-
Notifications
You must be signed in to change notification settings - Fork 0
/
lib_dasilva2024.py
345 lines (282 loc) · 12.4 KB
/
lib_dasilva2024.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
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
"""Implements an unpublished double dispersion automatic identification
method developed by da Silva.
"""
from datetime import timedelta
from intervaltree import IntervalTree
import numpy as np
import pandas as pd
from scipy.signal import find_peaks, savgol_filter
from progressbar import ProgressBar
DEFAULT_INTEGRAL_THRESHOLD = 0.1 # Default threshold for scores
INTERVAL_LENGTH = 60 # seconds
MIN_IFLUX_AT_PEAK = 10**(6.5) # Minimum ion flux at peak energy
MAX_PEAK_ENERGY = 10**(3.5) # Anayze energies under this level
MIN_MLAT = 50 # Minimum magnetic latitude (degrees)
EP_SMOOTH_WINDOW_SIZE = 11
OMNIWEB_FILL_VALUE = 9999 # fill value for msising omniweb data
def walk_and_integrate(dmsp_flux_fh, omniweb_fh, reverse_effect, threshold):
"""Walk through windows in the file, and estimate score for each window.
Args
dmsp_flux_fh: file handle returned by read_files()
omniweb_fh: file handle returned by read_files()
reverse_effect: Search for effects in the opposite direction with a
magnetic field set to the opposite of the coded threshold.
threshold: scalar score threshold
Returns
df_match: Pandas dataframe holding results in each row
"""
# Calculate coefficients -------------------------------------------------
mlt_coeff = (dmsp_flux_fh['mlt'] > 6) & (dmsp_flux_fh['mlt'] <= 18)
mlt_coeff = mlt_coeff.astype(float)
mlat_coeff = np.abs(dmsp_flux_fh['mlat']) > MIN_MLAT
mlat_coeff = mlat_coeff.astype(float)
mlat_dir_coeff = np.zeros(dmsp_flux_fh['t'].size)
mlat_dir_coeff[:-1] = -np.sign(np.diff(np.abs(dmsp_flux_fh['mlat'])))
mlat_dir_coeff[-1] = mlat_dir_coeff[-2]
if reverse_effect:
mlat_dir_coeff *= -1
# Iterate through time
# ------------------------------------------------------------------------
matching_intervals = IntervalTree()
bar = ProgressBar()
prog_bar_iter = bar(list(enumerate(dmsp_flux_fh['t'][:-1])))
for start_time_idx, start_time in prog_bar_iter:
# Set end time index ------------------------------------------------
end_time = start_time + timedelta(seconds=INTERVAL_LENGTH)
end_time_idx = dmsp_flux_fh['t'].searchsorted(end_time)
if end_time_idx > dmsp_flux_fh['t'].size:
end_time_idx = dmsp_flux_fh['t'].size
# Skip condition -----------------------------------------------------
i, j = start_time_idx, end_time_idx
if j - i < 8:
continue
if np.all(mlt_coeff[i:j] * mlt_coeff[i:j] == 0):
continue
# Do peakfinding -----------------------------------------------------
t, lower_Ep, upper_Ep = calculate_dual_Ep(
dmsp_flux_fh, start_i=i, stop_j=j
)
# Calculate Derivative of lower_Ep and upper_Ep ----------------------
# Convert to log space
lower_Ep = np.log10(lower_Ep)
upper_Ep = np.log10(upper_Ep)
# smooth
mask = np.ones(t.size, dtype=bool)
lower_Ep_smooth = clean_Ep(lower_Ep, mask, EP_SMOOTH_WINDOW_SIZE)
upper_Ep_smooth = clean_Ep(upper_Ep, mask, EP_SMOOTH_WINDOW_SIZE)
# Calculate difference in times (should be uniformly spaced)
dt = [delta.total_seconds() for delta in np.diff(t)]
dt.append(dt[-1]) # copy last value to retain shape
dt = np.array(dt)
# Calculate difference in Ep's
lower_Ep_deriv = np.zeros(t.size)
upper_Ep_deriv = np.zeros(t.size)
lower_Ep_deriv[:-1] = np.diff(lower_Ep_smooth) / dt[:-1]
upper_Ep_deriv[:-1] = np.diff(upper_Ep_smooth) / dt[:-1]
lower_Ep_deriv[-1] = lower_Ep_deriv[-2]
upper_Ep_deriv[-1] = upper_Ep_deriv[-2]
