Goal: Demonstrate a basic conda environment creation workflow by creating a Python environment that contains the numpy and pandas packages.
General Process
[]$ module purge []$ module load miniconda3/24.3.0 []$ conda search python []$ conda create -n py_env []$ conda activate py_env (py_env) []$ conda install python=3.12.4 (py_env) []$ python --version Python 3.12.4 (py_env) []$ conda search numpy (py_env) []$ conda install numpy (py_env) []$ python -c "import numpy; print(numpy.__version__)" 1.26.4 (py_env) []$ conda deactivate []$
Search for Packages
[]$ conda search python Loading channels: done # Name Version Build Channel python 2.7.13 hac47a24_15 pkgs/main ... python 3.12.3 h996f2a0_1 pkgs/main python 3.12.4 h5148396_1 pkgs/main
Notice: Although Python version 2 has been deprecated, it is still used for old packages/modules/scripts. You can create a conda environment that provides this old, no longer supported version.
This is how ARCC provides this version on the cluster.
Create an Environment
[]$ conda create -n py_env Channels: - defaults Platform: linux-64 Collecting package metadata (repodata.json): done Solving environment: done ## Package Plan ## environment location: /home/salexan5/.conda/envs/py_env Proceed ([y]/n)? y
Note the location that the environment will be saved to: environment location: /home/salexan5/.conda/envs/py_env
Create an Environment: Proceed
Preparing transaction: done Verifying transaction: done Executing transaction: done # # To activate this environment, use # # $ conda activate py_env # # To deactivate an active environment, use # # $ conda deactivate
Note the command required to activate your environment: conda activate py_env
This will be required every time you wish to use it.
Activate an Environment
Before you can use a conda environment it must be activated.
Every time you want to use it, you must remember to activate it.
[]$ conda activate py_env (py_env) []$
Note: See how the command line prompt has changed once the environment has been activated: (py_env) [...]
This indicates the name of the conda environment that is currently active.
Conda Install a Version of Python
Once a conda environment is active, you can start to use it, and one aspect of this is to install conda packages.
(py_env) []$ conda install python=3.12.4
Note all the addition dependencies/libraries that are being installed into the environment
Lets check the version installed, within are active environment.
(py_env) []$ python --version Python 3.12.4
Conda Install the numpy Package
Let’s first check if numpy is available as a package, and if it is, which versions are available.
(py_env) []$ conda search numpy Loading channels: done # Name Version Build Channel numpy 1.9.3 py27_nomklhbee5d10_3 pkgs/main ... numpy 1.26.4 py39heeff2f4_0 pkgs/main
(py_env) []$ conda install numpy
(Again) Note: See all the addition dependencies/libraries that are being installed into the environment
If no package version is defined, typically it’ll install the latest version.
Lets check the version installed, within are active environment.
(py_env) []$ python -c "import numpy; print(numpy.__version__)" 1.26.4
Packages are being updated on a frequent basis.
When this workshop was created
numpy/1.26.4was the latest version.As of 20240815, version
2.0.1is available, and in a few weeks a newer version will likely be available.If you do not define a version, then the latest will be installed.
Conda Deactivate your Environment
(py_env) []$ conda deactivate []$
Note how the command line prompt has changed, reverting back to before being activated: [...]
The conda environment is no longer activate and can not be used.
[]$ python --version Python 3.12.2 []$ python -c "import numpy; print(numpy.__version__)" Traceback (most recent call last): File "<string>", line 1, in <module> ModuleNotFoundError: No module named 'numpy'
(Re-)Using the Environment
To re-use a conda environment, you must activate it again.
But, you do not have to re-create it. Creation only happens once.
[]$ module purge []$ module load miniconda3/24.3.0 []$ conda activate py_env (py_env) []$ python --version Python 3.12.4 (py_env) []$ python -c "import numpy; print(numpy.__version__)" 1.26.4 (py_env) []$ conda deactivate []$
Note: The conda environment must be activated to use it.
Adding to an Existing Environment
Although we only need to create a conda environment once, we can add/update it every time we use it.
Let’s go back into are existing conda environment and add the pandas package.
[]$ module purge []$ module load miniconda3/24.3.0 []$ conda activate py_env (py_env) []$ conda install pandas (py_env) []$ python py_test.py Python: 3.12.4 | packaged by Anaconda, Inc. | (main, Jun 18 2024, 15:12:24) [GCC 11.2.0] Numpy: 1.26.4 Pandas: 2.2.2 (py_env) []$ conda deactivate
Prev | Workshop Home | Next |