Code Review in Academia Workshop by Amelia Edmondson-Stait

Pull requests

edmondsonstaita — 2025-02-10 10:07:06 UTC

Talked about how industry uses pull requests

GitHub issues

edmondsonstaita — 2025-02-21 15:37:36 UTC

For examples see: https://github.com/OxfordCodeReviewNet/forum

edmondsonstaita — 2025-02-21 15:39:23 UTC

Code formatters

2025-03-13 10:05:30 UTC

Code formatters (such as black for Python) can automatically check/apply style rules.

Do industry/particular organisations use a common coding style to help with coding review?

2025-03-13 10:19:29 UTC

Is chatGPT a helpful tool to review code?

2025-03-13 10:23:09 UTC

What are your thoughts?

How can we name/structure folders and files in an understandable way?

2025-03-13 10:34:42 UTC

What are docstrings?

2025-03-13 10:37:39 UTC

R linter package

2025-03-13 10:40:22 UTC

Python linter

2025-03-13 10:42:55 UTC

A basic linter for Python.

How can we make file paths useable by different coders?

2025-03-13 10:47:45 UTC

Unneeded code

2025-03-13 11:30:43 UTC

Look for code that doesn't get executed, functions or packages that are not used

Link

2025-03-13 11:32:12 UTC

What are guard clauses?

edmondsonstaita — 2025-03-13 11:34:23 UTC

Excellent resource for learning git/GitHub

edmondsonstaita — 2025-03-13 11:38:03 UTC

I reccomend data and software carpentries. They run some events in Edinburgh too. They teach good coding practices. Here is one on git and GitHub: https://datacarpentry.github.io/rr-version-control/01-git-github/index.html

We liked the comments that were next to the packages. It was helpful to see why each package was needed. However, some of the packages look like they may do similar things and may not be needed.

2025-03-13 11:48:05 UTC

The quality of the comments were a little inconsistent (sometimes very detailed, sometimes not). Sometimes they explained what the code was doing, and sometimes they explained why the code was being used.

2025-03-13 11:48:42 UTC

We liked that some of the code was written in the actual colours

2025-03-13 11:49:28 UTC

The indentations of some of the functions weren't always very clear

2025-03-13 11:49:55 UTC

Some of the code was repeated and could be turned into functions to make it more succinct

2025-03-13 11:50:25 UTC

The data is called 'data' which isn't very descriptive. There is also a function in R called 'data' which may also cause confusion.

2025-03-13 11:51:29 UTC

It would be good to have a comment about what the input data looks like. Then how it looks when you have cleaned it up at each step.

2025-03-13 11:55:49 UTC

You used the absolute path. This may make it difficult for other users to load in the data.

2025-03-13 11:56:11 UTC

At the end of the Python script there are three functions that could be combined and simplified

2025-03-13 12:13:46 UTC

There are a lot of modules that are imported but not used.

2025-03-13 12:15:22 UTC

Internal checks of the data were useful, however, more comments would aid the understandability of the code.

2025-03-13 12:17:08 UTC

The column name CRP is in double quotes (Line 73)

2025-03-13 12:18:02 UTC

The script uses a lot of interactive commands. It may be helpful to run this in a notebook.

2025-03-13 12:18:54 UTC

Don't include comments with actual counts of individuals when you are working with sensitive data.

2025-03-13 12:19:24 UTC

Line 73 you use the number 10. You could define this as a variable at the top of the script.

2025-03-13 12:20:26 UTC

lubridate and hms provide similar functions. are they both needed?

2025-03-13 12:25:39 UTC

since dplyr is used, could use bind_rows() instead of do.call(rbind, .)

2025-03-13 12:26:19 UTC

could show an example of expected JSON input structure

2025-03-13 12:27:32 UTC

How can we encourage coding review within our lab?

2025-03-13 12:29:47 UTC

ANSWERS TO PRACTICAL SESSION:

edmondsonstaita — 2025-03-13 12:32:11 UTC

https://github.com/AmeliaES/code-review-workshop/issues