JOSS Review about pymgpipe HOT 5 CLOSED

@cumbof thank you for taking the time to review our package. These points are very helpful- I will try my best to address each of them over the next couple of days.

from pymgpipe.

@cumbof I have made the suggested changes and updated the joss branch as well. Please don't hesitate to let me know if you have any additional questions, comments, concerns. Thank you!

from pymgpipe.

Seems great overall. Thanks for working on my comments @ym2877.

I only have a couple of extra comments:

1. you addressed point 3 by moving the .VERSION file into the pymgpipe/resources as specified in the setup.py file. However, I don't actually see any .VERSION file under that folder. I'm ok with taking track of the package version with the .VERSION file instead of specifying it in the pymgpipe/__init__.py file as I suggested. Just remember to push that file into the resources folder;
2. you should add a few sentences in the README.md about how you handle people's contributions (since this is open source) and how people should report issues. Look at the example here;
3. what about point 4? That's a very minor point compared to the others and it would be enough to have the package in PyPI. However, making it available in bioconda would be way better and it will allow people to avoid taking care of installing the external software dependencies by their own.

Point 1 and 2 must be addressed, while point 3 is strongly recommended but not mandatory.

from pymgpipe.

Okay, addressed points #1 and #2! Thanks again for your comments @cumbof. Point #1 was just an error, that reference to .VERSION was old and shouldn't have been there, so thank you for catching that!

With regard to #3, I will absolutely look into this. Although, with the way our dependencies are currently set up in setup.py, there is no need for users to install external dependencies on their own. The only steps they would need to run is

1. conda create -n my_env python=3.10
2. conda activate my_env
3. pip install pymgpipe

Maybe I am misunderstanding what you mean. Either way, I will look into publishing a bioconda package asap!

from pymgpipe.

Thanks for your quick reply @ym2877.

That's true, your tool doesn't have any non-python dependencies actually and everything is handled by pip. However, I would consider adding your tool to bioconda in case anyone wants to use your tool as a dependency along with other non-python tools.

Again, I don't consider this point as mandatory in the context of the JOSS review, but I would strongly recommend to consider addressing this extra step. Defining a conda recipe would be extremely easy at this point.

I'm going to close this issue since I don't have any other comments.

from pymgpipe.