import find_precursors
SKIP_STATES = ['NEW YORK']
months = ['August', 'July', 'June'];
metrics = ['corrcoef', 'r2_score'] # 'MAE', 'RMSE',
xr_score = xarray.copy() ; xr_score.attrs = {}
list_xr = [xr_score.copy().expand_dims('metric', axis=0) for m in metrics]
xr_score = xr.concat(list_xr, dim = 'metric')
xr_score['metric'] = ('metric', metrics)
list_xr = [xr_score.expand_dims('month', axis=0) for m in months]
xr_score = xr.concat(list_xr, dim = 'month')
xr_score['month'] = ('month', months)
for im, month in enumerate(xr_score.month.values):
for s, metric in enumerate(xr_score.metric.values):
int_metric = {}
for STATE in All_states:
if STATE not in df_score_States.index or STATE in SKIP_STATES:
continue
US_States_format = ' '.join([s.lower().capitalize() for s in STATE.split(' ')])
abbrev = enums.us_state_abbrev[US_States_format]
integer = States.__dict__[abbrev].real
score = df_score_States.loc[STATE].loc[month, metric]
int_metric[integer] = score
xr_score[im, s] = find_precursors.view_or_replace_labels(xarray.copy(),
list(int_metric.keys()),
list(int_metric.values()))
#%%
fig = plot_maps.plot_corr_maps(xr_score, col_dim='month', row_dim='metric',
size=4, clevels=np.arange(-.5,0.51,.1),
cbar_vert=0, hspace=.1, add_cfeature='LAKES')
's': size,
'zorder': 2,
'color': colors,
'edgecolors': 'black'
}
]
]]
# regions= list(np.unique(xrclustered)[~np.isnan(np.unique(xrclustered))])
# if region == 'USCA':
# dic = {4:3, 3:4}
# else:
# dic = {2:3, 3:2}
# xrclustered = find_precursors.view_or_replace_labels(xrclustered.copy(), regions,
# [int(dic.get(n, n)) for n in regions])
if region == 'USCA':
mask_cl_n = find_precursors.view_or_replace_labels(xrclustered.copy(),
[1, 5])
mask_cl_n = make_country_mask.binary_erosion(~np.isnan(mask_cl_n))
mask_cl_s = ~np.isnan(
find_precursors.view_or_replace_labels(xrclustered.copy(), [2]))
mask_cl = ~np.logical_or(mask_cl_n, mask_cl_s)
elif region == 'USCAnew':
mask_cl = find_precursors.view_or_replace_labels(xrclustered.copy(),
[1, 4])
mask_cl = np.isnan(mask_cl)
elif region == 'init':
mask_cl_e = find_precursors.view_or_replace_labels(xrclustered.copy(), [3])
mask_cl_e = make_country_mask.binary_erosion(~np.isnan(mask_cl_e))
mask_cl_w = ~np.isnan(
find_precursors.view_or_replace_labels(xrclustered.copy(), [1]))
mask_cl = ~np.logical_or(mask_cl_w, mask_cl_e)
def process(rg, lags, fold_method, crossyr):
import find_precursors, plot_maps
#Preprocess precursors
rg.pp_precursors(detrend=True,
anomaly=True,
selbox=None,
format_lon='west_east')
#set any nan value in ts to 0
# ds = rg.get_clust(format_lon='west_east')['ts'][:]
# ds = ds[np.where(np.isnan(rg.get_clust(format_lon='west_east')['ts'][:]))]
# rg.get_clust(format_lon='west_east')['ts'][np.where(np.isnan(rg.get_clust(format_lon='west_east')['ts'][:]))] = 0.0
# ts plot
rg.df_fullts.plot()
# define train and test periods
rg.traintest(method=fold_method, seed=1)
testyrs = rg._get_testyrs()
print(testyrs)
# save target region plot
target_cluster = int(rg.list_of_name_path[0][0])
xrclustered = rg.get_clust(format_lon='west_east')['xrclustered']
fig = plot_maps.plot_labels(
find_precursors.view_or_replace_labels(xrclustered,
regions=target_cluster))
fig.savefig(
os.path.join(rg.path_outsub1, 'target_cluster_{target_cluster}.jpeg'))
# calculate correlation maps
rg.calc_corr_maps()
# show correlation maps
rg.plot_maps_corr(kwrgs_plot={'clevels': np.arange(-.6, .61, 0.1)})
#
rg.cluster_list_MI()
# define period names
period_dict = {
'01': 'January',
'02': 'February',
'03': 'March',
'04': 'April',
'05': 'May',
'06': 'June',
'07': 'July',
'08': 'August',
'09': 'September',
'10': 'October',
'11': 'November',
'12': 'December'
}
periodnames = []
if crossyr:
for i in lags:
month_nr_str = i[0][i[0].find("-") + 1:i[0].find("-") + 1 +
2] #find first instace of "-" +2
periodnames.append(period_dict[month_nr_str])
else:
for i in lags:
month_nr_str = i[0][:2] #find first instace of "-" +2
periodnames.append(period_dict[month_nr_str])
for i in range(len(rg.list_for_MI)):
rg.list_for_MI[i].prec_labels['lag'] = ('lag', periodnames)
rg.list_for_MI[i].corr_xr['lag'] = ('lag', periodnames)
