I was just starting to learn Python libraries for data analysis (Numpy, Pandas and Matplotlib), but I came across my first problem - getting Matplotlib to work with virtualenv.
When working with Python, I always create a new virtual environment, get the right version of Python and libraries / dependencies into it. This time, the course required me to use Python3 , so I installed it through HomeBrew .
Task:
- Matplotlib must interact with the OS
- For this to happen, it needs a Python framework structure (system)
- ... which is not possible if it is inside virtualenv, which makes it use the python virtualenv assembly
The workaround that is supposed to be distributed is described in this link , but I'm not sure how to use it (OSX section).
My understanding of the solution:
- get the version of Python that I want to use, install it systematically, NOT in virtualenv
- create virtualenv, get the dependencies I need, this creates a virtual Python build
- somehow trick the system into using virtual dependencies using the Python system build
- this is done using a shell script (?) that seems to modify certain variables in the shell / terminal configuration file
Questions:
- Am I correct with the above explanation for myself?
- What is the right way to do this? from inside virtualenv, from outside ...?
- , Python? virtualenv ?
!