#! /usr/bin/env python
"""Some utilities for the landlab package.
Landlab utilities
+++++++++++++++++
.. autosummary::
~landlab.core.utils.radians_to_degrees
~landlab.core.utils.as_id_array
~landlab.core.utils.make_optional_arg_into_id_array
~landlab.core.utils.get_functions_from_module
~landlab.core.utils.add_functions_to_class
~landlab.core.utils.add_module_functions_to_class
~landlab.core.utils.strip_grid_from_method_docstring
~landlab.core.utils.argsort_points_by_x_then_y
~landlab.core.utils.sort_points_by_x_then_y
~landlab.core.utils.anticlockwise_argsort_points
~landlab.core.utils.get_categories_from_grid_methods
"""
import errno
import importlib
import inspect
import os
import pathlib
import re
import shutil
import numpy as np
import pkg_resources
SIZEOF_INT = np.dtype(int).itemsize
[docs]
class ExampleData:
def __init__(self, example, case=""):
self._example = example
self._case = case
self._base = pathlib.Path(
pkg_resources.resource_filename(
"landlab", str(pathlib.Path("data").joinpath(example, case))
)
)
@property
def base(self):
return self._base
[docs]
def fetch(self):
"""Fetch landlab example data files.
Examples
--------
>>> data = ExampleData("io/shapefile")
>>> sorted(data)
['methow', 'redb', 'soque']
>>> import os
>>> data.fetch() # doctest: +SKIP
>>> sorted(os.listdir()) # doctest: +SKIP
['methow', 'redb', 'soque']
"""
dstdir, srcdir = pathlib.Path("."), self.base
for dst in (dstdir / p for p in self):
if dst.exists():
raise FileExistsError(
"[Errno {errno}] File exists: {name}".format(
errno=errno.EEXIST, name=repr(dst.name)
)
)
for src in (srcdir / p for p in self):
if src.is_file():
shutil.copy2(src, ".")
elif src.is_dir():
shutil.copytree(src, src.name)
def __iter__(self):
for p in self.base.iterdir():
yield p.name
def __truediv__(self, path):
return self.base / path
def __str__(self):
return str(self.base)
def __repr__(self):
return f"ExampleData({self._example!r}, case={self._case!r})"
[docs]
def degrees_to_radians(degrees):
"""Convert compass-style degrees to radians.
Convert angles in degrees measured clockwise starting from north to
angles measured counter-clockwise from the positive x-axis in radians
Parameters
----------
degrees : float or ndarray
Converted angles in degrees.
Returns
-------
rads : float or ndarray
Angles in radians.
Examples
--------
>>> import numpy as np
>>> from landlab.core.utils import degrees_to_radians
>>> degrees_to_radians(90.0)
0.0
>>> degrees_to_radians(0.0) == np.pi / 2.0
True
>>> degrees_to_radians(-180.0) == 3.0 * np.pi / 2.0
True
>>> np.testing.assert_array_almost_equal(
... [np.pi, np.pi], degrees_to_radians([-90.0, 270.0])
... )
"""
rads = np.pi * np.array(degrees) / 180.0
return (5.0 * np.pi / 2.0 - rads) % (2.0 * np.pi)
[docs]
def radians_to_degrees(rads):
"""Convert radians to compass-style degrees.
Convert angles (measured counter-clockwise from the positive x-axis) in
radians to angles in degrees measured clockwise starting from north.
Parameters
----------
rads : float or ndarray
Angles in radians.
Returns
-------
degrees : float or ndarray
Converted angles in degrees.
Examples
--------
>>> import numpy as np
>>> from landlab.core.utils import radians_to_degrees
>>> radians_to_degrees(0.0)
90.0
>>> radians_to_degrees(np.pi / 2.0)
0.0
>>> radians_to_degrees(-3 * np.pi / 2.0)
0.0
>>> radians_to_degrees(np.array([-np.pi, np.pi]))
array([ 270., 270.])
"""
degrees = (5.0 * np.pi / 2.0 - rads) % (2.0 * np.pi)
return 180.0 / np.pi * degrees
[docs]
def as_id_array(array):
"""Convert an array to an array of ids.
