chaosinflesh / hokohoko Goto Github PK

The research code-base underwent a significant refactor under time pressure, which left the unit tests and documentation out-of-date.

When this project first launched, it was envisaged that it would provide the backing data (7 years of per-minute data across 50 symbols at the time). Alas, I no longer host the data - in fact, there may no longer exist a copy of it anywhere), and it was stored in a numpy NPZ file. This significantly reduces the usefulness of this library, as a current researcher would have to hack the code to use it.

So this needs to be updated to use a standard data format - and pandas seems a good target.