interpolate() - Code Metrics - Inspection of "Fixes" - Unidata/MetPy - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Pull Request — master (#456)

by Ryan

created 2017-06-13 18:38 UTC

interpolate() C

↳ Parent: Project

Complexity

Conditions

Size

Total Lines

Duplication

Lines	0
Ratio	0 %

Importance

Changes

Metric	Value
cc	7
dl	0
loc	94
rs	5.325
c	0
b	0
f	0

How to fix Long Method

# Copyright (c) 2008-2016 MetPy Developers.
# Distributed under the terms of the BSD 3-Clause License.
# SPDX-License-Identifier: BSD-3-Clause
"""Tools and calculations for assigning values to a grid."""

from __future__ import division

import numpy as np
from scipy.interpolate import griddata, Rbf
from scipy.spatial.distance import cdist

from . import interpolation, points
from ..package_tools import Exporter

exporter = Exporter(globals())


def calc_kappa(spacing, kappa_star=5.052):
    r"""Calculate the kappa parameter for barnes interpolation.

    Parameters
    ----------
    spacing: float
        Average spacing between observations
    kappa_star: float
        Non-dimensional response parameter. Default 5.052.

    Returns
    -------
        kappa: float

    """
    return kappa_star * (2.0 * spacing / np.pi)**2


def remove_observations_below_value(x, y, z, val=0):
    r"""Remove all x, y, and z where z is less than val.

    Will not destroy original values.

    Parameters
    ----------
    x: array_like
        x coordinate.
    y: array_like
        y coordinate.
    z: array_like
        Observation value.
    val: float
        Value at which to threshold z.

    Returns
    -------
    x, y, z
        List of coordinate observation pairs without
        observation values less than val.

    """
    x_ = x[z >= val]
    y_ = y[z >= val]
    z_ = z[z >= val]

    return x_, y_, z_


def remove_nan_observations(x, y, z):
    r"""Remove all x, y, and z where z is nan.

    Will not destroy original values.

    Parameters
    ----------
    x: array_like
        x coordinate
    y: array_like
        y coordinate
    z: array_like
        observation value

    Returns
    -------
    x, y, z
        List of coordinate observation pairs without
        nan valued observations.

    """
    x_ = x[~np.isnan(z)]
    y_ = y[~np.isnan(z)]
    z_ = z[~np.isnan(z)]

    return x_, y_, z_


def remove_repeat_coordinates(x, y, z):
    r"""Remove all x, y, and z where (x,y) is repeated and keep the first occurrence only.

    Will not destroy original values.

    Parameters
    ----------
    x: array_like
        x coordinate
    y: array_like
        y coordinate
    z: array_like
        observation value

    Returns
    -------
    x, y, z
        List of coordinate observation pairs without
        repeated coordinates.

    """
    coords = []
    variable = []

    for (x_, y_, t_) in zip(x, y, z):
        if (x_, y_) not in coords:
            coords.append((x_, y_))
            variable.append(t_)

    coords = np.array(coords)

    x_ = coords[:, 0]
    y_ = coords[:, 1]

    z_ = np.array(variable)

    return x_, y_, z_


@exporter.export
def interpolate(x, y, z, interp_type='linear', hres=50000,
                minimum_neighbors=3, gamma=0.25, kappa_star=5.052,
                search_radius=None, rbf_func='linear', rbf_smooth=0):
    r"""Interpolate given (x,y), observation (z) pairs to a grid based on given parameters.

    Parameters
    ----------
    x: array_like
        x coordinate
    y: array_like
        y coordinate
    z: array_like
        observation value
    interp_type: str
        What type of interpolation to use. Available options include:
        1) "linear", "nearest", "cubic", or "rbf" from Scipy.interpolate.
        2) "natural_neighbor", "barnes", or "cressman" from Metpy.mapping .
        Default "linear".
    hres: float
        The horizontal resolution of the generated grid. Default 50000 meters.
    minimum_neighbors: int
        Minimum number of neighbors needed to perform barnes or cressman interpolation for a
        point. Default is 3.
    gamma: float
        Adjustable smoothing parameter for the barnes interpolation. Default 0.25.
    kappa_star: float
        Response parameter for barnes interpolation, specified nondimensionally
        in terms of the Nyquist. Default 5.052
    search_radius: float
        A search radius to use for the barnes and cressman interpolation schemes.
        If search_radius is not specified, it will default to the average spacing of
        observations.
    rbf_func: str
        Specifies which function to use for Rbf interpolation.
        Options include: 'multiquadric', 'inverse', 'gaussian', 'linear', 'cubic',
        'quintic', and 'thin_plate'. Defualt 'linear'. See scipy.interpolate.Rbf for more
        information.
    rbf_smooth: float
        Smoothing value applied to rbf interpolation.  Higher values result in more smoothing.