Parameters
----------
array : ndarray
Array of IDs.
Returns
-------
ndarray
A, possibly new, array of IDs.
Examples
--------
>>> import numpy as np
>>> from landlab.core.utils import as_id_array
>>> x = np.arange(5)
>>> y = as_id_array(x)
>>> y
array([0, 1, 2, 3, 4])
>>> x = np.arange(5, dtype=int)
>>> y = as_id_array(x)
>>> y
array([0, 1, 2, 3, 4])
>>> x = np.arange(5, dtype=np.int32)
>>> y = as_id_array(x)
>>> y
array([0, 1, 2, 3, 4])
>>> y.dtype == int
True
>>> x = np.arange(5, dtype=np.int64)
>>> y = as_id_array(x)
>>> y
array([0, 1, 2, 3, 4])
>>> y.dtype == int
True
>>> x = np.arange(5, dtype=np.intp)
>>> y = as_id_array(x)
>>> y
array([0, 1, 2, 3, 4])
>>> y.dtype == int
True
>>> x = np.arange(5, dtype=np.intp)
>>> y = np.where(x < 3)[0]
>>> y.dtype == np.intp
True
>>> as_id_array(y).dtype == int
True
"""
try:
if array.dtype == int:
return array.view(int)
else:
return array.astype(int)
except AttributeError:
return np.asarray(array, dtype=int)
[docs]
def make_optional_arg_into_id_array(number_of_elements, *args):
"""Transform an optional argument into an array of element ids.
Many landlab functions an optional argument of element ids that tells the
function to operate only on the elements provided. However, if the argument
is absent, all of the elements are to be operated on. This is a convenience
function that converts such an arguments list into an array of elements
ids.
Parameters
----------
number_of_elements : int
Number of elements in the grid.
array : array_like
Iterable to convert to an array.
Returns
-------
ndarray
Input array converted to a numpy array, or a newly-created numpy
array.
Examples
--------
>>> import numpy as np
>>> from landlab.core.utils import make_optional_arg_into_id_array
>>> make_optional_arg_into_id_array(4)
array([0, 1, 2, 3])
>>> make_optional_arg_into_id_array(4, [0, 0, 0, 0])
array([0, 0, 0, 0])
>>> make_optional_arg_into_id_array(4, (1, 1, 1, 1))
array([1, 1, 1, 1])
>>> make_optional_arg_into_id_array(4, np.ones(4))
array([1, 1, 1, 1])
>>> make_optional_arg_into_id_array(4, 0)
array([0])
>>> make_optional_arg_into_id_array(4, np.array([[1, 2], [3, 4]]))
array([1, 2, 3, 4])
"""
if len(args) == 0:
ids = np.arange(number_of_elements, dtype=int)
elif len(args) == 1:
ids = as_id_array(np.asarray(args[0])).reshape((-1,))
else:
raise ValueError("Number of arguments must be 0 or 1.")
return ids
[docs]
def get_functions_from_module(mod, pattern=None, exclude=None):
"""Get all the function in a module.
Parameters
----------
mod : module
An instance of a module.
pattern : str, optional
Only get functions whose name match a regular expression.
exclude : str, optional
Only get functions whose name exclude the regular expression.
*Note* if both pattern and exclude are provided both conditions must be met.
Returns
-------
dict
Dictionary of functions contained in the module. Keys are the
function names, values are the functions themselves.
"""
funcs = {}
for name, func in inspect.getmembers(mod, inspect.isroutine):
if (pattern is None or re.search(pattern, name)) and (
exclude is None or re.search(exclude, name) is None
):
funcs[name] = func
return funcs
[docs]
def add_functions_to_class(cls, funcs):
"""Add functions as methods of a class.
Parameters
----------
cls : class
A class.
funcs : dict
Dictionary of function names and instances.
"""
for name, func in funcs.items():
setattr(cls, name, func)
[docs]
def add_module_functions_to_class(cls, module, pattern=None, exclude=None):
"""Add functions from a module to a class as methods.
Parameters
----------
cls : class
A class.
module : module
An instance of a module.
pattern : str, optional
Only get functions whose name match a regular expression.
exclude : str, optional
Only get functions whose name exclude the regular expression.
