forked from jajcayn/multi-scale
-
Notifications
You must be signed in to change notification settings - Fork 0
/
scale_nets-synchronization.py
185 lines (156 loc) · 7.64 KB
/
scale_nets-synchronization.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
from __future__ import print_function
from scale_network import ScaleSpecificNetwork
from datetime import date
from pathos.multiprocessing import Pool
import numpy as np
import pyclits as clt
import h5py
import matplotlib.pyplot as plt
plt.style.use('ipython')
def load_nino34_wavelet_phase(start_date, end_date, period, anom = False):
g = clt.data_loaders.load_enso_index('/Users/nikola/work-ui/data/nino34raw.txt', '3.4', start_date, end_date, anom=anom)
a = g.get_seasonality()
sg = clt.surrogates.SurrogateField()
sg.copy_field(g)
g.return_seasonality(a[0], a[1], a[2])
g.wavelet(period, period_unit="y", cut=2)
return g, g.phase.copy(), sg, a
def load_NAOindex_wavelet_phase(start_date, end_date, period, anom = False):
raw = np.loadtxt('/Users/nikola/work-ui/data/NAO.station.monthly.1865-2016.txt')
data = raw[:,1:].reshape((-1))
g = clt.geofield.DataField(data=np.array(data))
g.create_time_array(date_from=date(1865,1,1), sampling='m')
g.location = "NAO"
g.select_date(start_date, end_date)
if anom:
g.anomalise()
a = g.get_seasonality()
sg = clt.surrogates.SurrogateField()
sg.copy_field(g)
g.return_seasonality(a[0], a[1], a[2])
g.wavelet(period, period_unit="y", cut=2)
return g, g.phase.copy(), sg, a
def load_sunspot_number_phase(start_date, end_date, period, anom = False):
raw = np.loadtxt('/Users/nikola/work-ui/data/sunspot.monthly.1749-2017.txt')
data = []
time = []
for y in range(raw.shape[0]):
dat = float(raw[y, 3])
data.append(dat)
time.append(date(int(raw[y, 0]), int(raw[y, 1]), 1).toordinal())
g = clt.geofield.DataField(data=np.array(data), time=np.array(time))
g.location = "suspots"
g.select_date(start_date, end_date)
if anom:
g.anomalise()
a = g.get_seasonality()
sg = clt.surrogates.SurrogateField()
sg.copy_field(g)
g.return_seasonality(a[0], a[1], a[2])
g.wavelet(period, period_unit="y", cut=2)
return g, g.phase.copy(), sg, a
def load_pdo_phase(start_date, end_date, period, anom = False):
raw = np.loadtxt('/Users/nikola/work-ui/data/PDO.monthly.1900-2015.txt')
data = raw[:,1:].reshape((-1))
g = clt.geofield.DataField(data=np.array(data))
g.create_time_array(date_from=date(1900,1,1), sampling='m')
g.location = "PDO"
g.select_date(start_date, end_date)
if anom:
g.anomalise()
a = g.get_seasonality()
sg = clt.surrogates.SurrogateField()
sg.copy_field(g)
g.return_seasonality(a[0], a[1], a[2])
g.wavelet(period, period_unit="y", cut=2)
return g, g.phase.copy(), sg, a
def _compute_MI_synch(a):
ph, i, j, nao, nino, sunspots, pdo = a
# nao_s = clt.knn_mutual_information(ph, nao, k=128)
# nino_s = clt.knn_mutual_information(ph, nino, k=128)
# sunspot_s = clt.knn_mutual_information(ph, sunspots, k=128)
# pdo_s = clt.knn_mutual_information(ph, pdo, k=128)
nao_s = clt.mutual_information(ph, nao, algorithm='EQQ2', bins=8)
nino_s = clt.mutual_information(ph, nino, algorithm='EQQ2', bins=8)
sunspot_s = clt.mutual_information(ph, sunspots, algorithm='EQQ2', bins=8)
pdo_s = clt.mutual_information(ph, pdo, algorithm='EQQ2', bins=8)
return (i, j, nao_s, nino_s, sunspot_s, pdo_s)
WORKERS = 5
NUM_SURRS = 100
to_do_periods = np.arange(2,15.5,0.5)
net = ScaleSpecificNetwork('/Users/nikola/work-ui/data/NCEP/air.mon.mean.levels.nc',
'air', date(1950,1,1), date(2014,1,1), None, None,
level = 0, dataset="NCEP", sampling='monthly', anom=False)
synchronization = {}
for period in to_do_periods:
print("running for %.1f period..." % (period))
