jack-willturner commented 3 years ago

I personally prefer working with yaml files since I find them a little easier to read/write.

There are a few other small changes:

let Rich overflow vertically so we see updated results in the tables
add a -config shortcut flag to argparse because I'm tired of writing out the whole filepath_to_json_arguments_config
include a couple of small example configs - including an imagenet one which replicates the pytorch default script

sourcery-ai[bot] commented 3 years ago

Sourcery Code Quality Report

❌ Merging this PR will decrease code quality in the affected files by 0.40%.

Quality metrics	Before	After	Change
Complexity	10.57 🙂	10.61 🙂	0.04 👎
Method Length	102.80 🙂	105.87 🙂	3.07 👎
Working memory	13.57 😞	13.59 😞	0.02 👎
Quality	50.83% 🙂	50.43% 🙂	-0.40% 👎

Other metrics	Before	After	Change
Lines	524	542	18

Changed files	Quality Before	Quality After	Quality Change
train.py	41.23% 😞	41.20% 😞	-0.03% 👎
utils/arg_parsing.py	70.50% 🙂	67.90% 🙂	-2.60% 👎

Here are some functions in these files that still need a tune-up:

File	Function	Complexity	Length	Working Memory	Quality	Recommendation
utils/arg_parsing.py	process_args	13 🙂	156 😞	11 😞	48.02% 😞	Try splitting into smaller methods. Extract out complex expressions
train.py	housekeeping	7 ⭐	168 😞	9 🙂	56.28% 🙂	Try splitting into smaller methods
train.py	get_base_argument_parser	0 ⭐	234 ⛔	9 🙂	58.21% 🙂	Try splitting into smaller methods
train.py	train	1 ⭐	92 🙂	12 😞	65.49% 🙂	Extract out complex expressions
train.py	eval	1 ⭐	72 🙂	12 😞	68.42% 🙂	Extract out complex expressions

Legend and Explanation

The emojis denote the absolute quality of the code:

The 👍 and 👎 indicate whether the quality has improved or gotten worse with this pull request.

Please see our documentation here for details on how these metrics are calculated.

We are actively working on this report - lots more documentation and extra metrics to come!

Let us know what you think of it by mentioning @sourcery-ai in a comment.