Show HN: Laboratory, a Python port of GitHub's Scientist library for refactoring | Hacker News Reader

Show HN: Laboratory, a Python port of GitHub's Scientist library for refactoring

github.com

153 pointsbuttscicles13 comments

Open articleView on HN

Here's some other ports of Scientist I've come across as well:

PHP

https://github.com/LartTyler/Analyst

Javascript

https://github.com/trello/scientist

https://github.com/Runnable/scientist

https://github.com/jirwin/labrat

.NET

http://haacked.com/archive/2016/01/20/scientist/

I imagine this can only work on getters? How do you use this when code makes state changes?

It only should be used for code which doesn't have side effects. According to original github lib.

From Github's Scientist [1]:

Because `enabled?` and `run_if` determine when a candidate runs, it's impossible to guarantee that it will run every time. For this reason, Scientist is only safe for wrapping methods that aren't changing data.

When using Scientist, we've found it most useful to modify both the existing and new systems simultaneously anywhere writes happen, and verify the results at read time with `science`.

[1]: https://github.com/github/scientist/blob/master/README.md#de...

Before clicking the link I wondered how this would translate the block-based Ruby API into Python. Using Python's with statement is quite nice here, I think.

The with statement is one of my favorite additions to the language. It's a really flexible syntax to encapsulate control flow, exceptions, necessary cleanup steps… etc. etc.

I would argue that python provides even simpler syntax to enable functionality like this.

  @experiment.with(new_function)
  def old_function()
      ...

Agreed -- a decorator seems like a much cleaner way to handle this.

A decorator must be applied to a function or class, whereas a context manager can wrap an arbitrary section of code. So, it depends on how you want the tool to be used. I don't think there's a clear best design.

This was my thought as well, and it's also a bit more awkward if your function signatures differ. Luckily they aren't mutually exclusive and it should be easy to implement a decorator version.

Another Python one: https://github.com/staticshock/scientist.py

Cool! Thanks for the tool!

> Imagine you've implemented a complex caching strategy for some objects in your database and a stale cache is simply not acceptable.

I hope you have a formal specification of that cache. That will help ensure your design is correct in the first place before you even get to running it in code.

I shall be playing with this soon. Cheers!

Python? Isn't Python dying? ;) https://news.ycombinator.com/item?id=11100251

Reply on news.ycombinator.com