search

chrisquince / STRONG

Strain Resolution ON Graphs
MIT License
47 stars 9 forks source link

problem to get vbgmm module in python for concoct_refine #112

Closed yuxiangtan closed 3 years ago

yuxiangtan commented 3 years ago

Dear author,

In the concoct_refine, it need to import vbgmm. But I found it is not available from conda or pip, and it is not mentioned in the installation on how to get it. What can I do now?

Thank you!

Yuxiang

Sebastien-Raguideau commented 3 years ago

Dear Yuxiang,

I'm unsure why you would look at that in particular. During the install, Concoct is installed as part of the STRONG env. vbgmm is part of Concoct and installed together with it.

Your question must result from an issue, an error message? Can you explain why you try to import vbgmm? Did you activate the STRONG env before doing so?

Best, Seb

yuxiangtan commented 3 years ago

Thank you for your reply.

In the manual guideline, there is no Concoct installation. I installed concoct from conda. The error message is failed to import vbgmm in python. However, I could not find a way to install it.

I am assuming the concoct should be installed together with SPAdes in your auto installation?

Best,

Sebastien-Raguideau commented 3 years ago

Dear Yuxiang,

It's hidden but it is there. It is part of the conda env you install with : mamba env create -f conda_env.yaml

Again, the auto installation is just doing the same thing as documented under.

Ok, to help me understand better, you need to tell me what command you used to get that error message. Another question is about the environment. Did you activate the STRONG env before with : conda activate STRONG

If you did try to launch concoct from the correct env and do still get that error message, then it is rather related to issues with concoct install and looking at the corresponding github should help. I had a quick look and some people report that same issue, there. One proposed solution was to install the mkl lib while in the correct env. So that would mean you need something like:

conda activate STRONG
pip install mkl

Let me know how that works.

Best, Seb