Pseudocolor images with layered plotting?

[nfahlgren]

Over time we have added input options to the pcv.visualize.pseudocolor function and there's a potential for it to get unwieldy. I'm not sure if folks would see this as more user-friendly or not, but what if we created a (ideally) flexible layered plotting scheme?

class Pseudocolor:
    def figure(self):
        self.fig = plt.figure()
        return self
    def plot_layer(self, img, cmap):
        plt.imshow(img, cmap=cmap)
        return self
    def add_colorbar(self):
        plt.colorbar(fraction=0.033, pad=0.04)
        return self
    def add_title(self, title):
        plt.title(title)
        return self
    def remove_axes(self):
        plt.xticks([])
        plt.yticks([])
        return self

This class doesn't implement everything included in pseudocolor currently but it works for a demo. Each class method returns the instance of the class allowing you to chain methods together.

# Create a new instance
ps = Pseudocolor()

# Create an empty figure (not very interesting)
fig = ps.figure()
# Add a plotting layer
fig = fig.plot_layer(img=gray_img, cmap="gray")

Or chain methods together instead:

fig = ps.figure().plot_layer(img=gray_img, cmap="gray").plot_layer(img=np.ma.array(gray_img2, mask=~bin_clean.astype(bool)), cmap="viridis")

Remove the background layer and add a title and colorbar:

fig = lp.figure().plot_layer(img=np.ma.array(gray_img2, mask=~bin_clean.astype(bool)), cmap="viridis").add_title("Pseudocolor image on white background").add_colorbar()