-
Notifications
You must be signed in to change notification settings - Fork 0
/
probability_distribution_map_v2.py
227 lines (169 loc) · 9.13 KB
/
probability_distribution_map_v2.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
# coding=utf-8
import glob
from timeit import default_timer as timer
from typing import List
import dask.dataframe as dd
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import xarray as xr
from haversine import haversine
from matplotlib.colors import BoundaryNorm
from matplotlib.ticker import MaxNLocator
# from fast_histogram import histogram2d
import bathymetry
import common_tools_drift as ct
import config_plot
import config_sedimentdrift
from circle_of_distance import Circle_of_distance
class SedimentDistribution():
def __init__(self):
self.config = config_plot.ConfigPlot()
self.config_sedimentdrift = config_sedimentdrift.MartiniConf()
self.bath = bathymetry.Bathymetry(self.config)
self.bath.add_bathymetry_from_etopo1()
self.dx = None;
self.dy = None;
self.lon_bins = None;
self.lat_bins = None
def get_start(self, d, n):
# find index of the release event,
# first non masked element
arr = np.ma.masked_invalid(d.dif_depth[n].values)
if arr.count() == 0:
return None
return np.ma.flatnotmasked_edges(arr)[0]
def get_active_passive(self, status):
return ["active", "passive"][status]
def filter_data(self, df, filter_options: List):
np.warnings.filterwarnings('ignore')
# df = df[df.status == (filter_options["status"])]
if "diameter_max" and "diameter_min" in filter_options.keys() and not None:
df = df[(df.diameter >= float(filter_options["diameter_min"]))
& (df.diameter <= float(filter_options["diameter_max"]))]
# & (df.status == int(filter_options["status"]))]
if "selected_month" in filter_options.keys() and filter_options["selected_month"] is not None:
df = df[df.time.dt.month == int(filter_options["selected_month"])]
if "selected_day" in filter_options.keys() and not None:
df = df[df.time.dt.day == int(filter_options["selected_day"])]
return df
def return_ds(self, data):
return data
def get_indexes_of_last_valid_position(self, ds, var_name="density"):
# Assuming coordinates [time, trajectory]
print(ds)
masked_data = np.ma.masked_invalid(ds[var_name].values)
firstlast = np.ma.notmasked_edges(masked_data, axis=0)
return firstlast[1][0]
def extract_filtered_data(self, file_list: List, filter_options: List):
print("[Probability] Start filtering dataframe using dask")
start = timer()
# Get the data and group by trajectory
df = xr.open_mfdataset(file_list, concat_dim='trajectory', combine='nested').to_dataframe()
# Convert from multi-index (date - trajectory) to single index (date) to use Dask
df = df.reset_index(level="time")
# Convert to Dask dataframe with chunks
ddf = dd.from_pandas(df, 10)
# This one returns a Pandas dataframe
ddf = self.filter_data(ddf, filter_options).compute()
ddf = ddf.set_index(['time', ddf.index])
ds = ddf.groupby("trajectory").apply(self.return_ds).to_xarray()
end = timer()
print("[Probability] Finished filtering dataframe in {} seconds ".format(end - start))
indexes = self.get_indexes_of_last_valid_position(ds)
lons = np.ma.masked_invalid(ds.lon[indexes, :].values)
lats = np.ma.masked_invalid(ds.lat[indexes, :].values)
z = np.ma.masked_invalid(ds.z[indexes, :].values)
time = np.ma.masked_invalid(ds.time[indexes].values)
return np.array([np.array(xi) for xi in lats]), np.array([np.array(xi) for xi in lons]), np.array(
[np.array(xi) for xi in z]), np.array([np.array(xi) for xi in time])
def createBins(self):
print('func: createBins() => Creating bins for averaging')
dy = haversine((self.config.ymin, self.config.xmin), (self.config.ymax, self.config.xmin))
dx = haversine((self.config.ymin, self.config.xmin), (self.config.ymin, self.config.xmax))
print("Distance from minimum to maximim longitude binned area is %s km" % (dx))
print("Distance from minimum to maximim latitude binned area is %s km" % (dy))
nx = int(abs(dx / self.config.required_resolution))
ny = int(abs(dy / self.config.required_resolution))
# print("nx {} ny {}".format(nx, ny))
self.lon_bins = np.linspace(np.floor(self.config.xmin), np.ceil(self.config.xmax), nx, endpoint=True)
self.dx = len(self.lon_bins)
# print("DELTA X = {}".format(self.lon_bins))
self.lat_bins = np.linspace(np.floor(self.config.ymin), np.ceil(self.config.ymax), ny, endpoint=True)
# print("DELTA Y = {}".format(self.lat_bins))
self.dy = len(self.lat_bins)
print('=> created binned array of domain of grid cell size (%s,%s) with resolution %s' % (
self.dx, self.dy, self.config.required_resolution))
def get_density(self, lats, lons, nlevels):
density, xedges, yedges = np.histogram2d(lons.flatten(), lats.flatten(),
range=[[self.config.xmin, self.config.xmax],
[self.config.ymin, self.config.ymax]],
bins=[self.dx, self.dy])
total = np.sum(density)
density = (density / total) * 100.