*Note* if both pattern and exclude are provided both conditions must be met.
"""
(module, _) = os.path.splitext(os.path.basename(module))
mod = importlib.import_module("." + module, package="landlab.grid")
funcs = get_functions_from_module(mod, pattern=pattern, exclude=exclude)
strip_grid_from_method_docstring(funcs)
add_functions_to_class(cls, funcs)
[docs]
def strip_grid_from_method_docstring(funcs):
"""Remove 'grid' from the parameters of a dict of functions' docstrings.
Note that the docstring must be close to numpydoc standards for this to
work.
Parameters
----------
funcs : dict
Dictionary of functions to modify. Keys are the function names,
values are the functions themselves.
Examples
--------
>>> def dummy_func(grid, some_arg):
... '''A dummy function.
...
... Parameters
... ----------
... grid : ModelGrid
... A landlab grid.
... some_arg:
... An argument.
... '''
... pass
...
>>> funcs = {"dummy_func_to_demonstrate_docstring_modification": dummy_func}
>>> print(dummy_func.__doc__)
A dummy function.
<BLANKLINE>
Parameters
----------
grid : ModelGrid
A landlab grid.
some_arg:
An argument.
<BLANKLINE>
>>> strip_grid_from_method_docstring(funcs)
>>> print(dummy_func.__doc__)
A dummy function.
<BLANKLINE>
Parameters
----------
some_arg:
An argument.
<BLANKLINE>
"""
import re
for func in funcs.values():
# strip the entry under "Parameters":
func.__doc__ = re.sub("grid *:.*?\n.*?\n *", "", func.__doc__)
# # cosmetic magic to get a two-line signature to line up right:
match_2_lines = re.search(
func.__name__ + r"\(grid,[^\)]*?\n.*?\)", func.__doc__
)
try:
lines_were = match_2_lines.group()
except AttributeError: # no successful match
pass
else:
end_chars = re.search(r" .*?\)", lines_were).group()[4:]
lines_are_now = re.sub(r" .*?\)", " " + end_chars, lines_were)
func.__doc__ = (
func.__doc__[: match_2_lines.start()]
+ lines_are_now
+ func.__doc__[match_2_lines.end() :]
)
# Move "grid" in signature from an arg to the class position
func.__doc__ = re.sub(
func.__name__ + r"\(grid, ", "grid." + func.__name__ + "(", func.__doc__
)
[docs]
def argsort_points_by_x_then_y(points):
"""Sort points by coordinates, first x then y, returning sorted indices.
Parameters
----------
points : tuple of ndarray or ndarray of float, shape `(*, 2)`
Coordinates of points to be sorted. Sort by first coordinate, then
second.
Returns
-------
ndarray of int, shape `(n_points, )`
Indices of sorted points.
Examples
--------
>>> import numpy as np
>>> from landlab.core.utils import argsort_points_by_x_then_y
>>> points = np.zeros((10, 2))
>>> points[:, 0] = np.array([0.0, 0.0, 0.0, 1.0, 1.0, 1.0, 1.0, 2.0, 2.0, 2.0])
>>> points[:, 1] = np.array([0.0, 1.0, 2.0, -0.5, 0.5, 1.5, 2.5, 0.0, 1.0, 2.0])
>>> argsort_points_by_x_then_y(points)
array([3, 0, 7, 4, 1, 8, 5, 2, 9, 6])
>>> x = [0.0, 0.0, 0.0, 1.0, 1.0, 1.0, 1.0, 2.0, 2.0, 2.0]
>>> y = [0.0, 1.0, 2.0, -0.5, 0.5, 1.5, 2.5, 0.0, 1.0, 2.0]
>>> indices = argsort_points_by_x_then_y((x, y))
>>> indices
array([3, 0, 7, 4, 1, 8, 5, 2, 9, 6])
>>> argsort_points_by_x_then_y(np.array((x, y)))
array([3, 0, 7, 4, 1, 8, 5, 2, 9, 6])
"""
if isinstance(points, np.ndarray):
if points.shape[0] > points.shape[1]:
points = points.T
try:
return argsort_points_by_x_then_y((points[0, :], points[1, :]))
except IndexError:
return as_id_array([0])
else:
points = [np.asarray(coord) for coord in points]
a = points[0].argsort(kind="mergesort")
b = points[1][a].argsort(kind="mergesort")
return as_id_array(a[b])
[docs]
def sort_points_by_x_then_y(pts):
"""Sort points by coordinates, first x then y.
