def test_num_axes_hist(image_array_2bands, basic_image):
"""We expect one AxesSubplot object per band."""
f_1, ax_1 = ep.hist(basic_image)
assert len(f_1.axes) == 1
f_2, ax_2 = ep.hist(image_array_2bands)
assert len(f_2.axes) == 2
plt.close(f_1)
plt.close(f_2)
def test_single_hist_2titles(image_array_single_band):
"""Test that a single custom title works for one band hists."""
custom_title = ["Great hist", "another title"]
with pytest.raises(ValueError,
match="You have one array to plot but more than one"):
ep.hist(image_array_single_band, title=custom_title)
plt.close()
def test_number_of_hist_bins(basic_image):
"""Test that the number of bins is customizable."""
n_bins = [1, 10, 100]
for n in n_bins:
f, ax = ep.hist(basic_image, bins=n)
assert n == len(ax.patches)
plt.close()
def test_hist_number_of_columns(image_array_2bands):
"""Test that the col argument changes the number of columns."""
number_of_columns = [1, 2]
for n in number_of_columns:
f, ax = ep.hist(image_array_2bands, cols=n)
assert [a.numCols for a in ax] == [n] * 2
plt.close(f)
def viewhist(self): # Function to display the histogram in the frame and save it in the specified directory
if self.display.winfo_exists():
self.display.grid_forget()
self.display.destroy()
self.display = ttk.Frame(self)
self.display.grid(row = 0, column = 4,rowspan=2, sticky = 'nwes')
frame=self.display
band_path=[]
for file in self.histfiles:
band_path.append(file)
# dir = self.dir[0]
# band_path = glob.glob(os.path.join(dir,"*.tif"))
# band_path = glob.glob(os.path.join(dir,"*.TIF"))
band_path.sort()
band_stack, meta_data = es.stack(band_path, nodata=-9999)
size = len(band_path)
color_list = ["Indigo","Blue","Green","Yellow","Red","Maroon","Purple","Violet"]
titles = []
for i in range(size):
titles.append("File"+str(i+1))
self.figure,self.plot = ep.hist(band_stack, title = titles,figsize = (6,4))
self.canvas = FigureCanvasTkAgg(self.figure,frame)
self.toolbar = NavigationToolbar2Tk(self.canvas,frame)
self.toolbar.update()
self.canvas.get_tk_widget().pack()
def test_hist_partially_masked_array(image_array_3bands):
ones = np.ones(image_array_3bands[0].shape)
image_array_3bands[2] = ones
masked_arr = np.ma.masked_where(image_array_3bands == 1,
image_array_3bands)
f, ax = ep.hist(masked_arr, cols=3)
assert len(f.axes) == 3
def test_hist_fully_masked_array(image_array_single_band):
masked_arr = np.ma.masked_where(
image_array_single_band >= 0, image_array_single_band
)
with pytest.raises(ValueError, match="is not finite"):
f, ax = ep.hist(masked_arr)
plt.close(f)
def test_multiband_hist_title(image_array_2bands):
"""Test that custom titles work for multiband hists."""
custom_titles = ["Title 1", "Title 2"]
f, ax = ep.hist(image_array_2bands, title=custom_titles)
num_plts = image_array_2bands.shape[0]
assert [f.axes[i].get_title() for i in range(num_plts)] == custom_titles
plt.close(f)
def test_hist_1band_masked_array(image_array_single_band):
"""Ensure that a masked single band arr plots & the number of bins is correct"""
masked_arr = np.ma.masked_where(image_array_single_band == 4,
image_array_single_band)
nbins = 3
f, ax = ep.hist(masked_arr, title=["TITLE HERE"], bins=nbins)
assert len(f.axes) == 1
assert len(ax.patches) == nbins
plt.close(f)
def test_hist_plot_odd_axes(image_array_3bands):
"""A histogram with 2 cols and 3 plots should have 4 axes total.
this will ensure code coverage for clearing out those axes"""
ncols = 2
f, ax = ep.hist(image_array_3bands, cols=ncols)
assert len(f.axes) == ncols * 2
plt.close(f)
def test_hist_color_string(image_array_2bands):
f, ax = ep.hist(image_array_2bands, colors="blue")
colors = [a.patches[0].__dict__.get("_original_facecolor") for a in ax]
expected_colors = [
np.array([0.0, 0.0, 1.0, 1.0]),
np.array([0.0, 0.0, 1.0, 1.0]),
]
for i in range(2):
assert np.array_equal(colors[i], expected_colors[i])
plt.close(f)
def test_hist_masked_array(image_array_2bands):
""" Test that a masked 2 band array plots properly"""
masked_arr = np.ma.masked_where(image_array_2bands == 6,
image_array_2bands)
nbins = 6
f, ax = ep.hist(masked_arr, bins=nbins)
for a in ax:
assert len(a.patches) == 6
assert len(f.axes) == 2
plt.close(f)
def test_hist_color_multi_band(image_array_2bands):
"""Test that multiple colors work for multiband images."""
