internals: add path completion w tests

[Kappers]

As per 'Autocompletion broke completion for paths (file, destination, database)' #38

Any input would be appreciated, or style/whatever guidelines for the repo (does that information exist anywhere?) I haven't written tests in a bit too long....

This works, but looks a bit messy for deep directories (long paths) or long file names etc. At least it works!

[codecov]

Codecov Report

Merging #61 into master will increase coverage by 0.05%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master      #61      +/-   ##
==========================================
+ Coverage   97.85%   97.91%   +0.05%     
==========================================
  Files          20       20              
  Lines        1028     1055      +27     
==========================================
+ Hits         1006     1033      +27     
  Misses         22       22              

Impacted Files	Coverage Δ
bibo/bibo.py	97.82% <ø> (ø)
bibo/internals.py	92.50% <100.00%> (+0.53%)	⬆️
tests/bibo/test_internals.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 2bbf83d...851929b. Read the comment docs.

[Nagasaki45]

Looks great!

Note that there's an issue with trailing spaces when trying to expand a nested path, at least in zsh. Apparently that's one of the major issues with supporting path completion in multiple shells. Check out long discussion on pallets/click#1403. Anyway, for our purposes, as a temp solution, this is really good.

2 minor comments:

• Can you base future PRs on dev, not master. Everything pushed to master is released to pypi :-)
• Add the number of the issue at the end of the commit to create a link in github from the issue to the commit (with # before the number). If the issue is fixed by the commit add Fix #number.

Adding a note to include these in the dev docs.