New median_distance func for near neighbors

[leouieda]

The new median_distance function calculates the median near neighbor distance between each point in the given dataset. It automates the process of creating a kdtree and querying it on the same input points. It will take a median of the k nearest neighbors for each point and handle projections and reshaping of the output.

This type of function is useful to get a sense for how scattered are the data points and what would be a reasonable grid spacing or block mean spacing.

Fixes #127

Reminders

• Run make format and make check to make sure the code follows the style guide.
• Add tests for new features or tests that would have caught the bug that you're fixing.
• Add new public functions/methods/classes to doc/api/index.rst and verde/__init__.py.
• Write detailed docstrings for all functions/classes/methods.
• If adding new functionality, add an example to the docstring, gallery, and/or tutorials.

[leouieda]

@jessepisel I could use some feedback on this PR. I'm not entirely sure this is the best way of doing this sort of calculation or if the results are actually meaningful for setting grid spacings etc. What do you think?

[leouieda]

Actually, I'm gonna go ahead and merge this because I will need this function for some new gridders. But I'll wait a while before using it in the docs (#164).