-
Notifications
You must be signed in to change notification settings - Fork 1
/
figure_6.py
215 lines (205 loc) · 8.46 KB
/
figure_6.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
from collections import deque
from pathlib import Path
import json
import math
import numpy as np
import matplotlib
matplotlib.use("pgf")
matplotlib.rcParams.update(
{
"pgf.texsystem": "pdflatex",
"font.family": "serif",
"text.usetex": True,
"pgf.rcfonts": False,
}
)
import matplotlib.pyplot as plt
from matplotlib.colors import ListedColormap
# See https://jfly.uni-koeln.de/color/
CUD_CMAP = ListedColormap([
(0, 0, 0), # black
(0.9, 0.6, 0), # orange
(0.35, 0.7, 0.9), # sky blue
(0, 0.6, 0.5), # bluish green
(0.95, 0.9, 0.25), # yellow
(0, 0.45, 0.7), # blue
(0.8, 0.4, 0), # vermilion
(0.8, 0.6, 0.7) # reddish purple
], name="Color Universal Design")
DATA = {
0: [99.94],
30: [90.21, 85.55],
60: [80.74, 78.39, 77.92, 77.03],
90: [73.48],
120: [70.55, 68.38, 67.21, 66.67],
180: [40.37, 38.13, 37.53, 37.85, 41.82, 37],
}
PERSISTENCE = {
0: 100.00,
15: 90.86,
30: 81.38,
45: 68.67,
60: 56.11,
90: 38.08,
120: 27.73,
180: 17.10,
}
def shift_values(obj, shift):
dq = deque(obj)
dq.rotate(shift)
return np.array(list(dq))
if __name__ == '__main__':
import argparse
parser = argparse.ArgumentParser()
parser.add_argument("--aspect-ratio", default=3./5.)
parser.add_argument("--scale", default=1.0)
parser.add_argument("--markersize", default=1.3)
parser.add_argument("--dpi", default=300)
parser.add_argument("--legend-loc", default=9, type=int)
parser.add_argument("--format", nargs="+", default=["png"])
parser.add_argument("--name", default="untitled_figure")
parser.add_argument("-lt", "--lead-time", required=True, type=int)
parser.add_argument("--tag-name")
parser.add_argument("--styles", default=["science", "grid"])
parser.add_argument("--detailed", action="store_true")
parser.add_argument("--persistence")
parser.add_argument("--vis-per")
parser.add_argument("--wv-per")
parser.add_argument("--ir-per")
parser.add_argument("--json")
args = parser.parse_args()
plt.style.use(args.styles)
if args.tag_name is None:
args.tag_name = str(args.lead_time)
text_width = 5.45776
width = text_width * args.scale
height = width * args.aspect_ratio
fig, ax = plt.subplots(2 if args.detailed else 1, 1, sharex=True, figsize=(width, height * (2 if args.detailed else 1)), dpi=args.dpi)
if args.detailed:
fig.suptitle(f"{args.lead_time}min Performance")
fig.subplots_adjust(hspace=0)
(axes1, ax21) = ax
else:
axes1 = ax
axes12 = axes1.twinx()
axes12._get_lines.prop_cycler = axes1._get_lines.prop_cycler
axes12.grid(False)
vis_per_pt = Path(args.vis_per)
if not vis_per_pt.exists():
raise RuntimeError("VIS+PER file doesn't exists.")
with open(vis_per_pt, "r") as fh:
vis_per_js = json.load(fh)
wv_per_pt = Path(args.wv_per)
if not wv_per_pt.exists():
raise RuntimeError("WV+PER file doesn't exists.")
with open(wv_per_pt, "r") as fh:
wv_per_js = json.load(fh)
ir_per_pt = Path(args.ir_per)
if not ir_per_pt.exists():
raise RuntimeError("IR+PER file doesn't exists.")
with open(ir_per_pt, "r") as fh:
ir_per_js = json.load(fh)
if args.persistence:
json_per_pt = Path(args.persistence)
if not json_per_pt.exists():
raise RuntimeError("PER file doesn't exists.")
with open(json_per_pt, "r") as fh:
per_js = json.load(fh)
if args.json:
json_mod_pt = Path(args.json)
if not json_mod_pt.exists():
raise RuntimeError("PER file doesn't exists.")