# Set derivative = 0 at nulls
lower_Ep_deriv[np.isnan(lower_Ep_deriv)] = 0
upper_Ep_deriv[np.isnan(upper_Ep_deriv)] = 0
# Look up magnetic field value at center of the interval -------------
B_test_time = start_time + timedelta(seconds=INTERVAL_LENGTH/2)
B_i = omniweb_fh['t'].searchsorted(B_test_time)
if B_i == omniweb_fh['Bz'].size:
continue
Bx = omniweb_fh['Bx'][B_i]
By = omniweb_fh['By'][B_i]
Bz = omniweb_fh['Bz'][B_i]
if Bz > OMNIWEB_FILL_VALUE:
continue
#elif reverse_effect and Bz < 0:
# continue
elif not reverse_effect and Bz > 0:
continue
# Calculate integrand ----------------------------------------------------
lower_integrand = (
mlt_coeff[i:j] *
mlat_coeff[i:j] *
mlat_dir_coeff[i:j] *
lower_Ep_deriv
)
upper_integrand = (
mlt_coeff[i:j] *
mlat_coeff[i:j] *
mlat_dir_coeff[i:j] *
upper_Ep_deriv
)
# Calculate total score for upper and lower curves
lower_score = np.sum(lower_integrand * dt)
upper_score = np.sum(upper_integrand * dt)
if lower_score > threshold and upper_score > threshold:
matching_intervals[start_time:end_time] = {
'Bx': Bx,
'By': By,
'Bz': Bz,
't': t,
'lower_integrand': lower_integrand,
'upper_integrand': upper_integrand,
'lower_Ep': lower_Ep,
'upper_Ep': upper_Ep,
}
# Merge overlapping intervals into common intervals, retaining context
# associated with each.
# ------------------------------------------------------------------------
def reducer(current, new_data):
for key in new_data:
if key not in current:
current[key] = []
current[key].append(new_data[key])
return current
matching_intervals.merge_overlaps(
data_reducer=reducer, data_initializer={}
)
# Convert to pandas dataframe for easy output to terminal of matching
# intervals and associated metadata.
# ------------------------------------------------------------------------
df_match_rows = []
for interval in sorted(matching_intervals):
df_match_rows.append([
interval.begin, interval.end,
np.mean(list(interval.data['Bx'])),
np.mean(list(interval.data['By'])),
np.mean(list(interval.data['Bz'])),
interval.data['t'][0],
interval.data['lower_integrand'][0],
interval.data['upper_integrand'][0],
interval.data['lower_Ep'][0],
interval.data['upper_Ep'][0],
])
df_match = pd.DataFrame(df_match_rows, columns=[
'start_time', 'end_time', 'Bx_mean', 'By_mean', 'Bz_mean',
't', 'lower_integrand', 'upper_integrand', 'lower_Ep', 'upper_Ep',
])
return df_match
def calculate_dual_Ep(dmsp_flux_fh, start_i=None, stop_j=None, _cache={}):
"""Calculate two time series associated with lower and upper peak
energies.
Args
dmsp_flux_fh: Dicitonary mapping variable names to arrays, associated with
DMSP precipitating flux file.
start_i: Start index of the file to analyze, defaults to start of file
stop_j: Stop index (exclusive) to analyze, defaults to end of file
Returns
t: time axis associated with other return values, array of datetimes
lower_Ep: Lower curve of peak energies
upper_Ep: Upper curve of peak energies
"""
# Iinitialize default keyword arguments to full size of the time series
# available
# ------------------------------------------------------------------------
if start_i is None:
start_i = 0
if stop_j is None:
stop_j = dmsp_flux_fh['t'].size
# Initialize vairables that will be used during iteration
# ------------------------------------------------------------------------
t = dmsp_flux_fh['t']
E = dmsp_flux_fh['ch_energy']
lower_Ep = np.zeros(stop_j - start_i, dtype=float)
upper_Ep = np.zeros(stop_j - start_i, dtype=float)
lower_Flux = np.zeros(stop_j - start_i, dtype=float)
upper_Flux = np.zeros(stop_j - start_i, dtype=float)
last_filled = 'lower'
