pseudocolor function in visualization does not behave as expected

[DannieSheng]

I was trying to update the pseudocolor function by including the "showing bad mask" feature. However when I test on the current version of pseudocolor, the behavior is not as expected. Specifically, I tested:
# Pseudocolor the same image but include the mask and limit the range of values pseudo_img_masked = pcv.visualize.pseudocolor(gray_img=img, obj=None, mask=mask, background="white", cmap='viridis', min_value=30, max_value=200)
which is in our documentation, and a visualization of the pseudocolored plant in the mask area would be shown, with the background being white:

But what I actually got is this:

I am not sure if it is related to some Matplotlib backend settings. I also tried to use the source code of pcv.visualize.pseudocolor but it seemed that the it did not behave as expected.

[DannieSheng]

After testing several potential solutions, I figured out one solution that (hopefully) addresses the issues. But before I include it into my current PR #569 , I would like to share the code here because it might me some not trivia changes. Feel free to direct your self here: /shares/mgehan_share/hsheng/projects/test_plantcv/threshold_mask_bad/mask_bad_visualize/pseudocolor_update.ipynb
(Test data already saved where the notebook located)

But in brief, the ideas are:
If mask is given, and background is either 'white' or 'black', set the 'bad_color' for current Matplotlib color map to be background, i.e.

cmap_current = copy.copy(cm.get_cmap(cmap))
cmap_current.set_bad(color=background)
masked_img = np.ma.array(gray_img1, mask=~mask.astype(np.bool))
fig, ax = plt.subplots()
pseudo = ax.imshow(masked_img, cmap=cmap_current, vmin=min_value, vmax=max_value)

If bad_mask is given, plot those "bad" pixels on top of the pseudocolored image using plot, i.e.

bad_idx, bad_idy = np.where(bad_mask > 0)
bad_pixs = ax.plot(bad_idy, bad_idx, '.', color=bad_color)

[nfahlgren]

I get something different, though still not working correctly. Plotting on the background image works:

ps = pcv.visualize.pseudocolor(gray_img=gray_img, min_value=30, max_value=200, 
                               mask=bin_clean, background="image", cmap="viridis")

But the "white" or "black" background both show up black for me:

ps = pcv.visualize.pseudocolor(gray_img=gray_img, min_value=30, max_value=200, 
                               mask=bin_clean, background="image", cmap="white")

[nfahlgren]

I think the issue on my end is that when we set the background to all white or all black and use the gray colormap it always uses the first value of the colormap, which is black. So if we use the gray_r colormap for the white background it would fix it.

[nfahlgren]

I'm on matplotlib v3.3.4 for reference

[DannieSheng]

You are absolutely right! I think now I got the results we expected!

[DannieSheng]

addressed in #569