLoader(BaseWrfChemDataLoader):
window_lead_time=DEFAULT_WINDOW_LEAD_TIME,
external_coords_file: str = None,
wind_sectors=None, wind_sector_edge_dim_name=None, statistics_per_var: Dict = None,
- aggregation_sampling: str = None, time_zone: str = None, station: str = None,
- lazy_preprocessing: bool = False, **kwargs):
+ aggregation_sampling: str = None, time_zone: str = None, target_time_type: str = None,
+ station: str = None, lazy_preprocessing: bool = False, **kwargs):
"""
@@ -453,6 +454,7 @@ class SingleGridColumnWrfChemDataLoader(BaseWrfChemDataLoader):
self.statistics_per_var = statistics_per_var
self.aggregation_sampling = aggregation_sampling
self.time_zone = time_zone
+ self.target_time_type = target_time_type
self.station = station
self.lazy = lazy_preprocessing
@@ -696,10 +698,13 @@ class SingleGridColumnWrfChemDataLoader(BaseWrfChemDataLoader):
meta = pd.DataFrame({self.physical_x_coord_name: self.get_coordinates()['lon'].tolist(),
self.physical_y_coord_name: self.get_coordinates()['lat'].tolist()},
index=self.station)
- local_time_zone = self.get_local_time_zone_from_lat_lon(meta=meta)
+ if self.target_time_type is None:
+ target_time_type_or_zone = self.get_local_time_zone_from_lat_lon(meta=meta)
+ else:
+ target_time_type_or_zone = self.target_time_type
collector = []
for var in self.statistics_per_var.keys():
- collector.append(self.__toarstats_aggregation(data, local_time_zone, meta, target_sampling, var))
+ collector.append(self.__toarstats_aggregation(data, target_time_type_or_zone, meta, target_sampling, var))
sampling_data = xr.merge(collector).dropna(self.physical_t_coord_name)
# sampling_data.attrs = data.attrs
missing_squeezed_coords = data.coords._names - sampling_data.coords._names
@@ -708,7 +713,7 @@ class SingleGridColumnWrfChemDataLoader(BaseWrfChemDataLoader):
return sampling_data
- def __toarstats_aggregation(self, data, local_time_zone, meta, target_sampling, var):
+ def __toarstats_aggregation(self, data, target_time_type_or_zone, meta, target_sampling, var):
with TimeTracking(name=f"{self.station}: apply toarstats `{self.statistics_per_var[var]}({var})`"):
spatial_dims = list(remove_items(data[var].dims, self.physical_t_coord_name))
try:
@@ -718,7 +723,7 @@ class SingleGridColumnWrfChemDataLoader(BaseWrfChemDataLoader):
df = data[var].to_dataframe()[[var]].reset_index(level=spatial_dims)
df = df[[var] + spatial_dims]
- df = self.set_external_time_zone_and_convert_to_local_time_zone(df, local_time_zone)
+ df = self.set_external_time_zone_and_convert_to_targer_time(df, target_time_type_or_zone)
df = df.groupby(spatial_dims)
df = df.apply(self._toarstats_wrapper, sampling=target_sampling, statistics=self.statistics_per_var[var],
metadata=(meta[self.physical_y_coord_name], meta[self.physical_x_coord_name]))
@@ -730,21 +735,26 @@ class SingleGridColumnWrfChemDataLoader(BaseWrfChemDataLoader):
# collector.append(df)
return df
- def set_external_time_zone_and_convert_to_local_time_zone(self, data, local_time_zone):
+ def set_external_time_zone_and_convert_to_targer_time(self, data, target_time_type_or_zone):
"""
:param data:
:type data:
- :param local_time_zone:
- :type local_time_zone:
+ :param target_time_type_or_zone:
+ :type target_time_type_or_zone:
:return:
:rtype:
"""
hdata = data
# hdata = data.squeeze().to_pandas()
- hdata.index = self.set_time_zone(hdata.index)
- hdata.index = hdata.index.tz_convert(local_time_zone)
- logging.debug(f"Set local time zone to: {local_time_zone}")
+ if target_time_type_or_zone == "solar_time":
+ lon = self.get_coordinates()['lon']
+ toffset = dt.timedelta(hours=np.floor(lon/15.))
+ hdata.index = hdata.index + toffset
+ else:
+ hdata.index = self.set_time_zone(hdata.index)
+ hdata.index = hdata.index.tz_convert(target_time_type_or_zone)
+ logging.debug(f"Set local time zone to: {target_time_type_or_zone}")
return hdata
def get_local_time_zone_from_lat_lon(self, lat=None, lon=None, meta=None):
@@ -798,6 +808,7 @@ class DataHandlerSingleGridColumn(DataHandlerSingleStation):
date_format_of_nc_file=None,
as_image_like_data_format=True,
time_zone=None,
+ target_time_type=None,
input_output_sampling4toarstats : tuple = None,
**kwargs):
self.external_coords_file = external_coords_file
@@ -814,6 +825,7 @@ class DataHandlerSingleGridColumn(DataHandlerSingleStation):
self.as_image_like_data_format = as_image_like_data_format
self.time_zone = time_zone
+ self.target_time_type = target_time_type
self.input_output_sampling4toarstats = input_output_sampling4toarstats
super().__init__(*args, **kwargs)
@@ -899,6 +911,7 @@ class DataHandlerSingleGridColumn(DataHandlerSingleStation):
statistics_per_var=self.statistics_per_var,
aggregation_sampling=self.input_output_sampling4toarstats[1],
time_zone=self.time_zone,
+ target_time_type=self.target_time_type,
station=self.station,
lazy_preprocessing=True,
)
@@ -947,7 +960,7 @@ class DataHandlerSingleGridColumn(DataHandlerSingleStation):
# @TimeTrackingWrapper
# def prepare_and_apply_toarstats(self, data, meta, target_sampling="daily"):
# local_time_zone = self.get_local_time_zone_from_lat_lon(meta=meta)
- # hdata = self.set_external_time_zone_and_convert_to_local_time_zone(data, local_time_zone)
+ # hdata = self.set_external_time_zone_and_convert_to_targer_time(data, local_time_zone)
# hsampling_data = []
# for i, var in enumerate(hdata.columns):
# hdf = toarstats(target_sampling, self.statistics_per_var[var], hdata[var],
@@ -964,7 +977,7 @@ class DataHandlerSingleGridColumn(DataHandlerSingleStation):
#
# return self._force_dask_computation(sampling_data)
#
- # def set_external_time_zone_and_convert_to_local_time_zone(self, data, local_time_zone):
+ # def set_external_time_zone_and_convert_to_targer_time(self, data, local_time_zone):
# """
#
# :param data:
--
GitLab