Uplink library not in conda

Anvil Q&A
1

conda install anvil-uplink

cannot find the package. Whilst I can install it in each env I create I wanted it as a default package.

Anyone know how I can install it as such?

(I’m actually using miniconda).

edit
I believe it might be something to do with this : https://conda-forge.org/

2

Sorry to bring up an old thread, but I ran into the same issue today when setting up a project. anvil-uplink isn’t on Conda or Conda-Forge. Most likely because the Anvil team doesn’t want the overhead of managing multiple package channels. But if you are using Conda (as I do) you can specify a separate pip dependency when you create the Conda environment from a environment.yml file. Here is an example:

name: anvil
channels:
- conda-forge
dependencies:
- python>=3.5
- jupyterlab
- pandas
- numpy
- scikit-learn
- pip
- pip:
    - anvil-uplink

I’m sure you already figured this out, but for others searching who use Anaconda or Conda as their package manager this may be helpful.

3 Likes
3

I use anaconda and I install more packages with pip than with conda because conda has only the most common ones.

1 Like
4

I might be misunderstanding, but I use miniconda for my virtual environments (just because it seems quite neat), and I can pip install anvil uplink without any issues or playing with conda’s environment.

Is full fat conda different in that regard?

5

For the purposes of package management conda and miniconda are basically the same and commands and processes are identical. Fat conda just gives you all the data science goodies I spend way too much time on.

I should also note that if you add a pip package in a conda (or miniconda) environment it is tracked and updated along with the other packages. So a conda update --all will update pip packages as well.

1 Like
6

That I didn’t know.