def test_grid_by_width(self):
"""
Test error-free creation and return of a valid grid, with dimensions
provided as cell widths.
"""
grid = GridDimensions((180, 360), "width")
dims_by_width = grid.dims_by_width()
dims_by_width_expected = (180, 360)
self.assertEqual(dims_by_width, dims_by_width_expected)
def test_grid_count_to_width(self):
"""
Test error-free creation and return of a valid grid with dimensions
provided as cell counts, and the correct conversion of the grid to
a representation in cell widths.
"""
grid = GridDimensions((45, 180), "count")
dims_by_width = grid.dims_by_width()
dims_by_width_expected = (4, 2)
self.assertEqual(dims_by_width, dims_by_width_expected)
def test_tiny_grid_cells(self):
"""
Test correct conversion between grid representations when grid cell
widths are very small.
"""
grid = GridDimensions((2880, 2880), "count")
dims_by_width = grid.dims_by_width()
dims_by_width_expected = (0.0625, 0.125)
dims_by_count = grid.dims_by_count()
dims_by_count_expected = (2880, 2880)
self.assertEqual(dims_by_width, dims_by_width_expected)
self.assertEqual(dims_by_count, dims_by_count_expected)
def test_max_grid_dimensions(self):
"""
Test that a grid is successfully created when the maximum valid cell
widths are provided.
"""
grid = GridDimensions((180, 360), "width")
dims_by_width = grid.dims_by_width()
dims_by_width_expected = (180, 360)
dims_by_count = grid.dims_by_count()
dims_by_count_expected = (1, 1)
self.assertEqual(dims_by_width, dims_by_width_expected)
self.assertEqual(dims_by_count, dims_by_count_expected)
def test_float_grid_dims(self):
"""
Test the error-free creation of a valid grid with floating point
grid cell widths, and sensible conversion to a representation in
cell counts.
"""
grid = GridDimensions((0.5, 1.5), "width")
dims_by_width = grid.dims_by_width()
dims_by_width_expected = (0.5, 1.5)
dims_by_count = grid.dims_by_count()
dims_by_count_expected = (360, 240)
self.assertEqual(dims_by_width, dims_by_width_expected)
self.assertEqual(dims_by_count, dims_by_count_expected)
class ModelImageRenderer:
"""
A converter between gridded climate data and image visualizations of the
data. Displays data in the form of nested lists or arrays as an overlay
on a world map projection.
"""
def __init__(self: 'ModelImageRenderer', data: np.ndarray) -> None:
"""
Instantiate a new ModelImageReader.
Data to display is provided through the data parameter. This parameter
may either by an array, or an array-like structure such as a nested
list. the There must be three dimensions to the data: time first,
followed by latitude and longitude.
:param data:
An array-like structure of numeric values, representing temperature
over a globe
"""
self._data = data
self._grid = GridDimensions((len(data), len(data[0])), "count")
# Some parameters for the visualization are also left as attributes.
self._continent_linewidth = 0.5
self._lat_long_linewidth = 0.1
def save_image(
self: 'ModelImageRenderer',
out_path: str,
min_max_grades: Tuple[int, int] = (-60, 60)) -> None:
"""
Produces a .PNG formatted image file containing the gridded data
overlaid on a map projection.
The map is displayed in equirectangular projection, labelled with
lines of latitude and longitude at the intersection between
neighboring grid cells. The grid cells are filled in with a colour
denoting the magnitude of the temperature at that cell.
The optional min_max_grades parameter specifies the lower bound and
the upper bound of the range of values for which different colours
are assigned. Any grid cell with a temperature value below the first
element of min_max_grades will be assigned the same colour. The same
applies to any cell with a temperature greater than min_max_grades[1].
The default boundaries are -60 and 40.
The image is saved at the specified absolute or relative filepath.
:param out_path:
The location where the image file is created
:param min_max_grades:
A tuple containing the boundary values for the colorbar shown in
the image file
"""
if len(min_max_grades) != 2:
raise ValueError("Color grade boundaries must be given in a tuple"
"of length 2 (is length {})".format(
len(min_max_grades)))
# Create an empty world map in equirectangular projection.
map = Basemap(llcrnrlat=-90,
llcrnrlon=-180,
urcrnrlat=90,
urcrnrlon=180)
map.drawcoastlines(linewidth=self._continent_linewidth)
