oschwengers / bakta

Rapid & standardized annotation of bacterial genomes, MAGs & plasmids
GNU General Public License v3.0
448 stars 55 forks source link

MacOS M2 cannot install bakta using conda. #261

Closed YuLanSky closed 11 months ago

YuLanSky commented 11 months ago

Describe the bug When I was using my old computer (macOS Intel chip), Bakta ran smoothly. However, I recently upgraded to a new computer with an M2 chip and found that I cannot install Bakta using conda.

The reason is that conda cannot find the piler-cr package in any channel.

When I searched for the piler-cr package in the bioconda database, I found that it only supports Linux-64 and OSX-64 (Intel), so it cannot be installed.

My question is: M1/M2 chips have been around for a while now, so I was wondering how others are running Bakta on M1/M2 chip computers? Thank you

(H) ➜  ~ micromamba install bakta
bioconda/osx-arm64                                          Using cache
bioconda/noarch                                             Using cache
conda-forge/osx-arm64                                       Using cache
conda-forge/noarch                                          Using cache
pkgs/main/osx-arm64                                           No change
pkgs/r/osx-arm64                                              No change
pkgs/r/noarch                                                 No change
pkgs/main/noarch                                              No change
error    libmamba Could not solve for environment specs
    The following package could not be installed
    └─ bakta is not installable because it requires
       └─ piler-cr, which does not exist (perhaps a missing channel).
critical libmamba Could not solve for environment specs
oschwengers commented 11 months ago

Hi @YuLanSky , thanks for reaching out and asking. I guess, this is rather a general issue with still lacking support for M1/M2 in Bioconda. Hence, may I kindly refer you to the Bioconda GH repository: https://github.com/bioconda/bioconda-recipes/

Until there is a general upstream fix for that and if your analyses do not depend on CRISPR arrays, you could build your own Conda environment for Bakta w/o Piler-CR and then skip this step in Bakta via --skip-crispr. Since I do not have a Mac do not know about Bioconda's potential M1/M2 support route, I'm very sorry that I cannot provide more support in this case. But I'm pretty sure that many people are interest in that and that Bioconda maintainers can give more answers.

Best regards Oliver