For long I have been using miniconda/mambaconda to manage different python
enviroments and R environments. Basically conda does a very good job. Creating
separate envrironments with different versions of the same packages/flavors,
importing/exporting and migrating enrivoments when needed, keeping system clean
and untouched and all without the need for root access. It can be slow when
solving dependencies, which is when mamba comes to save the day. All work
together has been a fairly easy and smooth experience for me. However, the more
packages I have to install in an environment, the more complex the dependency
has became. I started to encounter cases where conda/mamba had no problem
solving dependencies and installing packages, but conflicts were reported when I
tried to actually importing and using packages in a python session. I started to
created separate environments where after creation of a basic environment, I use
pip to install packages I need and avoid using conda/mamba as far as I possible.
In cases I am using right now, pip seems to always work.
