Installation

The installation procedure depends on what version you’d like to install.

Activate e3sm_unified environment

If you have an account on one of the E3SM supported machines (NERSC, Compy, Acme1, LCRC, Cooley, Rhea), you can access e3sm_diags by activating e3sm_unified, which is a conda environment that pulls together Python and other E3SM analysis tools such as e3sm_diags, mpas-analysis, NCO, cdat and processflow.

The paths to e3sm_unified activation scripts are machine dependent:

Compy
source /share/apps/E3SM/conda_envs/load_latest_e3sm_unified_compy.sh
NERSC
source /global/common/software/e3sm/anaconda_envs/load_latest_e3sm_unified_cori-haswell.sh
LCRC
source /lcrc/soft/climate/e3sm-unified/load_latest_e3sm_unified_chrysalis.sh
Cooley
source /lus/theta-fs0/projects/ccsm/acme/tools/e3sm-unified/load_latest_e3sm_unified_cooley.sh
acme1
source /usr/local/e3sm_unified/envs/load_latest_e3sm_unified_acme1.sh

Change .sh to .csh for csh shells. Note that e3sm_unified’s development cycle is not in phase with e3sm_diags, therefore the version of e3sm_diags included may not be the latest. To install latest stable releases, refer to following:

Installation in a Conda Environment

If the E3SM Unified environment doesn’t serve your needs, you can alternatively install the latest version in your own custom conda environment.

First, activate conda or install it if not available. Details vary on the machine.

Compy

module load anaconda3/2019.03
source /share/apps/anaconda3/2019.03/etc/profile.d/conda.sh

NERSC

module load python/3.7-anaconda-2019.10
source /global/common/cori_cle7/software/python/3.7-anaconda-2019.10/etc/profile.d/conda.sh

Others/Local

If the system doesn’t come with conda pre-installed, follow these instructions:

  1. Download Conda

    Linux
    wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
    
    MacOS
    wget https://repo.anaconda.com/miniconda/Miniconda3-latest-MacOSX-x86_64.sh
    
  2. Install Conda

    Linux
    bash ./Miniconda3-latest-Linux-x86_64.sh
    
    MacOS
    bash ./Miniconda3-latest-MacOSX-x86_64.sh
    
    • Do you wish the installer to initialize Miniconda3 by running conda init? [yes|no] yes

3. If you are working on a machine/network that intercepts SSL communications (such as acme1), you will get an SSL error unless you disable the SSL verification:

conda config --set ssl_verify false
binstar config --set ssl_verify False
  1. Configure Conda channels

    conda config --add channels conda-forge
    conda config --set channel_priority strict
    
  2. Once conda is properly working, you can install the (a) Latest Stable Release or create a (b) Development Environment.

(a) Latest Stable Release

  1. Follow “Others/Local” section for installing Conda.

  2. Get the yml file to create an environment.

    wget https://raw.githubusercontent.com/E3SM-Project/e3sm_diags/main/conda-env/prod.yml
    
  3. Change prefix in that file to be your conda prefix. Typically, this will be ~/miniconda3/envs/e3sm_diags_env.

  4. Remove any cached conda packages. This will ensure that you always get the latest packages

    conda clean --all
    
  5. Use conda to create a new environment with E3SM Diags (e3sm_diags) included.

    • Tip: Add the flag -n <name_of_env> to customize the name of the environment

    conda env create -f conda-env/prod.yml
    conda activate e3sm_diags_env
    

(b) Development Environment

Unlike the latest stable release (i.e., the user environment), the development environment does not include E3SM Diags (e3sm-diags). Instead, the developer will pip install . to build e3sm-diags with changes (see step 6 below).

Note

The dev environment includes quality assurance (QA) tools such as code formatters, linters, and pre-commit. You must use the dev environment for all contributions because these QA tools are enforced using pre-commit checks in the continuous integration/continuous deployment build.

  1. Follow “Others/Local” section for installing conda.

  2. Clone your fork and keep it in sync with the main repo’s main

    # Go to https://github.com/E3SM-Project/e3sm_diags
    # Click "Fork" in the upper right hand corner. This will fork the main repo.
    # Click the green "Code" button
    # Choose the HTTPS or SSH option.
    # (To use the SSH option, you need to have a SSH connection to GitHub set up).
    # Click the clipboard icon to copy the path.
    # On your command line:
    git clone <path>
    git remote -v
    # You should see your fork listed as `origin`
    

    or if you already have a clone of your fork, rebase your fork on the main repo’s main to keep it in sync:

    # Add the main repo as a remote.
    # You can call it anything but "upstream" is recommended.
    # We'll use `<upstream-origin>` here.
    git remote add <upstream-origin> https://github.com/E3SM-Project/e3sm_diags.git
    
    # Fetch all the branches of that remote into remote-tracking branches
    git fetch <upstream-origin>
    
    # Make sure that you're on your main branch:
    git checkout main
    
    # Rewrite your main branch so that any of your commits that
    # aren't already in <upstream-origin>/main are replayed on top of that branch:
    git rebase <upstream-origin>/main
    
    # Push your main branch to your GitHub fork:
    # Note that <fork-origin> should be `origin` if you cloned your fork as above.
    git push -f <fork-origin> main
    

    Checkout a new branch from main.

    git checkout -b <branch-name> main
    
  3. Remove any cached conda packages. This will ensure that you always get the latest packages.

    conda clean --all
    
  4. Enter the fork directory.

    cd e3sm_diags
    
  5. Use conda to create a new dev environment (e3sm_diags is not included in this environment).

    • Tip: Add the flag -n <name_of_env> to customize the name of the environment

    conda env create -f conda-env/dev.yml
    conda activate e3sm_diags_env_dev
    
  6. Install pre-commit.

    pre-commit install
    
  7. Make the desired changes to E3SM Diags, then rebuild and install with:

    pip install .
    
  8. Check that tests pass: ./tests/test.sh. This takes about 4 minutes.

  9. Commit changes and make sure pre-commit checks pass
    git commit -m "..."
    
    pre-commit Output

    pre-commit Output