Today I went down a long rabbit hole trying to solve a linting error for a coworker. Our team at ShopRunner uses numpy docstrings, and we check our code with flake8 and flake8-docstrings.

My coworker had written a method that looked like this:

def my_sweet_method(self, examples):   """
   This is a dope docstring.
   
   Parameters
   ----------
   examples
       str inputs to be tokenized   Returns
   -------
   Tokenized strings
   """
   ...

When we ran flake8, we kept getting this error “D414 Section has no content” on the first line of the docstring (“This is a dope docstring”). …

Before RC

I first heard about the Recurse Center in 2016. RC hosts self-directed educational retreats for batches of programmers; basically go hang out in Brooklyn with cool people who like learning (yes, please!). I knew I would love to attend one of their retreats, but it was not feasible for me to go to NYC for 6 or 12 weeks (how could I possibly leave my dog for that long!?). …

This is a cautionary tale of how I spent a large amount of money running PySpark clusters and what I discovered about deploying a recommendation engine at scale. I’ll describe how I finally chased down an intermittent error and lessons learned along the way including why you can’t always write PySpark code like Python code and how important it is to check third party library integrations.

Originally published at https://shoprunnerblog.wordpress.com on March 4, 2019.

This past October, I traveled to NYC to give a talk about spending a lot of ShopRunner’s money.

Here’s the video: https://www.youtube.com/watch?v=SFkmrgfwEyE