|
1
|
|
|
# Copyright (c) 2008-2015 MetPy Developers. |
|
2
|
|
|
# Distributed under the terms of the BSD 3-Clause License. |
|
3
|
|
|
# SPDX-License-Identifier: BSD-3-Clause |
|
4
|
|
|
"""Contains a collection of generally useful calculation tools.""" |
|
5
|
|
|
|
|
6
|
|
|
import numpy as np |
|
7
|
|
|
|
|
8
|
|
|
from ..package_tools import Exporter |
|
9
|
|
|
|
|
10
|
|
|
exporter = Exporter(globals()) |
|
11
|
|
|
|
|
12
|
|
|
|
|
13
|
|
|
@exporter.export |
|
14
|
|
|
def resample_nn_1d(a, centers): |
|
15
|
|
|
"""Return one-dimensional nearest-neighbor indexes based on user-specified centers. |
|
16
|
|
|
|
|
17
|
|
|
Parameters |
|
18
|
|
|
---------- |
|
19
|
|
|
a : array-like |
|
20
|
|
|
1-dimensional array of numeric values from which to |
|
21
|
|
|
extract indexes of nearest-neighbors |
|
22
|
|
|
centers : array-like |
|
23
|
|
|
1-dimensional array of numeric values representing a subset of values to approximate |
|
24
|
|
|
|
|
25
|
|
|
Returns |
|
26
|
|
|
------- |
|
27
|
|
|
An array of indexes representing values closest to given array values |
|
28
|
|
|
""" |
|
29
|
|
|
ix = [] |
|
30
|
|
|
for center in centers: |
|
31
|
|
|
index = (np.abs(a - center)).argmin() |
|
32
|
|
|
if index not in ix: |
|
33
|
|
|
ix.append(index) |
|
34
|
|
|
return ix |
|
35
|
|
|
|
|
36
|
|
|
|
|
37
|
|
|
@exporter.export |
|
38
|
|
|
def nearest_intersection_idx(a, b): |
|
39
|
|
|
"""Determine the index of the point just before two lines with common x values. |
|
40
|
|
|
|
|
41
|
|
|
Parameters |
|
42
|
|
|
---------- |
|
43
|
|
|
a : array-like |
|
44
|
|
|
1-dimensional array of y-values for line 1 |
|
45
|
|
|
b : array-like |
|
46
|
|
|
1-dimensional array of y-values for line 2 |
|
47
|
|
|
|
|
48
|
|
|
Returns |
|
49
|
|
|
------- |
|
50
|
|
|
An array of indexes representing the index of the values |
|
51
|
|
|
just before the intersection(s) of the two lines. |
|
52
|
|
|
""" |
|
53
|
|
|
# Difference in the two y-value sets |
|
54
|
|
|
difference = a - b |
|
55
|
|
|
|
|
56
|
|
|
# Determine the point just before the intersection of the lines |
|
57
|
|
|
# Will return multiple points for multiple intersections |
|
58
|
|
|
sign_change_idx, = np.nonzero(np.diff(np.sign(difference))) |
|
59
|
|
|
|
|
60
|
|
|
return sign_change_idx |
|
61
|
|
|
|
|
62
|
|
|
|
|
63
|
|
|
@exporter.export |
|
64
|
|
|
def find_intersections(x, a, b, direction='all'): |
|
65
|
|
|
"""Calculate the best estimate of intersection. |
|
66
|
|
|
|
|
67
|
|
|
Calculates the best estimates of the intersection of two y-value |
|
68
|
|
|
data sets that share a common x-value set. |
|
69
|
|
|
|
|
70
|
|
|
Parameters |
|
71
|
|
|
---------- |
|
72
|
|
|
x : array-like |
|
73
|
|
|
1-dimensional array of numeric x-values |
|
74
|
|
|
a : array-like |
|
75
|
|
|
1-dimensional array of y-values for line 1 |
|
76
|
|
|
b : array-like |
|
77
|
|
|
1-dimensional array of y-values for line 2 |
|
78
|
|
|
direction : string |
|
79
|
|
|
specifies direction of crossing. 'all', 'increasing' (a becoming greater than b), |
|
80
|
|
|
or 'decreasing' (b becoming greater than a). |
|
81
|
|
|
|
|
82
|
|
|
Returns |
|
83
|
|
|
------- |
|
84
|
|
|
A tuple (x, y) of array-like with the x and y coordinates of the |
|
85
|
|
|
intersections of the lines. |
|
86
|
|
|
""" |
|
