enzbus
enzbus
Is it (relatively) easy to do? In the simplest case, a primal infeasibility certificate is a dual vector y with non-zero values on the constraints that cause infeasibility. Users could...
Even after considerable effort in data cleaning developed for 1.2.0, the London strategy still has problems with data quality. This is slowing down development of other non-US strategies, assuming (maybe...
Today's sp500 execution failed because Yahoo Finance didn't update the price of one symbol. DownloadedMarketData canceled the whole last row of returns, since it applies a dropna on it. The...
Some individual test modules enforce `warnings='error'` in `unittest.main` but that is not done in `__main__.py`. Most warnings have been fixed already, with this we'll get a Github CI failure any...
Hello @tschm and @kasperjo. I've been looking at this repository and your paper and I think it would be easy to write a class that wraps the functionality you provide...
using composition to break forecasters into elementary ones; will add logic to manage "workspace" in BaseForecast so each is evaled only once; need to untangle EMW-MW logic which was too...
Due to CVXPY which was compiled only against the Numpy 1.X ABI, not prepared for the transition, and doesn't publish it in the PyProject requirements. Consider making a `1.3.3` release...
_Rationale_ Only primal infeasibility benefits from managing the problem formulation. Dual infeasibility (unboundedness) is typically due to bugs in the user syntax. Primal infeasibility instead should have an automatic exception...
This defeats the purpose of one of the design decisions of Cvxportfolio: Data management code should be the same, as far as possible, in the context of back-tests and live...
This is a github issue, and there's little that can be done. See: - https://github.com/github/markup/issues/1798 - https://github.com/github/markup/issues/1796 - https://github.com/orgs/community/discussions/113792 - https://github.com/orgs/community/discussions/86715 ...