_, nao_ph, sg_nao, a_nao = load_NAOindex_wavelet_phase(date(1950,1,1), date(2014,1,1), period, anom=False)
_, nino_ph, sg_nino, a_nino = load_nino34_wavelet_phase(date(1950,1,1), date(2014,1,1), period, anom=False)
_, sunspots_ph, sg_sunspots, a_sunspots = load_sunspot_number_phase(date(1950,1,1), date(2014,1,1), period, anom=False)
_, pdo_ph, sg_pdo, a_pdo = load_pdo_phase(date(1950,1,1), date(2014,1,1), period, anom=False)
pool = Pool(WORKERS)
net.wavelet(period, period_unit='y', cut=2, pool=pool)
args = [(net.phase[:, i, j], i, j, nao_ph, nino_ph, sunspots_ph, pdo_ph) for i in range(net.lats.shape[0]) for j in range(net.lons.shape[0])]
result = pool.map(_compute_MI_synch, args)
synchs = np.zeros((4, net.lats.shape[0], net.lons.shape[0]))
synchs_surrs = np.zeros((NUM_SURRS, 4, net.lats.shape[0], net.lons.shape[0]))
for i, j, naos, ninos, suns, pdos in result:
synchs[0, i, j] = naos
synchs[1, i, j] = ninos
synchs[2, i, j] = suns
synchs[3, i, j] = pdos
for surr in range(NUM_SURRS):
sg_nao.construct_fourier_surrogates(algorithm='FT')
sg_nao.add_seasonality(a_nao[0], a_nao[1], a_nao[2])
sg_nao.wavelet(period, period_unit="y", cut=2)
sg_nino.construct_fourier_surrogates(algorithm='FT')
sg_nino.add_seasonality(a_nino[0], a_nino[1], a_nino[2])
sg_nino.wavelet(period, period_unit="y", cut=2)
sg_sunspots.construct_fourier_surrogates(algorithm='FT')
sg_sunspots.add_seasonality(a_sunspots[0], a_sunspots[1], a_sunspots[2])
sg_sunspots.wavelet(period, period_unit="y", cut=2)
sg_pdo.construct_fourier_surrogates(algorithm='FT')
sg_pdo.add_seasonality(a_pdo[0], a_pdo[1], a_pdo[2])
sg_pdo.wavelet(period, period_unit="y", cut=2)
args = [(net.phase[:, i, j], i, j, sg_nao.phase, sg_nino.phase, sg_sunspots.phase, sg_pdo.phase) for i in range(net.lats.shape[0]) for j in range(net.lons.shape[0])]
result = pool.map(_compute_MI_synch, args)
for i, j, naos, ninos, suns, pdos in result:
synchs_surrs[surr, 0, i, j] = naos
synchs_surrs[surr, 1, i, j] = ninos
synchs_surrs[surr, 2, i, j] = suns
synchs_surrs[surr, 3, i, j] = pdos
if (surr%20 == 0):
print("...%d/%d surrs done..." % (surr, NUM_SURRS))
pool.close()
pool.join()
synchronization[period] = synchs
synchronization["%.1f_surrs" % (period)] = synchs_surrs
hf = h5py.File('networks/phase_synch_eqq_bins=8_all_periods_%dFTsurrs.h5' % (NUM_SURRS))
for k in synchronization:
hf.create_dataset(str(k), data=synchronization[k])
hf.close()
## spectra
# nino, _ = load_nino34_wavelet_phase(date(1950,1,1), date(2014,1,1), 4, anom=False)
# suns, _ = load_sunspot_number_phase(date(1950,1,1), date(2014,1,1), 4, anom=False)
# pdo, _ = load_pdo_phase(date(1950,1,1), date(2014,1,1), 4, anom=False)
# scales = np.arange(2, 15*12+1)
# wvlt_power = np.zeros((4,scales.shape[0]))
# for sc in range(scales.shape[0]):
# nao.wavelet(scales[sc], period_unit='m', cut=24, save_wave=True)
# wvlt_power[0,sc] = np.sum(np.power(np.abs(nao.wave), 2)) / float(nao.wave.shape[0])
# nino.wavelet(scales[sc], period_unit='m', cut=24, save_wave=True)
# wvlt_power[1,sc] = np.sum(np.power(np.abs(nino.wave), 2)) / float(nino.wave.shape[0])
# suns.wavelet(scales[sc], period_unit='m', cut=24, save_wave=True)
# wvlt_power[2,sc] = np.sum(np.power(np.abs(suns.wave), 2)) / float(suns.wave.shape[0])
# pdo.wavelet(scales[sc], period_unit='m', cut=24, save_wave=True)
# wvlt_power[3,sc] = np.sum(np.power(np.abs(pdo.wave), 2)) / float(pdo.wave.shape[0])
# tits = ["NAO", "NINO3.4", "sunspot #", "PDO"]
# for j, tit in zip(range(4), tits):
# plt.subplot(2,2,j+1)
# plt.plot(scales, wvlt_power[j, :])
# plt.title(tit)
# plt.xticks(np.arange(10, scales.shape[0], 12), [i/12 for i in scales[10::12]], rotation = 30)
# # print(j, tit)
# if j > 1:
# plt.xlabel("scale [year]")
# plt.show()