Good then it solved with fixing source code,but this can be painful,so look into virtual environment.
Many of these data science stuff like pinkfish which trow together at one time with the use of 40 third-party libraries.
Then some years later someone try to run this in there main Python installation,then the version pain and error starts
So look into virtual environment conda dos a great job here.
Look at this Post where run the whole R-stack(100 of packages) in a conda environment.
Also not running Anaconda,so do Python now comes with virtual environment build in,look at this post.
Many of these data science stuff like pinkfish which trow together at one time with the use of 40 third-party libraries.
Then some years later someone try to run this in there main Python installation,then the version pain and error starts
So look into virtual environment conda dos a great job here.
Look at this Post where run the whole R-stack(100 of packages) in a conda environment.
Also not running Anaconda,so do Python now comes with virtual environment build in,look at this post.