# Iterate through the timesteps. Performs peakfinding and branches logic
# based on number of peaks.
# ------------------------------------------------------------------------
for ii, i in enumerate(range(start_i, stop_j)):
data = dmsp_flux_fh['ion_d_ener'][:, i]
key = (id(dmsp_flux_fh['ion_d_ener']), i)
if key in _cache:
smoothed = _cache[key]
else:
smoothed = savgol_filter(data, 5, 2)
_cache[key] = smoothed
#smoothed = savgol_filter(data, 5, 2)
#smoothed = np.convolve(data, kernel, mode='same')
#smoothed = data
peaks = find_peaks(smoothed)[0]
peaks = np.array([peak for peak in peaks if
data[peak] > MIN_IFLUX_AT_PEAK and
E[peak] < MAX_PEAK_ENERGY])
if len(peaks) == 0:
lower_Ep[ii] = np.nan
lower_Flux[ii] = np.nan
upper_Ep[ii] = np.nan
upper_Flux[ii] = np.nan
last_filled = 'none'
elif len(peaks) == 1:
Enext = E[peaks[0]]
Fluxnext = data[peaks[0]]
cond0 = (ii == 0)
cond1 = (np.isnan(lower_Ep[ii-1]) and np.isnan(upper_Ep[ii-1]))
cond2 = (np.isfinite(lower_Ep[ii-1]) and lower_Ep[ii-1] >= Enext)
cond3 = (np.isnan(upper_Ep[ii-1]))
cond4 = (last_filled in ('none', 'lower', 'both'))
if ((cond0 or cond1 or cond2 or cond3) and cond4):
lower_Ep[ii] = Enext
lower_Flux[ii] = Fluxnext
upper_Ep[ii] = np.nan
upper_Flux[ii] = np.nan
last_filled = 'lower'
else:
lower_Ep[ii] = np.nan
lower_Flux[ii] = np.nan
upper_Ep[ii] = Enext
upper_Flux[ii] = Fluxnext
last_filled = 'upper'
elif len(peaks) >= 2:
# Pick peaks with top-2 fluxes
I = np.argsort(smoothed[peaks])[-2:]
low_E, hi_E = sorted(E[peaks[I]])
low_Flux, hi_Flux = data[peaks][np.argsort(E[peaks[I]])]
# if high is closer to last lower line, then probably a false
# signal
if np.abs(hi_E - lower_Ep[ii-1]) < np.abs(low_E - lower_Ep[ii-1]):
lower_Ep[ii] = hi_E
upper_Ep[ii] = np.nan
last_filled = 'lower'
else:
lower_Ep[ii], upper_Ep[ii] = low_E, hi_E
lower_Flux[ii], upper_Flux[ii] = low_Flux, hi_Flux
last_filled = 'both'
#lower_Ep = np.log10(lower_Ep)
#upper_Ep = np.log10(upper_Ep)
# Filter out lone points -------------------------------------------------
# for i in range(1, lower_Ep.size - 1):
# isfin1 = np.isfinite(lower_Ep[i-1:i+2])
# isfin2 = np.isfinite(upper_Ep[i-1:i+2])
# if np.all(isfin1 == [False, True, False]):
# lower_Ep[i] = np.nan
# if np.all(isfin2 == [False, True, False]):
# upper_Ep[i] = np.nan
# Return
return t[start_i:stop_j], lower_Ep, upper_Ep
#@jit(nopython=True)
def clean_Ep(Ep, keep_mask, window_size):
"""Smooth Ep with a mask of points to include in moving average.
Arguments
Ep: floating point numpy array
keep_mask: points to include in moving average
window_size: integer, must be odd
Returns
Smoothed Ep array
"""
assert (window_size is None) or (window_size % 2 == 1),\
'Window size must be odd'
Ep_clean = Ep.copy()
for i in range(Ep.size):
if not keep_mask[i]:
Ep_clean[i] = np.nan
elif (window_size is None) and keep_mask[i]:
Ep_clean[i] = Ep[i]
elif (window_size is not None):
total = 0.0
count = 0
for di in range(-window_size//2, window_size//2 + 1):
if i + di > 0 and i + di < Ep.size and keep_mask[i + di]:
total += Ep[i + di]
count += 1
if count > 0: # else left as nan
Ep_clean[i] = total / count
return Ep_clean