[Chicago] merge_closest package, like VLOOKUP( ... ,TRUE)

[Robin Fishbein]

Hello again, ChiPy!
My first time making a Python package and a GitHub repo:https://github.com/RobinFiveWords/merge_closest/

It's a single function that merges pandas DataFrames with a closest-without-going-over match, like Excel's VLOOKUP function when the last argument is TRUE. I couldn't find existing code out there, so I ended up writing it...and then rewriting it when I discovered numpy.searchsorted().
If this exact thing already exists out there, please break it to me gently. =]
I'd very much like to learn how to build better packages. I imagine there's somewhere other than the __init__.py file that the public (only) function should go, but I'm not sure what is a standard and what is personal preference. Tests, certainly, although it isn't clear to me what tests to write for a single function. Any comments or suggestions are very welcome.
Thanks!!-Robin Fishbein
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/chicago/attachments/20160920/b322b482/attachment.html>