def __init__(self, file_list, figsize=None):
if figsize is None:
figsize = (12, 6)
self.figsize = figsize
self.files = od()
for fp in file_list:
self.files[os.path.basename(fp)] = fp
self._f = list(self.files.keys())
self.dropdown_left = ipw.Dropdown(options=self._f, value=self._f[0])
self.dropdown_right = ipw.Dropdown(options=self._f, value=self._f[0])
self.dropdown_left.observe(self.on_change_left)
self.dropdown_right.observe(self.on_change_right)
self.data = pya.ColocatedData()
self.output_left = ipw.Output()
self.output_right = ipw.Output()
layout_left = ipw.VBox([self.dropdown_left, self.output_left])
layout_right = ipw.VBox([self.dropdown_right, self.output_right])
self.layout = ipw.HBox([layout_left, layout_right])
self.plot_scatter_left(file_list[0])
self.plot_scatter_right(file_list[1])
def compute_statistics(self, crop_value_range_var=True):
if not len(self._meta_info) == self.number_of_files:
self.read_meta_from_files()
d = pya.ColocatedData()
f = ProgressBarLabelled(
len(self.files), 'Computing statistics ({} '
'files)'.format(self.number_of_files))
for i, file in enumerate(self.files):
data = d.read_netcdf(file)
if crop_value_range_var:
var = pya.Variable(data.meta['var_name'][1])
self._meta_info[i].update(
data.calc_statistics(lowlim=var.lower_limit,
highlim=var.upper_limit))
else:
self._meta_info[i].update(data.calc_statistics())
if f is not None:
f.value += 1
self._stats_computed = True
def read_meta_from_files(self):
results = []
self._stats_computed = False
data = pya.ColocatedData()
f = ProgressBarLabelled(
len(self.files),
'Reading meta from {} files'.format(self.number_of_files))
my_all = []
vo_all = []
for file in self.files:
info = data.get_meta_from_filename(file)
info['model_id'] = info['data_source'][1]
info['obs_id'] = info['data_source'][0]
info['year'] = info['start'].year
info['file'] = file
vo = [info['var_name'], info['obs_id']]
my = [info['model_id'], info['year']]
found = False
for _vo in vo_all:
if _vo == vo:
found = True
break
if not found:
vo_all.append(vo)
found = False
for _my in my_all:
if _my == my:
found = True
break
if not found:
my_all.append(my)
results.append(info)
if f is not None:
f.value += 1
self._meta_info = results
self.var_obs_combinations = vo_all
self.model_year_combinations = my_all
def read(self, filename):
return pya.ColocatedData(filename)
def compute_statistics_table(self,
var_obs_combinations=None,
model_year_combinations=None,
crop_value_range_var=True):
if var_obs_combinations is None:
var_obs_combinations = self.var_obs_combinations
if model_year_combinations is None:
model_year_combinations = self.model_year_combinations
meta_info = self._meta_info
header = [
'model_id', 'year', 'var_name', 'obs_id', 'ts_type', 'ts_type_src',
'nmb', 'rms', 'R', 'R_kendall', 'fge'
]
data = []
f = ProgressBarLabelled(
len(meta_info),
'Computing statistics table ({} files)'.format(len(meta_info)))
d = pya.ColocatedData()
for info in meta_info:
vo_match, my_match = False, False
for vo in var_obs_combinations:
if [info['var_name'], info['obs_id']] == vo:
vo_match = True
break
for my in model_year_combinations:
if [info['model_id'], info['year']] == my:
my_match = True
break
if vo_match and my_match:
if not 'nmb' in info:
dat = d.read_netcdf(info['file'])
if crop_value_range_var:
var = pya.Variable(dat.meta['var_name'][1])
# =============================================================================
# print(dat.meta['var_name'][1], var.lower_limit,
# var.upper_limit)
# =============================================================================
info.update(
dat.calc_statistics(lowlim=var.lower_limit,
highlim=var.upper_limit))
else:
info.update(dat.calc_statistics())
file_data = [
info['model_id'], info['year'], info['var_name'],
info['obs_id'], info['ts_type'], info['ts_type_src'],
info['nmb'], info['rms'], info['R'], info['R_kendall'],
info['fge']
]
data.append(file_data)
if f is not None:
f.value += 1
df = pd.DataFrame(data, columns=header)
df.set_index(['model_id', 'year', 'var_name', 'obs_id'], inplace=True)
df.sort_index(inplace=True)
self.stats_table = df
return df