Skip to content

Anaconda

Anaconda is a distribution of the Python language aimed for scientific computing (data science, machine learning applications, large-scale data processing, predictive analytics, etc.). It aims to simplify package management and deployment. The distribution includes data-science packages suitable for Windows, Linux, and macOS.

Managing enviroments

Enviroments info

$ conda env list
or 
$ conda info --envs

Create a new enviroment

Create an environment: 

$ conda create --name venv_name

Create an environment with a specific version of Python: 

$ conda create -n venv_name python=3.6

To activate this environment: 

$ conda activate venv_name

To deactivate an active environment: 

$ conda deactivate

To remove an environment: 

$ conda remove --name myenv --all

To specify the location of the new enviroment use: 

$ conda create --prefix /tmp/test-env

Create an enviroment from a file

Export your active environment to a new file: 

$ conda env export > environment.yml

Create from environment.yml: 

$ conda env create -f environment.yml
The first line of the yml file sets the new environment's name.

Create a requirements file

To create an Anaconde requirements file use: 

$ conda list -e > requirements.txt
Note: This requirements.txt is not compatible with pip.

The resulting file can be used to create a conda virtual environment with: 

$ conda create --name venv_name --file requirements.txt

However, this output isn't in the right format for pip. If you want a file which you can use to create a pip virtual environment you can install pip within the conda environment, the use pip to create requirements.txt.

$ conda activate venv_name
$ conda install pip
$ pip freeze > requirements.txt

Clone an existing enviroment

For instance, to clone the base enviroment use: 

$ conda create --name venv_name --clone base

Add a folder to the Anaconda path

By adding a folder to the Anaconda path you can call its Python scripts from any location in the Anaconda Prompt. 

cd your_folder
conda-develop .

or 

conda-develop /path/to/module/

Set environment variables

To list any variables: 

$ conda env config vars list

To set environment variables: 

$ conda env config vars set my_var=value

Once you have set an environment variable, you have to reactivate your environment with conda activate venv_name. To check if the environment variable has been set, run echo my_var or conda env config vars list.

Declare environment variables in the environment.yml: 

name: venv_name
channels:
  - conda-forge
  - defaults
dependencies:
  - python=3.7
  - codecov
variables:
  VAR1: valueA
  VAR2: valueB

Rollback Anaconda environment

Rollback Anaconda environment

Sometimes updating your environment can bring unintended consequences. Fortunatly Anaconda allows you to list the changes to your environment and revert to a previous point.

List the history of each change to the current environment:

$ conda list --revisions

You'll get an output like this:

     ...

2020-06-24 12:47:58  (rev 12)
     _anaconda_depends  {2019.03 (defaults/win-64) -> 2020.02 (defaults/win-64)}
     astropy  {4.0 (defaults/win-64) -> 4.0.1.post1 (defaults/win-64)}
     atomicwrites  {1.3.0 (defaults/win-64) -> 1.4.0 (defaults/noarch)}
     autopep8  {1.4.4 (defaults/noarch) -> 1.5.3 (defaults/noarch)}
     backcall  {0.1.0 (defaults/win-64) -> 0.2.0 (defaults/noarch)}
     bcrypt  {3.1.7 (defaults/win-64) -> 3.1.7 (defaults/win-64)}
     beautifulsoup4  {4.8.2 (defaults/win-64) -> 4.9.1 (defaults/win-64)}
     bitarray  {1.2.1 (defaults/win-64) -> 1.2.2 (defaults/win-64)}
     ...
     xz  {5.2.4 (defaults/win-64) -> 5.2.5 (defaults/win-64)}
     yapf  {0.28.0 (defaults/noarch) -> 0.29.0 (defaults/noarch)}
     zipp  {2.2.0 (defaults/noarch) -> 3.1.0 (defaults/noarch)}
     zlib  {1.2.11 (defaults/win-64) -> 1.2.11 (defaults/win-64)}
     zstd  {1.3.7 (defaults/win-64) -> 1.4.4 (defaults/win-64)}
    -backports.os-0.1.1 (defaults/win-64)
    +brotlipy-0.7.0 (defaults/win-64)
    +importlib-metadata-1.6.1 (defaults/win-64)
    +prompt-toolkit-3.0.5 (defaults/noarch)
    +regex-2020.5.14 (defaults/win-64)
    +threadpoolctl-2.1.0 (defaults/noarch)
    +toml-0.10.1 (defaults/noarch)

2020-06-24 14:57:05  (rev 13)
    +plotly-4.8.1 (plotly/noarch)
    +retrying-1.3.3 (defaults/win-64)

2020-11-02 14:49:53  (rev 14)
     _anaconda_depends  {2020.02 (defaults/win-64) -> 2020.07 (defaults/win-64)}
     asn1crypto  {1.3.0 (defaults/win-64) -> 1.4.0 (defaults/noarch)}
     astroid  {2.3.3 (defaults/win-64) -> 2.4.2 (defaults/win-64)}
     astropy  {4.0.1.post1 (defaults/win-64) -> 4.0.2 (defaults/win-64)}
     attrs  {19.3.0 (defaults/noarch) -> 20.2.0 (defaults/noarch)}
     autopep8  {1.5.3 (defaults/noarch) -> 1.5.4 (defaults/noarch)}
     bcrypt  {3.1.7 (defaults/win-64) -> 3.2.0 (defaults/win-64)}
     ...
    +nest-asyncio-1.4.1 (defaults/noarch)
    +tifffile-2020.10.1 (defaults/win-64)
    +typed-ast-1.4.1 (defaults/win-64)
    +zope-1.0 (defaults/win-64)
    +zope.event-4.5.0 (defaults/win-64)
    +zope.interface-5.1.2 (defaults/win-64)

It list each revision along with updated packages (old version -> new version) and newly added packages (the ones with + symbol). Now you can safely rollback to previous versions of your environment by using conda install -–revision revision number. For instance:

conda install --revision 13

References

  • https://stackoverflow.com/a/37008663/5240904
  • https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html