f, ax = ep.hist(image_array_2bands, colors=["red", "blue"])
colors = [a.patches[0].__dict__.get("_original_facecolor") for a in ax]
expected_colors = [
np.array([1.0, 0.0, 0.0, 1.0]),
np.array([0.0, 0.0, 1.0, 1.0]),
]
for i in range(2):
assert np.array_equal(colors[i], expected_colors[i])
def test_hist_bbox(basic_image):
"""Test that the bbox dimensions are customizable."""
f, ax = ep.hist(basic_image, figsize=(50, 3))
bbox = str(f.__dict__.get("bbox_inches"))
assert bbox == "Bbox(x0=0.0, y0=0.0, x1=50.0, y1=3.0)"
def test_hist_color_single_band(basic_image):
"""Check that the color argument works for single bands."""
f, ax = ep.hist(basic_image, colors=["red"])
facecolor = ax.patches[0].__dict__.get("_original_facecolor")
assert np.array_equal(facecolor, np.array([1.0, 0.0, 0.0, 1.0]))
def test_num_plot_titles_mismatch_hist(image_array_2bands):
"""Raise an error if the number of titles != number of bands."""
with pytest.raises(ValueError, match="number of plot titles"):
ep.hist(image_array_2bands, title=["One", "too", "many"])
def test_single_hist_title(basic_image):
"""Test that custom titles work for one band hists."""
custom_title = "Great hist"
f, ax = ep.hist(basic_image, title=[custom_title])
assert ax.get_title() == custom_title
def test_hist_1band_range(image_array_single_band):
"""Ensure that the range limits are reflected in the plot"""
f, ax = ep.hist(image_array_single_band, hist_range=(2, 5))
range_check = ax.get_xlim()
assert range_check[0] > 1.5 and range_check[1] < 5.5
stretch=True,
extent=extent,
str_clip=0.5,
title="RGB Image of Un-cropped Raster",
)
plt.show()
################################################################################
# Explore the Range of Values in the Data
# ---------------------------------------
# You can explore the range of values found in the data using the EarthPy ``hist()``
# function. Do you notice any extreme values that may be impacting the stretch
# of the image?
ep.hist(array, title=["Band 1", "Band 2", "Band 3"])
plt.show()
###########################################################################
# No Data Option
# ---------------
# ``es.stack()`` can handle ``nodata`` values in a raster. To use this
# parameter, specify ``nodata=``. This will mask every pixel that contains
# the specified ``nodata`` value. The output will be a numpy masked array.
os.chdir(os.path.join(et.io.HOME, "earth-analytics"))
array_nodata, raster_prof_nodata = es.stack(stack_band_paths, nodata=-9999)
# View hist of data with nodata values removed
ep.hist(
array_nodata,
arr_st, meta = es.stack(landsat_paths_pre)
# Import the landsat qa layer
with rio.open(
"data/cold-springs-fire/landsat_collect/LC080340322016070701T1-SC20180214145604/crop/LC08_L1TP_034032_20160707_20170221_01_T1_pixel_qa_crop.tif"
) as landsat_pre_cl:
landsat_qa = landsat_pre_cl.read(1)
landsat_ext = plotting_extent(landsat_pre_cl)
###############################################################################
# Plot Histogram of Each Band in Your Data
# ----------------------------------------
# You can view a histogram for each band in your dataset by using the
# ``hist()`` function from the ``earthpy.plot`` module.
ep.hist(arr_st)
plt.show()
###############################################################################
# Customize Histogram Plot with Titles and Colors
# -----------------------------------------------
ep.hist(
arr_st,
colors=["blue"],
title=[
"Band 1",
"Band 2",
"Band 3",
"Band 4",
"Band 5",
sns.set(font_scale=1.5, style="whitegrid")
# Open data and assign negative values to nan
with rio.open(str(p)) as src:
dem = src.read(1, masked=True)
std = np.std(dem)
print(std)
# fig, ax = plt.subplots(figsize = (10, 5))
# im = ax.imshow(dem.squeeze())
# ep.colorbar(im)
# ax.set(title="Diff")
# ax.set_axis_off()
# plt.show()
# View object dimensions
fig, ax = ep.hist(dem,
colors=['grey'],
title="3σ histogram %s" % p.name,
xlabel='Band value (meters)',
bins=20,
hist_range=(-3 * std, 3 * std))
plt.text(0.5,
0.5,
r'$n=%d,\ \sigma=%.3f$' % (len(dem), std),
transform=ax.transAxes,
bbox=dict(facecolor='red', ))
plt.grid(True)
plt.savefig('histogram-%s.png' % p.name, )
def test_hist_plot_1_dim(image_array_2bands):
"""A single dimensional array should plot properly."""