    Returns
    -------
    grid_x: (N, 2) ndarray
        Meshgrid for the resulting interpolation in the x dimension
    grid_y: (N, 2) ndarray
        Meshgrid for the resulting interpolation in the y dimension ndarray
    img: (M, N) ndarray
        2-dimensional array representing the interpolated values for each grid.

    """
    grid_x, grid_y = points.generate_grid(hres, points.get_boundary_coords(x, y))

    if interp_type in ['linear', 'nearest', 'cubic']:
        points_zip = np.array(list(zip(x, y)))
        img = griddata(points_zip, z, (grid_x, grid_y), method=interp_type)

    elif interp_type == 'natural_neighbor':
        img = interpolation.natural_neighbor(x, y, z, grid_x, grid_y)

    elif interp_type in ['cressman', 'barnes']:
        ave_spacing = np.mean((cdist(list(zip(x, y)), list(zip(x, y)))))

        if search_radius is None:
            search_radius = ave_spacing

        if interp_type == 'cressman':
            img = interpolation.inverse_distance(x, y, z, grid_x, grid_y, search_radius,
                                                 min_neighbors=minimum_neighbors,
                                                 kind=interp_type)
        else:
            kappa = calc_kappa(ave_spacing, kappa_star)
            img = interpolation.inverse_distance(x, y, z, grid_x, grid_y, search_radius,
                                                 gamma, kappa, min_neighbors=minimum_neighbors,
                                                 kind=interp_type)

    elif interp_type == 'rbf':
        # 3-dimensional support not yet included.
        # Assign a zero to each z dimension for observations.
        h = np.zeros((len(x)))

        rbfi = Rbf(x, y, h, z, function=rbf_func, smooth=rbf_smooth)

        # 3-dimensional support not yet included.
        # Assign a zero to each z dimension grid cell position.
        hi = np.zeros(grid_x.shape)
        img = rbfi(grid_x, grid_y, hi)

    else:
        raise ValueError('Interpolation option not available. '
                         'Try: linear, nearest, cubic, natural_neighbor, '
                         'barnes, cressman, rbf')