with open(json_mod_pt, "r") as fh:
mod_js = json.load(fh)
sft = int(args.lead_time) // 15
lns = []
for js_tag, js, clr, mrk in (
("VIS+PER", vis_per_js, matplotlib.colors.to_hex(CUD_CMAP.colors[1]), "o"),
("WV+PER", wv_per_js, matplotlib.colors.to_hex(CUD_CMAP.colors[2]), "^"),
("IR+PER", ir_per_js, matplotlib.colors.to_hex(CUD_CMAP.colors[3]), "*"),
("Full Model", mod_js, matplotlib.colors.to_hex(CUD_CMAP.colors[5]), "s")
):
h_aggr = {(h, m): np.array(js[h][m]["csi"]) for h in sorted(js, key=int) for m in sorted(js[h], key=int)}
tp_aggr = {(h, m): np.array(js[h][m]["tp"]).sum(axis=-1) for h in sorted(js, key=int) for m in sorted(js[h], key=int)}
fp_aggr = {(h, m): np.array(js[h][m]["fp"]).sum(axis=-1) for h in sorted(js, key=int) for m in sorted(js[h], key=int)}
fn_aggr = {(h, m): np.array(js[h][m]["fn"]).sum(axis=-1) for h in sorted(js, key=int) for m in sorted(js[h], key=int)}
ticks = list(":".join(list(t)) for t in h_aggr)
the_mean = shift_values([h_aggr[t].mean() for t in h_aggr], sft)
the_std = shift_values([h_aggr[t].std() for t in h_aggr], sft)
tp_mean = shift_values([tp_aggr[t].mean() for t in tp_aggr], sft)
tp_std = shift_values([tp_aggr[t].std() for t in tp_aggr], sft)
fp_mean = shift_values([fp_aggr[t].mean() for t in fp_aggr], sft)
fp_std = shift_values([fp_aggr[t].std() for t in fp_aggr], sft)
fn_mean = shift_values([fn_aggr[t].mean() for t in fn_aggr], sft)
fn_std = shift_values([fn_aggr[t].std() for t in fn_aggr], sft)
mean_plt = axes1.plot(ticks, the_mean, label=f"{js_tag}", color=clr, marker=mrk, markersize=args.markersize)
axes1.fill_between(
ticks,
the_mean - the_std,
the_mean + the_std,
alpha=0.1,
edgecolor=mean_plt[0].get_color(),
facecolor=mean_plt[0].get_color(),
)
lns += mean_plt
per_plt, mean_per_plt = None, None
if args.persistence:
per_aggr = {(h, m): np.array(per_js[h][m]["csi"][0]) for h in sorted(js, key=int) for m in sorted(js[h], key=int)}
persistence = shift_values(per_aggr.values(), sft)
if persistence is not None:
per_plt = axes1.plot(ticks, persistence, label="Persistence", color=matplotlib.colors.to_hex(CUD_CMAP.colors[6]), marker="d", markersize=args.markersize)
linets = [np.array(js[h][m]["linets"])[0].mean() for h in sorted(js, key=int) for m in sorted(js[h], key=int)]
linets_std = [np.array(js[h][m]["linets"])[0].std() for h in sorted(js, key=int) for m in sorted(js[h], key=int)]
linets = shift_values(linets, sft)
linet_plt = axes12.plot(ticks, linets, label="LINET", linestyle=":", color=matplotlib.colors.to_hex(CUD_CMAP.colors[0]))
if args.detailed:
_h = next(iter(js))
_m = next(iter(js[_h]))
no_models = len(js[_h][_m]["csi"])
model_plts = []
for m_idx in range(no_models):
m_fp = {k: fp_aggr[k][m_idx] for k in fp_aggr}
m_fn = {k: fn_aggr[k][m_idx] for k in fn_aggr}
_ = ax21.plot(ticks, shift_values(m_fp.values(), sft), linestyle='dashed', label=f"FP{m_idx}")
model_plts.extend(_)
_ = ax21.plot(ticks, shift_values(m_fn.values(), sft), linestyle=(0, (3, 1, 1, 1)), color=_[0].get_color(), label=f"FN{m_idx}")
model_plts.extend(_)
ax21.set_ylim((0, 1.2 * np.nanmax([list(fn_aggr.values()), list(fp_aggr.values())])))
plt.xticks(ticks[::4], [str(l).split(':')[0] for l in ticks[::4]])
plt.xlim((ticks[0], ticks[-1]))
axes1.set_ylim((0, 100))
axes12.set_ylim((0, 1.03 * np.nanmax(linets)))
lns += linet_plt
if args.persistence and per_plt is not None:
lns += per_plt
legend1 = axes1.legend(
lns, [l.get_label() for l in lns],
fancybox=False,
edgecolor="black",
loc=args.legend_loc,
ncol=math.ceil(len(lns)/2),
bbox_to_anchor=(0.5, 1.2),
fontsize=8,
)
legend1.get_frame().set_linewidth(0.3)
axes1.set_ylabel("Critical Success Index (CSI) [\\%]")
axes12.set_ylabel("Avg. Lightning Events [\\#]")
if args.detailed:
legend2 = ax21.legend(
model_plts, [l.get_label() for l in model_plts],
fancybox=False,
edgecolor="black",
loc=9,
ncol=no_models,
fontsize=8,
)
legend2.get_frame().set_linewidth(0.3)
ax21.set_xlabel("Daytime")
axes1.set_title('Comparison with Other Methods')
ax21.set_title('Detailed Model Performance(s)')
ax21.set_ylabel("Events in Samples [\\#]")
else:
axes1.set_xlabel("Daytime")
plt.tight_layout()
for frmt in args.format:
plt.savefig(f"{args.name}.{frmt}")