array_1_dim = image_array_2bands.ravel()
f, ax = ep.hist(array_1_dim)
assert len(f.axes) == 1
plt.close(f)
# Creating a raster histogram to investigate the differences in elevation
# In[15]:
lidar_dem_im #values in array
# In[16]:
# Plotting the histogram
ep.hist(lidar_dem_im[~lidar_dem_im.mask].ravel(),
bins=100,
title="Lee Hill Road - Digital elevation (terrain) model - \nDistribution of elevation values")
plt.show()
# We can see here that there are some differences in elevation. Most raster cells have values around 1,700 meters, but there are also many values that are between 1,800 - 1,950 meters tall.
# Now, I'm adjusting the plotting extent, so that we can "zoom up" on the data.
#
# (This just means that we readjust the bounding box of the raster, and plot a smaller area. This way, we can see more detail on some of the cells.)
# In[17]:
# Define a spatial extent that is "smaller" - so that this is what we're zooming up on in the final product.
def test_hist_axes(image_array_2bands):
f, ax = ep.hist(image_array_2bands, xlabel="xlabel", ylabel="ylabel")
for i in range(2):
assert ax[i].get_xlabel() == "xlabel"
assert ax[i].get_ylabel() == "ylabel"
def test_hist_single_band_3_dims(image_array_single_band_3dims):
f, ax = ep.hist(image_array_single_band_3dims)
assert len(f.axes) == 1
plt.close(f)
def test_title_string(basic_image):
"""Test that custom titles work for one band hists."""
custom_title = "Great hist"
f, ax = ep.hist(basic_image, title=custom_title)
assert ax.get_title() == custom_title
plt.close(f)
ax.set_title("Digital Elevation Model \n Shuttle Radar Topography Mission (SRTM), NASA \n , Western Uganda", fontsize=24)
plt.show()
lidar_dem_im #values in array
# In the original GeoTIFF, I was having trouble visualizing (in a 2D method) the elevation for the raster histogram.
#
# All I 'knew' about this data was the following:
# * The Rwenzori mountains are the tallest mountain range in Africa
# * The main town in the region, Fort Portal, Uganda, is *roughly* the same height as Denver, CO, USA.
#
# As a result, I thought this would be a particularly useful exercise to see just HOW MUCH elevation differences there would be!
ep.hist(lidar_dem_im[~lidar_dem_im.mask].ravel(),
bins=100,
title="Digital Elevation Model \n Shuttle Radar Topography Mission (SRTM), NASA \n , Western Uganda \nDistribution of elevation values")
plt.show()
# The above isn't very useful - we see the "5000 [m]" entry, but it's an outlier. I'm curious to see if there's a distribution of heights listed.
ep.hist(lidar_dem_im[~lidar_dem_im.mask].ravel(),
bins=500,
title="Digital Elevation Model \n Shuttle Radar Topography Mission (SRTM), NASA \n , Western Uganda \nDistribution of elevation values (500 bins)")
plt.show()
#downgrading the number of bins shown
ep.hist(lidar_dem_im[~lidar_dem_im.mask].ravel(),
bins=50,
def test_hist_plot_1_band_array(basic_image):
"""A 2 dim image should plot properly returning a single axis object"""
f, ax = ep.hist(basic_image)
assert len(f.axes) == 1
plt.close(f)
"midnightblue",
"Blue",
"Green",
"Red",
"Maroon",
"Purple",
"Violet",
]
# Create the list of titles for each band. The titles and colors listed
# in this example reflect the order and wavelengths of the Landsat 8 bands
titles = ["Ultra Blue", "Blue", "Green", "Red", "NIR", "SWIR 1", "SWIR 2"]
# Plot the histograms with the color and title lists you just created
# sphinx_gallery_thumbnail_number = 5
ep.hist(array_stack, colors=colors_list, title=titles)
plt.show()
###############################################################################
# Customize Bin Size and Arrangement of Histograms
# -------------------------------------------------
#
# You can customize the number of bins each histogram plot uses to group the
# data it is plotting. The default number is 20. This can be adjusted to match
# the data you are trying to display. Additionally, you can change the
# arrangement of the image overall by modifying the number of columns used
# to plot the data.
# Plot each histogram with 50 bins, arranged across three columns
ep.hist(array_stack, bins=50, cols=3)
plt.show()
def test_hist_multiband_array(image_array_3bands):
f, ax = ep.hist(image_array_3bands, hist_range=(2, 5), cols=3)
for a in ax:
range_check = a.get_xlim()
assert range_check[0] > 1.5 and range_check[1] < 5.5
