infer-actively / pymdp

A Python implementation of active inference for Markov Decision Processes
MIT License
419 stars 83 forks source link

Consider shifting from raw print statements to using canonical logging #141

Open scott-carroll-verses-ai opened 3 weeks ago

scott-carroll-verses-ai commented 3 weeks ago

I see that there are a lot of print statements build into the library (e.g. print(f'Factorized version: lnA at time {t}: {lnA}') in pymdp/algos/mmp.py). It would be preferrable to use a canonical logging library. This is a little article making the case for using logging rather than raw print statements: https://www.loggly.com/blog/4-reasons-a-python-logging-library-is-much-better-than-putting-print-statements-everywhere/

conorheins commented 2 weeks ago

Thanks a lot for this suggestion @scott-carroll-verses-ai , will look into this :)