Comprehensive fillna

[ivanzvonkov]

Issue:
Inference fails when the input tif file contains a band that has all nan values (with the exception of slope).

This is a prerequisite for making the next Hawaii map.

Additionally, the current solution used to fill nan slope values requires passing the mean slope value in functions such as process_test_file and _find_matching_point where the mean slope is not relevant.

Code changes:

• A more general fillna function which fills nan values for all scenarios (see tests)
• Implementation is simplified by applying fillna immediately after loading the tif file
• The not so robust regex start date extraction is removed since it is no longer absolutely necessary for loading tifs

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[gabrieltseng]

Doing this at tif loading time is so much cleaner