Parameters
----------
pts : Nx2 NumPy array of float
(x,y) points to be sorted
Returns
-------
pts : Nx2 NumPy array of float
sorted (x,y) points
Examples
--------
>>> import numpy as np
>>> from landlab.core.utils import sort_points_by_x_then_y
>>> pts = np.zeros((10, 2))
>>> pts[:, 0] = np.array([0.0, 0.0, 0.0, 1.0, 1.0, 1.0, 1.0, 2.0, 2.0, 2.0])
>>> pts[:, 1] = np.array([0.0, 1.0, 2.0, -0.5, 0.5, 1.5, 2.5, 0.0, 1.0, 2.0])
>>> pts = sort_points_by_x_then_y(pts)
>>> pts
array([[ 1. , -0.5],
[ 0. , 0. ],
[ 2. , 0. ],
[ 1. , 0.5],
[ 0. , 1. ],
[ 2. , 1. ],
[ 1. , 1.5],
[ 0. , 2. ],
[ 2. , 2. ],
[ 1. , 2.5]])
"""
indices = argsort_points_by_x_then_y(pts)
pts[:, 0] = pts[indices, 0]
pts[:, 1] = pts[indices, 1]
return pts
[docs]
def anticlockwise_argsort_points(pts, midpt=None):
"""Argort points into anticlockwise order around a supplied center.
Sorts CCW from east. Assumes a convex hull.
Parameters
----------
pts : Nx2 NumPy array of float
(x,y) points to be sorted
midpt : len-2 NumPy array of float (optional)
(x, y) of point about which to sort. If not provided, mean of pts is
used.
Returns
-------
pts : N NumPy array of int
sorted (x,y) points
Examples
--------
>>> import numpy as np
>>> from landlab.core.utils import anticlockwise_argsort_points
>>> pts = np.zeros((4, 2))
>>> pts[:, 0] = np.array([-3.0, -1.0, -1.0, -3.0])
>>> pts[:, 1] = np.array([-1.0, -3.0, -1.0, -3.0])
>>> sortorder = anticlockwise_argsort_points(pts)
>>> np.all(sortorder == np.array([2, 0, 3, 1]))
True
"""
if midpt is None:
midpt = pts.mean(axis=0)
assert len(midpt) == 2
theta = np.arctan2(pts[:, 1] - midpt[1], pts[:, 0] - midpt[0])
theta = theta % (2.0 * np.pi)
sortorder = np.argsort(theta)
return sortorder
[docs]
def anticlockwise_argsort_points_multiline(pts_x, pts_y, out=None):
"""Argort multi lines of points into CCW order around the geometric center.
This version sorts columns of data in a 2d array. Sorts CCW from east
around the geometric center of the points in the row.
Assumes a convex hull.
Parameters
----------
pts_x : rows x n_elements array of float
rows x points_to_sort x x_coord of points
pts_y : rows x n_elements array of float
rows x points_to_sort x y_coord of points
out : rows x n_elements (optional)
If provided, the ID array to be sorted
Returns
-------
sortorder : rows x n_elements NumPy array of int
sorted (x,y) points
Examples
--------
>>> import numpy as np
>>> from landlab.core.utils import anticlockwise_argsort_points_multiline
>>> pts = np.array([[1, 3, 0, 2], [2, 0, 3, 1]])
>>> pts_x = np.array([[-3.0, -1.0, -1.0, -3.0], [-3.0, -1.0, -1.0, -3.0]])
>>> pts_y = np.array([[-1.0, -3.0, -1.0, -3.0], [-3.0, -1.0, -3.0, -1.0]])
>>> sortorder = anticlockwise_argsort_points_multiline(pts_x, pts_y, out=pts)
>>> np.all(sortorder == np.array([[2, 0, 3, 1], [1, 3, 0, 2]]))
True
>>> np.all(pts == np.array([[0, 1, 2, 3], [0, 1, 2, 3]]))
True
"""
nrows = pts_x.shape[0]
midpt = np.empty((nrows, 2), dtype=float)
midpt[:, 0] = pts_x.mean(axis=1)
midpt[:, 1] = pts_y.mean(axis=1)
theta = np.arctan2(
pts_y - midpt[:, 1].reshape((nrows, 1)), pts_x - midpt[:, 0].reshape((nrows, 1))
)
theta = theta % (2.0 * np.pi)
sortorder = np.argsort(theta)
if out is not None:
out[:] = out[np.ogrid[:nrows].reshape((nrows, 1)), sortorder]
return sortorder
[docs]
def get_categories_from_grid_methods(grid_type):
"""Create a dict of category:[method_names] for a LL grid type.
