I have a conda environment M in which I had installed scikit-learn. I had done this installation by going to the local scikit repository at path "D" (where D is data/retry_runs/repo_testbed_dir/scikit-learn__scikit-learn__0.22__14706) in my machine and had then run pip install -e ..
Now, when I am installing a new conda environment named N, the scikit-learn being used in N is the one at path D. I have verified this by running pip list in environment N and here is the output:
scikit-learn 0.22.dev0 <Directory D's absolute path>
I am making some unrelated changes in directory D for a different task and do not want it to interfere with the task I am performing using environment N. As a result, I want the scikit-learn installation in environment N to NOT use the one already locally available at directory D but to use the unmodified version of scikit-learn mentioned in the requirements.txt of environment N.
How can I achieve this ? (I want to achieve the same not just for scikit-learn but also for all the dependencies mentioned in the requirements.txt)
Here are some additional details about my system:
Distributor ID: Ubuntu
Description: Ubuntu 20.04.6 LTS
Release: 20.04
Codename: focal