How do I get specific information from 'conda info'?-CodePudding

I'm trying to write a script that checks specific information of the conda that's installed on the user's computer.

For example, consider the call of conda info below. For my use, I just need the type of conda (miniforge3), the channel URLs, and the platform.

How would someone go about getting specific information from a data dump like this?

Is there a way to parse this data in a bash script?

If not, is there someway I can call something like conda info.platform and have it echo osx-arm64?

    higgsy@Higgsys-MBP Desktop % conda info

     active environment : None
            shell level : 0
       user config file : /Users/higgsy/.condarc
 populated config files : /Users/higgsy/miniforge3/.condarc
                          /Users/higgsy/.condarc
          conda version : 4.12.0
    conda-build version : not installed
         python version : 3.9.13.final.0
       virtual packages : __osx=12.5=0
                          __unix=0=0
                          __archspec=1=arm64
       base environment : /Users/higgsy/miniforge3  (writable)
      conda av data dir : /Users/higgsy/miniforge3/etc/conda
  conda av metadata url : None
           channel URLs : https://conda.anaconda.org/conda-forge/osx-arm64
                          https://conda.anaconda.org/conda-forge/noarch
          package cache : /Users/higgsy/miniforge3/pkgs
                          /Users/higgsy/.conda/pkgs
       envs directories : /Users/higgsy/miniforge3/envs
                          /Users/higgsy/.conda/envs
               platform : osx-arm64
             user-agent : conda/4.12.0 requests/2.27.1 CPython/3.9.13 Darwin/21.6.0 OSX/12.5
                UID:GID : 501:20
             netrc file : None
           offline mode : False

CodePudding user response：

I solved it with the help of @j_b and @Larme.

If you want to isolate something from conda info, you can pipe it into grep and awk.

For example, the example below finds the line with platform in it with grep and then removes platform : with awk using sub

conda info | grep 'platform' | awk '{sub(/platform.:./,""); print}'
osx-arm64

Similarly, the channel URLs can be found the same way.

conda info | grep 'channel URLs' | awk '{sub(/channel.URLs.:./,""); print}'
https://conda.anaconda.org/conda-forge/osx-arm64

Thanks again, @Larme and @j_b

CodePudding user response：

conda info | mawk '_{exit}_ =!_<(NF=NF)' OFS= FS='^.  platform[ :] '

osx-arm64

CodePudding user response：

Avoid parsing formatted output

It is better not to rely on processing formatted shell output. This particular output from Conda is meant for users to read, not to be programmatically parsed. As such, it is subject to change.

Programmatic formats

JSON

Instead, one can obtain all the same information formatted as JSON with conda info --json. This will give even more information and may be overkill, but it will be reliably formatted. I would go this route if the processing is not happening in shell, but in a full-featured scripting language (e.g., Python, JavaScript).

YAML, with specific queries

Alternatively, the conda config --show command can query an arbitrary set of configuration values. For example, if you wanted just the platform, it goes by subdir:

> conda config --show subdir
subdir: osx-64

or both platform and channels:

> conda config --show subdir channels
subdir: osx-64
channels:
  - conda-forge
  - bioconda
  - defaults

Note that this output is all valid YAML.

Conda "type" is not a thing

The Conda configuration does not have a setting for the installer of origin. Moreover, all Conda installations can effectively be converted to another variant with a matter of package installs and adding settings to .condarc.

For example, I have an old Miniconda3 installation on a machine, but it is essentially a Mambaforge installation, since conda-forge channel is prioritized and mamba is installed. I'm not sure what you plan to do with such a classification, but it is not clearly defined.