Looks in the final line of the docstrings
of class methods and properties for a catgory declaration, "LLCATS: ".
It then creates and returns a dict with keys found as categories and
values that are lists of the names of methods that have that category.
Currently defined LLCATS are:
- DEPR : deprecated
- GINF : information about the grid as a whole
- NINF : information about nodes
- LINF : information about links
- PINF : information about patches
- CINF : information about cells
- FINF : information about faces
- CNINF : information about corners
- FIELDIO : methods to access and change fields
- FIELDADD : methods to create new fields/delete old fields
- FIELDINF : information about fields (keys, names, etc)
- GRAD : methods for gradients, fluxes, divergences and slopes
- MAP : methods to map from one element type to another
- BC : methods to interact with BCs
- SURF : methods for surface analysis (slope, aspect, hillshade)
- SUBSET : methods to indentify part of the grid based on conditions
- CONN : method describing the connectivity of one element to another
(i.e., 'links_at_node')
- MEAS : method describing a quantity defined on an element (i.e.,
'length_of_link')
- OTHER : anything else
Parameters
----------
grid_type : str
String of grid to inspect. Options are 'ModelGrid', 'RasterModelGrid',
'HexModelGrid', 'RadialModelGrid', 'VoronoiDelaunayGrid', or
'NetworkModelGrid'.
Returns
-------
cat_dict : dict
Dictionary with cats as keys and lists of method name strings as
values.
grid_dict : dict
Dictionary with method name strings as keys and lists of cats as
values.
FAILS : dict of dicts
contains any problematic LLCAT entries. Keys: 'MISSING' - list of names
of any public method or property without an LLCAT declared.
"""
import inspect
import re
from copy import copy
from landlab import FramedVoronoiGrid
from landlab import HexModelGrid
from landlab import ModelGrid
from landlab import NetworkModelGrid
from landlab import RadialModelGrid
from landlab import RasterModelGrid
from landlab import VoronoiDelaunayGrid
grid_str_to_grid = {
"ModelGrid": ModelGrid,
"RasterModelGrid": RasterModelGrid,
"HexModelGrid": HexModelGrid,
"RadialModelGrid": RadialModelGrid,
"VoronoiDelaunayGrid": VoronoiDelaunayGrid,
"NetworkModelGrid": NetworkModelGrid,
"FramedVoronoiGrid": FramedVoronoiGrid,
}
grid_dict = {}
cat_dict = {}
FAILS = {"MISSING": []}
grid = grid_str_to_grid[grid_type]
funcs = {}
for name, func in inspect.getmembers(grid):
funcs[name] = func
for method_name in funcs.keys():
if method_name[0] == "_":
continue
else:
method_doc = funcs[method_name].__doc__
try:
cat_str = re.search("LLCATS:.+", method_doc)
except TypeError:
pass
else:
if cat_str is None:
FAILS["MISSING"].append(method_name)
continue
cats = cat_str.group().split()[1:]
grid_dict[method_name] = copy(cats)
for cat in cats:
try:
cat_dict[cat].append(method_name)
except KeyError:
cat_dict[cat] = [method_name]
return cat_dict, grid_dict, FAILS
if __name__ == "__main__":
import doctest
doctest.testmod()