1

I have recently installed Anaconda3 on my Windows machine. I need to install Yellowbrick and followed their instructions on the quickstart page. Conda is not on my system's PATH, according to installation instructions.

From the anaconda prompt, I tried to install yellowbrick but it never goes beyond the 'solving environment' part. I ran the installer command with the --debug flag and it seems to be pointing conflicts on every single of the 408 dependencies associated with it. I set the channel priority as strict according to this answer and nothing changed.

When I try to install yellowbrick (through pip) on my Linux machine, it works without a problem.

I have tried to install plotly the same way and it worked. However, pipenv has the same problems, and it never goes past the 'solving environment` step either.

I tried it on two different machines and the result is the same. One runs Windows 10, the other Windows 11. Please help as I'm clueless what to do next.

EDIT: Error output below.

Collecting package metadata (current_repodata.json): done Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source. Collecting package metadata (repodata.json): done Solving environment: / Found conflicts! Looking for incompatible packages. This can take several minutes. Press CTRL-C to abort.

The "several minutes" mentioned turned into a whole night and counting.

Otávio Monteagudo
  • 696
  • 4
  • 15

1 Answers1

1

My guess is that you are trying to install Yellowbrick in the base Anaconda installation. That almost never goes well because of the huge number of packages already in the base environment. You should create a new environment using Conda and install Yellowbrick in there.

First open the Anaconda Prompt from the Windows Start Menu. Then use the following commands:

First we create a new environment with the packages we want. The environment is name "myenv". You can change that to whatever you want. Here we are asking for pandas, yellowbrick, and umap-learn. All of the dependencies (python, numpy, ect) will be solved by Conda.

conda create -n myenv pandas conda-forge::yellowbrick conda-forge::umap-learn -y

After that finishes, we activate the environment and run all your code from within it.

conda activate myenv

If you want additional packages, such as Jupyter notebooks, you can install them using Conda.

conda install -n myenv jupyter
James
  • 32,991
  • 4
  • 47
  • 70
  • The same problem persists, however, instead of taking an endless amount of time checking each package now it fails fast with an UnsatisfiableError . – Otávio Monteagudo Jul 20 '22 at 09:02
  • 1
    It looks like Yellowbrick is also available in the conda-forge channel (newer version too). I updated my answer above. The first command works for me on Windows 10. – James Jul 20 '22 at 09:09
  • Yep, it worked. I'll mark your answer as the right one. – Otávio Monteagudo Jul 21 '22 at 10:24