87
|
|
|
# Find the index of the points just before the intersection(s) |
|
88
|
|
|
nearest_idx = nearest_intersection_idx(a, b) |
|
89
|
|
|
next_idx = nearest_idx + 1 |
|
90
|
|
|
|
|
91
|
|
|
# Determine the sign of the change |
|
92
|
|
|
sign_change = np.sign(a[next_idx] - b[next_idx]) |
|
93
|
|
|
|
|
94
|
|
|
# x-values around each intersection |
|
95
|
|
|
x0 = x[nearest_idx] |
|
96
|
|
|
x1 = x[next_idx] |
|
97
|
|
|
|
|
98
|
|
|
# y-values around each intersection for the first line |
|
99
|
|
|
a0 = a[nearest_idx] |
|
100
|
|
|
a1 = a[next_idx] |
|
101
|
|
|
|
|
102
|
|
|
# y-values around each intersection for the second line |
|
103
|
|
|
b0 = b[nearest_idx] |
|
104
|
|
|
b1 = b[next_idx] |
|
105
|
|
|
|
|
106
|
|
|
# Calculate the x-intersection. This comes from finding the equations of the two lines, |
|
107
|
|
|
# one through (x0, a0) and (x1, a1) and the other through (x0, b0) and (x1, b1), |
|
108
|
|
|
# finding their intersection, and reducing with a bunch of algebra. |
|
109
|
|
|
delta_y0 = a0 - b0 |
|
110
|
|
|
delta_y1 = a1 - b1 |
|
111
|
|
|
intersect_x = (delta_y1 * x0 - delta_y0 * x1) / (delta_y1 - delta_y0) |
|
112
|
|
|
|
|
113
|
|
|
# Calculate the y-intersection of the lines. Just plug the x above into the equation |
|
114
|
|
|
# for the line through the a points. One could solve for y like x above, but this |
|
115
|
|
|
# causes weirder unit behavior and seems a little less good numerically. |
|
116
|
|
|
intersect_y = ((intersect_x - x0) / (x1 - x0)) * (a1 - a0) + a0 |
|
117
|
|
|
|
|
118
|
|
|
# Make a mask based on the direction of sign change desired |
|
119
|
|
|
if direction == 'increasing': |
|
120
|
|
|
mask = sign_change > 0 |
|
121
|
|
|
elif direction == 'decreasing': |
|
122
|
|
|
mask = sign_change < 0 |
|
123
|
|
|
elif direction == 'all': |
|
124
|
|
|
return intersect_x, intersect_y |
|
125
|
|
|
else: |
|
126
|
|
|
raise ValueError('Unknown option for direction: {0}'.format(str(direction))) |
|
127
|
|
|
return intersect_x[mask], intersect_y[mask] |
|
128
|
|
|
|
|
129
|
|
|
|
|
130
|
|
|
@exporter.export |
|
131
|
|
|
def interpolate_nans(x, y, kind='linear'): |
|
132
|
|
|
"""Interpolate NaN values in y. |
|
133
|
|
|
|
|
134
|
|
|
Interpolate NaN values in the y dimension. Works with unsorted x values. |
|
135
|
|
|
|
|
136
|
|
|
Parameters |
|
137
|
|
|
---------- |
|
138
|
|
|
x : array-like |
|
139
|
|
|
1-dimensional array of numeric x-values |
|
140
|
|
|
y : array-like |
|
141
|
|
|
1-dimensional array of numeric y-values |
|
142
|
|
|
kind : string |
|
143
|
|
|
specifies the kind of interpolation x coordinate - 'linear' or 'log' |
|
144
|
|
|
|
|
145
|
|
|
Returns |
|
146
|
|
|
------- |
|
147
|
|
|
An array of the y coordinate data with NaN values interpolated. |
|
148
|
|
|
""" |
|
149
|
|
|
x_sort_args = np.argsort(x) |
|
150
|
|
|
x = x[x_sort_args] |
|
151
|
|
|
y = y[x_sort_args] |
|
152
|
|
|
nans = np.isnan(y) |
|
153
|
|
|
if kind is 'linear': |
|
154
|
|
|
y[nans] = np.interp(x[nans], x[~nans], y[~nans]) |
|
155
|
|
|
elif kind is 'log': |
|
156
|
|
|
y[nans] = np.interp(np.log(x[nans]), np.log(x[~nans]), y[~nans]) |
|
157
|
|
|
else: |
|
158
|
|
|
raise ValueError('Unknown option for kind: {0}'.format(str(kind))) |
|
159
|
|
|
return y[x_sort_args] |
|
160
|
|
|
|
Unidata /
MetPy
