Reputation: 299
I would like to install the TensorClus library. To do so, I do
pip install TensorClus
When I do this, however, Python complains, saying
tensorclus depends on numpy 1.18.3, and pandas 1.0.3.
However, when I check to see which versions of these I have via pip show numpy
and pip show pandas
, I'm told that I have 1.22.1 numpy
and pandas
1.2.4. I have Python 3.9, and my libraries seem to be up to date. What might the problem be?
EDIT: My coworker found the answer. Apparently the solution was difficult. When he shares it with me tomorrow, I will post the exact solution since it would presumably benefit others.
Upvotes: 0
Views: 66
Reputation: 94931
The problem with the library is that it specifies the exact versions up to the last component instead of providing reasonable ranges. To install it you need to install these exact versions. The simplest way to do that is to allow pip
to upgrade/downgrade already installed versions:
pip install --upgrade TensorClus
Can be shortened to just
pip install -U TensorClus
PS. And my advice is to always use virtual environments — virtualenv
, venv
, virtualenvwrapper
, pyenv-virtualenv
, whatever.
Upvotes: 1
Reputation:
Have you tried to downgrade "NumPy" and "Pandas" to the required versions?
pip3 install numpy==1.18.3 pip3 install pandas==1.0.3
Upvotes: 0