    return grid_x, grid_y, img


1			# Copyright (c) 2008-2016 MetPy Developers.
2			# Distributed under the terms of the BSD 3-Clause License.
3			# SPDX-License-Identifier: BSD-3-Clause
4			"""Tools and calculations for assigning values to a grid."""
5
6			from __future__ import division
7
8			import numpy as np
9			from scipy.interpolate import griddata, Rbf
10			from scipy.spatial.distance import cdist
11
12			from . import interpolation, points
13			from ..package_tools import Exporter
14
15			exporter = Exporter(globals())
16
17
18			def calc_kappa(spacing, kappa_star=5.052):
19			r"""Calculate the kappa parameter for barnes interpolation.
20
21			Parameters
22			----------
23			spacing: float
24			Average spacing between observations
25			kappa_star: float
26			Non-dimensional response parameter. Default 5.052.
27
28			Returns
29			-------
30			kappa: float
31
32			"""
33			return kappa_star * (2.0 * spacing / np.pi)**2
34
35
36			def remove_observations_below_value(x, y, z, val=0):
37			r"""Remove all x, y, and z where z is less than val.
38
39			Will not destroy original values.
40
41			Parameters
42			----------
43			x: array_like
44			x coordinate.
45			y: array_like
46			y coordinate.
47			z: array_like
48			Observation value.
49			val: float
50			Value at which to threshold z.
51
52			Returns
53			-------
54			x, y, z
55			List of coordinate observation pairs without
56			observation values less than val.
57
58			"""
59			x_ = x[z >= val]
60			y_ = y[z >= val]
61			z_ = z[z >= val]
62
63			return x_, y_, z_
64
65
66			def remove_nan_observations(x, y, z):
67			r"""Remove all x, y, and z where z is nan.
68
69			Will not destroy original values.
70
71			Parameters
72			----------
73			x: array_like
74			x coordinate
75			y: array_like
76			y coordinate
77			z: array_like
78			observation value
79
80			Returns
81			-------
82			x, y, z
83			List of coordinate observation pairs without
84			nan valued observations.
85
86			"""
87			x_ = x[~np.isnan(z)]
88			y_ = y[~np.isnan(z)]
89			z_ = z[~np.isnan(z)]
90
91			return x_, y_, z_
92
93
94			def remove_repeat_coordinates(x, y, z):
95			r"""Remove all x, y, and z where (x,y) is repeated and keep the first occurrence only.
96
97			Will not destroy original values.
98
99			Parameters
100			----------
101			x: array_like
102			x coordinate
103			y: array_like
104			y coordinate
105			z: array_like
106			observation value
107
108			Returns
109			-------
110			x, y, z
111			List of coordinate observation pairs without
112			repeated coordinates.
113
114			"""
115			coords = []
116			variable = []
117
118			for (x_, y_, t_) in zip(x, y, z):
119			if (x_, y_) not in coords:
120			coords.append((x_, y_))
121			variable.append(t_)
122
123			coords = np.array(coords)
124
125			x_ = coords[:, 0]
126			y_ = coords[:, 1]
127
128			z_ = np.array(variable)
129
130			return x_, y_, z_
131
132
133			@exporter.export
134			def interpolate(x, y, z, interp_type='linear', hres=50000,
135			minimum_neighbors=3, gamma=0.25, kappa_star=5.052,
136			search_radius=None, rbf_func='linear', rbf_smooth=0):
137			r"""Interpolate given (x,y), observation (z) pairs to a grid based on given parameters.
138
139			Parameters
140			----------
141			x: array_like
142			x coordinate
143			y: array_like
144			y coordinate
145			z: array_like
146			observation value
147			interp_type: str
148			What type of interpolation to use. Available options include:
149			1) "linear", "nearest", "cubic", or "rbf" from Scipy.interpolate.
150			2) "natural_neighbor", "barnes", or "cressman" from Metpy.mapping .
151			Default "linear".
152			hres: float
153			The horizontal resolution of the generated grid. Default 50000 meters.
154			minimum_neighbors: int
155			Minimum number of neighbors needed to perform barnes or cressman interpolation for a
156			point. Default is 3.
157			gamma: float
158			Adjustable smoothing parameter for the barnes interpolation. Default 0.25.
159			kappa_star: float
160			Response parameter for barnes interpolation, specified nondimensionally
161			in terms of the Nyquist. Default 5.052
162			search_radius: float
163			A search radius to use for the barnes and cressman interpolation schemes.
164			If search_radius is not specified, it will default to the average spacing of
165			observations.
166			rbf_func: str
167			Specifies which function to use for Rbf interpolation.
168			Options include: 'multiquadric', 'inverse', 'gaussian', 'linear', 'cubic',
169			'quintic', and 'thin_plate'. Defualt 'linear'. See scipy.interpolate.Rbf for more
170			information.
171			rbf_smooth: float
172			Smoothing value applied to rbf interpolation. Higher values result in more smoothing.
173
174			Returns
175			-------
176			grid_x: (N, 2) ndarray
177			Meshgrid for the resulting interpolation in the x dimension
178			grid_y: (N, 2) ndarray
179			Meshgrid for the resulting interpolation in the y dimension ndarray
180			img: (M, N) ndarray
181			2-dimensional array representing the interpolated values for each grid.
182
183			"""
184			grid_x, grid_y = points.generate_grid(hres, points.get_boundary_coords(x, y))
185
186			if interp_type in ['linear', 'nearest', 'cubic']:
187			points_zip = np.array(list(zip(x, y)))
188			img = griddata(points_zip, z, (grid_x, grid_y), method=interp_type)
189
190			elif interp_type == 'natural_neighbor':
191			img = interpolation.natural_neighbor(x, y, z, grid_x, grid_y)
192
193			elif interp_type in ['cressman', 'barnes']:
194			ave_spacing = np.mean((cdist(list(zip(x, y)), list(zip(x, y)))))
195
196			if search_radius is None:
197			search_radius = ave_spacing
198
199			if interp_type == 'cressman':
200			img = interpolation.inverse_distance(x, y, z, grid_x, grid_y, search_radius,
201			min_neighbors=minimum_neighbors,
202			kind=interp_type)
203			else:
204			kappa = calc_kappa(ave_spacing, kappa_star)
205			img = interpolation.inverse_distance(x, y, z, grid_x, grid_y, search_radius,
206			gamma, kappa, min_neighbors=minimum_neighbors,
207			kind=interp_type)
208
209			elif interp_type == 'rbf':
210			# 3-dimensional support not yet included.
211			# Assign a zero to each z dimension for observations.
212			h = np.zeros((len(x)))
213
214			rbfi = Rbf(x, y, h, z, function=rbf_func, smooth=rbf_smooth)
215
216			# 3-dimensional support not yet included.
217			# Assign a zero to each z dimension grid cell position.
218			hi = np.zeros(grid_x.shape)
219			img = rbfi(grid_x, grid_y, hi)
220
221			else:
222			raise ValueError('Interpolation option not available. '
223			'Try: linear, nearest, cubic, natural_neighbor, '
224			'barnes, cressman, rbf')
225
226			return grid_x, grid_y, img
227

Unidata / MetPy

Pull Request — master (#456)

interpolate() C

Complexity

Size

Duplication

Importance

How to fix Long Method

Long Method

Duplication Side-by-Side

Filter issues like