Dependency for checking the accuracy of the example code not specified

[XingerTang]

All the accuracy.py in the example folder requires the pandas library to check the accuracy of the output, however, the dependency is not specified anywhere. I'm wondering if we can state it somewhere.

[gregorgorjanc]

Should we add requirements.txt file for this case so we don’t bring all that package in just for example? Also a README.txt on how to prepare/install and run examples.

…

-- With regards!

----------------------------------------------------------------- University of Edinburgh Gregor Gorjanc, PhD The Roslin Institute Group leader & Royal Society Fellow Easter Bush @***@***.***<https://twitter.com/GregorGorjanc> EH25 9RG @***@***.******@***.***> Scotland, UK www.ed.ac.uk/roslin/highlanderlab<http://www.ed.ac.uk/roslin/highlanderlab>

----------------------------------------------------------------- On Thu, Jul 13 2023 at 4:31 pm, XingerTang ***@***.******@***.***>> wrote: This email was sent to you by someone outside the University. You should only click on links or attachments if you are certain that the email is genuine and the content is safe. All the accuracy.py in the example folder requires the pandas library to check the accuracy of the output, however, the dependency is not specified anywhere. I'm wondering if we can state it somewhere. — Reply to this email directly, view it on GitHub<#4>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AAFRGSVIMFHPVSSZB2VUECDXQAIF3ANCNFSM6AAAAAA2JD5NQ4>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***> The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. Is e buidheann carthannais a th’ ann an Oilthigh Dhùn Èideann, clàraichte an Alba, àireamh clàraidh SC005336.

[XingerTang]

@gregorgorjanc

Yes, I think that would help.

I'll work on it after the transfer to tuple datatype is completed.

[gregorgorjanc]

Could we routinely run examples on a push via GitHub action - as a part of a suite of tests? We will need to have this infrastructure in place at some point anyway to catch bug’s quickly and enable safe code modification, refactoring, …

…

-- With regards!

----------------------------------------------------------------- University of Edinburgh Gregor Gorjanc, PhD The Roslin Institute Group leader & Royal Society Fellow Easter Bush @***@***.***<https://twitter.com/GregorGorjanc> EH25 9RG @***@***.******@***.***> Scotland, UK www.ed.ac.uk/roslin/highlanderlab<http://www.ed.ac.uk/roslin/highlanderlab>

----------------------------------------------------------------- On Thu, Jul 13 2023 at 7:50 pm, XingerTang ***@***.******@***.***>> wrote: This email was sent to you by someone outside the University. You should only click on links or attachments if you are certain that the email is genuine and the content is safe. Yes, I think that would help. I'll work on it after the transfer to tuple datatype is completed. — Reply to this email directly, view it on GitHub<#4 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AAFRGSXCBYAEGXVYVRM43I3XQA7O5ANCNFSM6AAAAAA2JD5NQ4>. You are receiving this because you commented.Message ID: ***@***.***> The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. Is e buidheann carthannais a th’ ann an Oilthigh Dhùn Èideann, clàraichte an Alba, àireamh clàraidh SC005336.

[XingerTang]

@gregorgorjanc We can create a workflow that uses pytest to test the code, the workflow can be triggered by pushes or pull requests as usual. But we also need to write test functions with expected outputs as well.

[gregorgorjanc]

I like the sound of pytest - how would this look like compared to the example and tests we have in some of our packages? I suggest we try and do most/all these implementations in AlphaPeel first and then we port them to other packages. Otherwise we will loose track of what we have done where;)

[XingerTang]

@gregorgorjanc
According to the homepage of pytest, its features are

• Detailed info on failing assert statements (no need to remember self.assert* names)
• Auto-discovery of test modules and functions
• Modular fixtures for managing small or parametrized long-lived test resources
• Can run unittest (or trial), nose test suites out of the box
• Python 3.8+ or PyPy3
• Rich plugin architecture, with over 850+ external plugins and thriving community

I regard pytest as a tool designed specifically for the tests of the Python code. It can run all the test code at once as long as the functions and files name are written in the correct format and output a detailed and helpful report about the error raised by the test functions. It can also be integrated into the workflow. And I think with pytest we can probably get rid of the bash script and write everything in Python.

[gregorgorjanc]

That would be great @XingerTang! We will still need examples script though. How much work is there for this? Can you sketch out an example?

[XingerTang]

@gregorgorjanc I think I can try to start with test 1 of the functional tests of Alpha Peel to see how it works.