Handle errors in findDepth endpoint

[Waidhoferj]

The find depth functionality takes in a URL to an image stored in Cloudinary. It should download the image, run the depth estimation algorithm on it and return a 2D array of depth values. Most of this already works, but some edge cases aren't covered.

• What if URL doesn't point to a valid Cloudinary resource (video instead of image)?
• What if no URL is provided?
• What if the model fails to process the time?

Handle these edge cases in the findDepth function in main.py

[snekiam]

@bjtat this should help you get started on this:
You can run the server by running python3 main.py -l, which will use the local mocked database.
You can make a call to the API by using Postman, which I'd recommend as a nice GUI for making post/get requests - you can download it here

Here's my setup to hit the findDepth() function with a valid URL:

This should get you started running that function, and you'll be able to feed it bad URLs, stuff like that to test error handling - just change the raw data being provided by postman. I'm not sure if we have any non-image files in Cloudinary, but you could test with a non-cloudinary URL to a non-image resource (you can use this as an MP4 if you wanna test with non-images).

Let me know if you have any questions!

[snekiam]

Timeouts might be something we have to define in Gunicorn, or whatever we use for our production hosting.