print("Total number of points {} percentage sum {}".format(total, np.sum(density)))
density = np.ma.masked_where(density == 0, density)
levels = MaxNLocator(nbins=nlevels).tick_values(0.1, 2)
levels = np.arange(0.1, 4, 0.1)
cmap = plt.get_cmap('Spectral_r')
norm = BoundaryNorm(levels, ncolors=cmap.N, clip=True)
density = density.T
Xd, Yd = np.meshgrid(xedges, yedges)
return Xd, Yd, density
def create_distributional_map(self, filelist, filter_options):
lats, lons, depths, times = self.extract_filtered_data(filelist, filter_options)
start_date = pd.to_datetime(times[0])
end_date = pd.to_datetime(times[-1])
output_filename = ct.create_animation_or_png_filename(start_date, end_date, "clay", filter_options, "png")
self.createBins()
# confobj.mymap.drawparallels(confobj.lat_bins,linewidth=0.2, fmt='%g'+ 'E', fontsize=5, color='gray', labels=[True,False,False,False])
# confobj.mymap.drawmeridians(confobj.lon_bins,linewidth=0.2, fmt='%g'+ 'E', fontsize=5, color='gray',labels=[False,False,False,True])
nlevels = 10
Xd, Yd, density = self.get_density(lats, lons, nlevels)
levels = MaxNLocator(nbins=nlevels).tick_values(0, 20)
print(np.shape(Xd), np.shape(Yd), np.shape(density))
cmap = plt.get_cmap('PiYG')
# cplot = self.config.ax.contourf(Xd[:-1, :-1],
# Yd[:-1, :-1],
# density,
# cmap=cmap,
# levels=levels,
# edgecolors='face',
# linewidths=0.1)
cplot = self.config.ax.pcolormesh(Xd,
Yd,
density,
cmap=cmap,
edgecolors='face',
linewidths=0.1)
# cs = confobj.mymap.contourf(Xd[:-1, :-1],Yd[:-1, :-1], density, cmap=confobj.cmap,levels = 50) #,edgecolors='face',linewidths=0.1) #,alpha=0.8) #norm=norm,
# Seed area drawn as a circle
cs = Circle_of_distance()
X, Y = cs.create_circle_with_radius(self.config_sedimentdrift.st_lats[0],
self.config_sedimentdrift.st_lons[0],
self.config_sedimentdrift.release_radius/1000.)
self.config.ax.plot(self.config_sedimentdrift.st_lons[0],
self.config_sedimentdrift.st_lats[0],
marker='D',
color='r',
markersize=1)
plt.colorbar(cplot, shrink=0.7)
self.config.ax.plot(X, Y, marker=None, color='w', linewidth=1.0)
# plt.title("{} to {}".format(start_date, end_date))
print("Printing figure to file {} ".format(output_filename))
plt.savefig(output_filename, format='png', dpi=300)
# plt.show()
def main():
infilenames = glob.glob("output/*.nc")
# Loop over two size categories to se where particles end up depending on diameter
for diameter_min, diameter_max in zip([0, 100.e-6],[100.e-6, 350.e-6]):
filter_options = {"diameter_min": diameter_min,
"diameter_max": diameter_max,
"selected_month": None}
# "selected_day": 7,
# "status": status}
#for infile in infilenames:
distribution = SedimentDistribution()
distribution.create_distributional_map(infilenames, filter_options)
if __name__ == "__main__":
main()