Ioannis Filippidis

Results 195 comments of Ioannis Filippidis

Yet another observation in support of using the imperative mood when titling commit messages: the changes to the LaTeX kernel and to LaTeX packages in the LaTeX news have such...

Thank you for revising the changes, I will review the revised changes.

A note that while reviewing the changes of this pull request, I had encountered issues in a function in the call graph of the function `region_diff` (I think it is...

Following the [previous](https://github.com/tulip-control/polytope/issues/41#issuecomment-279903918) comment and to simplify `__repr__`, I suggest omitting all attributes that are computed automatically from other values. These comprise dependent information (for the sake of the argument...

The alternative sounds better, it has been implemented in 479e9ce4604adf9aec3ed0b41edc65e399f47494 (on branch `overhaul`). It remains to clear memoized values also for `Polytope`, when one of its `ConvexPolytope` objects is modified.

The Trove classifiers for Python 3.7 and 3.8 could be included in `polytope == 0.2.2`, #58.

Added Python 3.7 and 3.8 to test environments in commit 6b855cda9b4cae066c79cc2a10a94e99c2a8df1c and the Trove classifiers for Python 3.7 and 3.8 in commit 215143c9b0712499d25b14cf9bb6c09d66fe4d47.

For memory profiling one candidate to consider using is [`memprof`](https://pypi.python.org/pypi/memprof).

An [interesting comparison](https://scaron.info/blog/linear-programming-in-python-with-cvxopt.html) of `cvxopt` with `pulp`, its default solver, and with `glpk` concludes: ``` The bottom line is: cvxopt_glpk is 2 to 10 times faster than cvxopt, cvxopt_glpk and...

Within the module `setup.py`, `install_requires` contains `numpy >= 1.20.3`, on branch `dev`. The version `numpy == 1.20.3` [supports Python 3.7](https://pypi.org/project/numpy/1.20.3/). The earliest version of Python that is tested by the...