Statistics for CVXPY
Ideally, problems specified in CVXPY that are verified as convex are solved every time. In practice, various issues can cause CVXPY to fail to solve a problem. This project tries to provide a mechanism for users of CVXPY to upload problem data for bug reports and for further analysis.
We intend to maintain a suite of problems for which various (versions) of solvers are run against and produce up-to-date benchmarking for optimization problems in the wild.
For users of optimization, this would provide a quick way to see which solvers are robust and which ones are fast. It would provide a quick indicator, much like CI/CD does for software, as to whether one can expect CVXPY to reasonable solve your problem.
For researchers in optimization, this would provide a repository of test problems that are encountered in the real world.
Note that we adopt an opt-in model. If you wish to contribute a failing problem, simply install cvxstats and upload your problem.
A future version may let you create a github issue with CVXPY directly.
Installation
pip install cvxstats
Usage
import cvxstats
import cvxpy
p = ...
cvxstats.upload(p)
Website
Somewhere...