# View correlation regions
rg.quick_view_labels(mean=True, save=True)
rg.plot_maps_corr(save=True)
# Handle precursor regions
rg.get_ts_prec()
count = rg._df_count # how many times is each precursor regions found in the different training sets
print(count)
df_prec_regions = find_precursors.labels_to_df(
rg.list_for_MI[0].prec_labels)
df_prec_regions # center lat,lon coordinates and size (in number of gridcells)
return rg
def plot_ss2(agg_level, skillscores, col_wrap, metric=None):
#%%
import find_precursors
cluster_nc_path = get_list_of_name_path(agg_level, 1)[0][1]
ds = core_pp.import_ds_lazy(cluster_nc_path, format_lon='west_east')
cluster_labels_org = ds.coords['cluster']
ds = ds['xrclustered']
#create list of skill score names
skillscores_multi_idx = skillscores.index.levels
ss_list = []
for i in skillscores_multi_idx[1:][0]:
for j in skillscores_multi_idx[1:][1]:
ss_name = '{}_{}'.format(i, j)
ss_list.append(ss_name)
if metric is not None: #only apply single metric
ss_list = [metric]
#add dimensions and coordinates
xr_score = ds.copy()
xr_score.attrs = {}
list_xr = [xr_score.copy().expand_dims('metric', axis=0) for m in ss_list]
xr_score = xr.concat(list_xr, dim='metric')
xr_score['metric'] = ('metric', ss_list)
list_xr = [
xr_score.copy().expand_dims('target_month', axis=0)
for m in skillscores.columns
]
xr_score = xr.concat(list_xr, dim='target_month')
xr_score['target_month'] = ('target_month', skillscores.columns)
#replace labels with skillscores
for metric_nr, metric in enumerate(xr_score.metric.values):
test_or_train = metric[:metric.find("_")]
ss = metric[metric.find("_") + 1:]
for month_nr, month in enumerate(xr_score.target_month.values):
#slice over metric, month in skill score df
metric_cluster_dict = skillscores[month].xs(
(test_or_train, ss), level=(1, 2)).to_dict()
#replace cluster_labels with their skill score
cluster_labels_new = [
metric_cluster_dict.get(x, x)
for x in cluster_labels_org.values
]
#set all non replaced values of cluster labels to np.nan
cluster_labels_new = [
np.nan if isinstance(x, np.int32) else x
for x in cluster_labels_new
]
#replace values
xarr_labels_to_replace = ds
xr_score[month_nr,
metric_nr] = find_precursors.view_or_replace_labels(
xarr_labels_to_replace,
regions=list(cluster_labels_org.values),
replacement_labels=cluster_labels_new)
#set col wrap and subtitles
col_wrap = col_wrap #int determines nr of cols
import math
subtitles = [[] for i in range(
int(math.ceil(xr_score.target_month.values.size / col_wrap)))]
total_nr_fields = col_wrap * len(subtitles)
j = -1
for i, month in enumerate(xr_score.target_month.values):
if i % col_wrap == 0:
j += 1
subtitles[j].append('{}, {}'.format(month, metric))
if i == max(
list(enumerate(xr_score.target_month.values))
)[0] and total_nr_fields > xr_score.target_month.values.size:
for k in range(total_nr_fields -
xr_score.target_month.values.size):
subtitles[j].append('0')
#plot
fig = plot_maps.plot_corr_maps(xr_score,
col_dim='target_month',
row_dim='metric',
size=4,
clevels=np.arange(-.5, 0.51, .1),
cbar_vert=-0.1,
hspace=-0.2,
subtitles=subtitles,
col_wrap=col_wrap)
#%%
return fig
'fontsize': 18
}
}
save = True
# rg.plot_maps_corr(var='z500', save=save,
# min_detect_gc=min_detect_gc,
# kwrgs_plot=kwrgs_plot,
# append_str=''.join(map(str, z500_green_bb))+TV+str(cluster_label))
z500 = rg.list_for_MI[0]
xrvals, xrmask = RGCPD._get_sign_splits_masked(z500.corr_xr, min_detect_gc,
z500.corr_xr['mask'])
fig = plot_maps.plot_corr_maps(xrvals, xrmask, **kwrgs_plot)
rg.get_clust()
xrclustered = find_precursors.view_or_replace_labels(rg.ds['xrclustered'],
cluster_label)
fig.axes[0].contour(xrclustered.longitude,
xrclustered.latitude,
np.isnan(xrclustered),
transform=ccrs.PlateCarree(),
levels=[0, 2],
linewidths=1,
linestyles=['solid'],
colors=['white'])
filename = os.path.join(rg.path_outsub1,
'z500vsmx2t_' + rg.hash + '_' + str(cluster_label))
fig.savefig(filename + rg.figext, bbox_inches='tight')
#%% upon request of reviewer, using a smaller bounding box plot
# kwrgs_plot.update({'drawbox':[(0,0), (155,300,20,73)]})
