Do not use this matlab function
You should use miss_hit instead which has a lot more features and is actively maintained.
If however for some reason that I don't quite understand you still wanted to use
check_my_code, then please read on from there.
If you have python installed.
pip install miss_hitThen from your terminal command line.
mh_style --fix path_to_your_matlab_project
mh_metric --ci path_to_your_matlab_project
mh_lint path_to_your_matlab_projectThis will install and run the awesome miss_hit python package that will lint and reformat of your MATLAB code.
It was created that by someone (Florian Schanda) who actually knows what he's doing.
It also integrates fairly well with pre-commit to make sure you only commit clean code.
Because we all need pointers when it comes to make our codebase better.
A lot of us (at least in neuroscience) learn MATLAB by aping scripts we have borrowed from someone: one day you were given some .m files, a tap on the shoulder and "good luck" and you have been reverse engineering the wheel on what it means to write good code since then.
One of the reason for this is that people, whose code we are learning from, are as likely as you to have taken 'How to write good code 101' (i.e very unlikely but don't feel bad those classes are usually not proposed in most neuroscience departments anyway).
All of this is the fastest way to end with a codebase of MATLAB scripts of 1000 lines or above that that have more loose ends than a bowl of spaghetti and are a nightmare to debug or to read (for others or for you in 6 months).
In case you want to avoid this, this function can help you by pointing the files that are getting too complex or have too few comments.
The McCabe complexity of a function is presents how many paths one can take
while navigating through the conditional statements of your function (if,
switch, ...). If it gets above 10 you enter the danger zone. If you are above
15 you want to seriously consider
refactoring your code with
sub-functions for example.
Once you have downloaded (or clones) this repository and added its content to the Matlab path (see here), you can use the function any time by typing this in the MATLAB prompt:
check_my_codeThat will give you the The McCabe complexity of all the .m files in the
current directory. It will also return the complexity of the sub-functions in
each file.
This will also check the proportion of lines with comments in each file (might overestimate it). In general you might want to try to be around 20%.
It will then list the functions that do not meet the requirements you have set for your projects. You can then use this information to figure out which function you should refactor first.
The idea for this is partly inspired by this "equivalent" in python
BOOLEAN: if set to true this will check the .m files in all the sub-folders.
(default: false)
1 X 2 ARRAY: Thresholds for the acceptable McCabe complexity before triggering a
warning. Having 2 values lets you decide a zone of complexity that is high but
acceptable and another that is too high. (default: [15 20])
1 X 2 ARRAY : thresholds for the acceptable percentage of comments in a file
before triggering a warning. Having 2 values lets you decide levels that are low
but acceptable and another that is too low. (default: [20 10])
BOOLEAN this will print a file with the overall error code ; mostly used for automation for now. (default: true)
an array with [cplx_error_code comment_error_code] where each value is 0 if
there is no file that is too complex or has too few comments and is 1 otherwise
a n X 2 cell listing of all the function in {i,1} and subfunction in {i,2} tested. If the function is the main function of a file then {i,1} and {i,2} are the same.
an array with the complexity of each function and subfunction
an array with the percentage of comment in each file
Some aspects of this function will require Matlab 2017a (recursive search through sub-folders) or above to work.
Also because octave does not have a linter, so this will only work with Matlab.
Click on the Clone or download button and then on Download ZIP. Unzip the
downloaded file and add the content of the zipped folder to your Matlab path.
If you use Git you can get this repository by typing this in a terminal:
git clone https://github.com/Remi-Gau/matlab_checkcode.gitOtherwise you could fork this repository onto your github account by clikcing
the fork on the top right of the screen and then clone that copy onto your
computer by typing.
git clone https://github.com/YOUR_GITHUB_USERNAME/matlab_checkcode.gitIn either case you will then need to add the newly created folder to your matlab path.
If you use the matlab package manager, to simply download this repository by typing this in the Matlab prompt:
mpm install matlab_checkcode -u https://github.com/Remi-Gau/matlab_checkcode.gitThis will add check_my_code to the Matlabpath, but you will have to save the
path if you want to make this permanent or run mpm init next time your start
matlab.
It relies on the linter used natively by MATLAB so it could also be extended to check all the messages relating to all the little other issues in your code that you have not told MATLAB to ignore.
Also because octave does not have a linter, this will only work with Matlab. 😭
Because we don't brush our teeth just the day before we go to the dentist.
If the check_my_code function is in the MATLAB path, you can automate its
usage if you use Git for your project.
I have created a git hook that will execute the check_my_code every time you
try to push to your remote repository. If your code is not up to the standard
then the push will be aborted.
To use this here is what you need to do:
- Copy this file into your project/.git/hooks
- Rename it to pre-push
- You might need to modify the
alias matlabline to point this script to where Matlab is on your computer. - Make this file executable with
chmod +x .git/hooks/pre-push - Now your code quality will be checked when you push your code to your remote.