# Construct a grid from the horizontal and vertical sizes of the cells.
grid_by_width = self._grid.dims_by_width()
lat_val = -90.0
lats = []
while lat_val <= 90:
lats.append(lat_val)
lat_val += grid_by_width[0]
lon_val = -180.0
lons = []
while lon_val <= 180:
lons.append(lon_val)
lon_val += grid_by_width[1]
map.drawparallels(lats, linewidth=self._lat_long_linewidth)
map.drawmeridians(lons, linewidth=self._lat_long_linewidth)
x, y = map(lons, lats)
# Overlap the gridded data on top of the map, and display a colour
# legend with the appropriate boundaries.
img = map.pcolormesh(x, y, self._data, cmap=plt.cm.get_cmap("jet"))
map.colorbar(img)
plt.clim(min_max_grades[0], min_max_grades[1])
# Save the image and clear added components from memory
plt.savefig(out_path)
plt.close()
class ModelImageRenderer:
"""
A converter between gridded climate data and image visualizations of the
data. Displays data in the form of nested lists or arrays as an overlay
on a world map projection.
"""
def __init__(self: 'ModelImageRenderer',
data: np.ndarray) -> None:
"""
Instantiate a new ModelImageReader.
Data to display is provided through the data parameter. This parameter
may either by an array, or an array-like structure such as a nested
list. the There must be three dimensions to the data: time first,
followed by latitude and longitude.
:param data:
An array-like structure of numeric values, representing temperature
over a globe
"""
self._data = data
self._grid = GridDimensions((len(data), len(data[0])), "count")
# Some parameters for the visualization are also left as attributes.
self._continent_linewidth = 0.5
self._lat_long_linewidth = 0.1
def save_image(self: 'ModelImageRenderer',
out_path: str,
min_max_grades: Tuple[float, float] = (-8, 8)) -> None:
"""
Produces a .PNG formatted image file containing the gridded data
overlaid on a map projection.
The map is displayed in equirectangular projection, labelled with
lines of latitude and longitude at the intersection between
neighboring grid cells. The grid cells are filled in with a colour
denoting the magnitude of the temperature at that cell.
The optional min_max_grades parameter specifies the lower bound and
the upper bound of the range of values for which different colours
are assigned. Any grid cell with a temperature value below the first
element of min_max_grades will be assigned the same colour. The same
applies to any cell with a temperature greater than min_max_grades[1].
The default boundaries are -60 and 40.
The image is saved at the specified absolute or relative filepath.
:param out_path:
The location where the image file is created
:param min_max_grades:
A tuple containing the boundary values for the colorbar shown in
the image file
"""
if min_max_grades is not None and len(min_max_grades) != 2:
raise ValueError("Color grade boundaries must be given in a tuple"
"of length 2 (is length {})"
.format(len(min_max_grades)))
# Create an empty world map in equirectangular projection.
map = Basemap(llcrnrlat=-90, llcrnrlon=-180,
urcrnrlat=90, urcrnrlon=180)
map.drawcoastlines(linewidth=self._continent_linewidth)
# Construct a grid from the horizontal and vertical sizes of the cells.
grid_by_width = self._grid.dims_by_width()
lat_val = -90.0
lats = []
while lat_val <= 90:
lats.append(lat_val)
lat_val += grid_by_width[0]
lon_val = -180.0
lons = []
while lon_val <= 180:
lons.append(lon_val)
lon_val += grid_by_width[1]
map.drawparallels(lats, linewidth=self._lat_long_linewidth)
map.drawmeridians(lons, linewidth=self._lat_long_linewidth)
x, y = map(lons, lats)
# Overlap the gridded data on top of the map, and display a colour
# legend with the appropriate boundaries.
img_bin = map.pcolormesh(x, y, self._data, cmap=plt.cm.get_cmap("jet"))
map.colorbar(img_bin)
plt.clim(min_max_grades[0], min_max_grades[1])
fig = plt.gcf()
fig.canvas.draw()
pixels = fig.canvas.tostring_rgb()
img = np.fromstring(pixels, dtype=np.uint8, sep='')
img = img.reshape(fig.canvas.get_width_height()[::-1] + (3,))
img = img[118:-113, 80:-30, :]
alphas = np.ones(img.shape[:2], dtype=np.uint8) * 255
alphas[:, -65:-57] = 0
alphas[:9, :-43] = 0
alphas[-8:, :-43] = 0
for i in range(img.shape[0]):
for j in range(img.shape[1] - 43, img.shape[1]):
pixel = tuple(img[i, j, :])
# if sum(pixel) == 765:
if sum(pixel) < 450 and j >= img.shape[1] - 33:
img[i, j, :] = [147, 149, 152]
elif j >= img.shape[1] - 33 or i < 9 or i > img.shape[0] - 9:
alphas[i, j] = 0
img = np.dstack((img, alphas))
plt.imsave(fname=out_path, arr=img)
plt.close()
