Creating a repo for test/doc image/data

[DannieSheng]

I have been thinking about having a "library" / repository for all test/doc image/data, such that people working on the same function, either simultaneously or not, can easily find the image/data used for testing, generating sample images for the documentation page.
I've discussed it with @nfahlgren , and here are some useful features for this repository (or library) he mentioned to be useful:

1. Having all original source data files
2. Jupyter notebooks that convert source files into the forms used in the docs/tests (e.g. scaled down in size or converted between types)
3. A nice table or other resource that has the source and final file names and a description of what the data is. E.g. table to understand/organize test data
   helpful to understand test data:
   https://docs.google.com/spreadsheets/d/1cfNqEvvkGYdlzIV62qZOSZNJJJwVysHEi4n0FQAUbTE/edit?usp=sharing

We can think of other useful features of this repo.

Another thing is related to tests in PlantCV. I started using some synthetic data for tests. Though I am not sure if it is reasonable to do that. For example, to test the pcv.viduslize.time_lapse_video function in the PR #762 , a folder of test images is required. Instead of having a folder of test images stored in the test_data folder, I generate a folder of images and store in the cache_dir. e.g.:

def test_plantcv_visualize_time_lapse_video_cases():
    # Test cache directory
    cache_dir = os.path.join(TEST_TMPDIR, 'visualize_time_lapse_video_case1')
    os.mkdir(cache_dir)
    # Generate 3 test images and saved in cache_dir
    for i in range(3):
        temp_img = pcv.transform.rescale(np.random.rand(3,3))
        pcv.print_image(temp_img, filename=os.path.join(cache_dir, f"img{i}.png"))

[DannieSheng]

Try not to use PlantCV functions in tests

[nfahlgren]

I created the repository for this: https://github.com/danforthcenter/plantcv-data