various improvements

[bertsky]

Fixes #6 and #9. If this all-at-once PR is inconvenient, I can also split it up functionally.

I've also changed the st.image(figure.savefig()) to st.pyplot(figure) because that seemed faster. It depends on the image resolution and dpi parameter for savefig though. Still, it seems more logical that way. The single parameter to control the size/speed trade-off would be figsize then. Perhaps we should make that configurable?

I have also added an input field to look up images by file_name (which is helpful for navigation in large datasets), and an option to make the category view show only that single category.

[bertsky]

The single parameter to control the size/speed trade-off would be figsize then. Perhaps we should make that configurable?

Added that as well. This really helps browsing faster through a large dataset (with lower resolution), while still allowing high-res when needed.

I have also redefined the GT annotation type now: pred was completely redundant, but what's really needed is a distinction between false negatives / GT misses and GT in general. For example, one might want to view just TP/FP/TN, or just GT.

Lastly, I found a way to deal with overlapping labels by integrating a nice tool for mag(net)ic artist placement in matplotlib:

(I tried with mpl_connect("motion_notify_event", hover_func) but found that st.pyplot does not wrap events at all. Perhaps a mouse hover functionality could be arrived at via plotly instead of matplotlib?)

[i008]

Amazing contribiutions! thanks a lot. One issue i found is when using the select image by path:
when i just use the path printed (full path to image) i get:

Any idea whats happening?

[bertsky]

Any idea whats happening?

Oh, sorry. I confused image_id with image number there (again). Should be fixed now.

[i008]

Great, thanks a lot!