Author ArthurGoldberg
Recipients ArthurGoldberg, Sergey.Kirpichev, THRlWiTi, Trip.Volpe, ajaksu2, alex, amaury.forgeotdarc, barry, belopolsky, brett.cannon, diana, eric.araujo, eric.snow, ethan.furman, flox, nedbat, pgimeno, rhettinger, serhiy.storchaka, terry.reedy, tshepang, vstinner
Date 2018-08-28.22:59:17
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1535497157.27.0.56676864532.issue2506@psf.upfronthosting.co.za>
In-reply-to
Content
I'm another user of Ned's coverage tool. Our team at the Mount Sinai School of Medicine is building tools to model the dynamics of biochemistry inside individual cells. Our short term aims are to better understanding microbiology and model microorganisms so they can be engineered to more effectively produce drugs and do other industrial tasks. Long term, we seek to build genetically personalized models of human cells which can be used to improve the medical care of cancer and other illnesses. We're funded by several agencies of the federal government. Our field is called whole-cell modeling.

We use Python because it provides a wide array of powerful tools we can reuse to reduce our development time, enables us to rapidly prototype software to test and advance our modeling ideas, and is fun to program. Using git, pip, coverage, GitHub, CircleCI, Docker and other tools we've built a robust development environment that enables multiple people to contribute to advancing our tools for whole-cell modeling. We strongly emphasize software engineering because the data we use is large, incomplete and inconsistent, and our models are complex and difficult to train, verify and validate. We want to have a high level of confidence in our tested code so that if we have trouble with a model we can focus on checking the data and understanding the model design. Coverage testing is an important part of our software engineering. We test both line and branch coverage.

While working today on our simulator I found code that should have been fully covered except for a # pragma no cover, but was not fully covered. I reported it to Ned (https://github.com/nedbat/coveragepy/issues/697) who reproduced it in a simpler example and pointed out that this "Add mechanism to disable optimizations" issue contributed to the problem.

I realize that approximately 0.0% of Python users work on whole-cell modeling, which diminishes the importance of this use case. But Python is widely used in computational biomedicine, which represents many more users. Case in point -- I've created and teach a course in Biomedical Software Engineering which uses Python and teaches coverage testing to masters, PhD, and MD/PhD students.

We'd appreciate your help improving Ned's coverage tool. You can learn more about us at http://www.karrlab.org/ and https://github.com/KarrLab.

Regards
Arthur
History
Date User Action Args
2018-08-28 22:59:17ArthurGoldbergsetrecipients: + ArthurGoldberg, barry, brett.cannon, rhettinger, terry.reedy, amaury.forgeotdarc, belopolsky, vstinner, ajaksu2, nedbat, eric.araujo, alex, flox, THRlWiTi, ethan.furman, tshepang, eric.snow, serhiy.storchaka, diana, Trip.Volpe, pgimeno, Sergey.Kirpichev
2018-08-28 22:59:17ArthurGoldbergsetmessageid: <1535497157.27.0.56676864532.issue2506@psf.upfronthosting.co.za>
2018-08-28 22:59:17ArthurGoldberglinkissue2506 messages
2018-08-28 22:59:17ArthurGoldbergcreate