Repository: https://github.com/chrisjsewell/aiida-install
The tutorial is fully reporoducible using a VS Code Deveopment Container. It is run using Linux, but should also be applicable to other Operating Systems (MacOS, Windows Subsystem for Linux, etc).
This tutorial is based on the information at: https://aiida.readthedocs.io/projects/aiida-core/en/latest/intro/get_started.html, where you can find further information. (see also the troubeshooting section: https://aiida.readthedocs.io/projects/aiida-core/en/latest/intro/troubleshooting.html)
This is important, the ensure the environment remains correct, it is recommended for any Python based projects.
Using Python venv is the simplest way to achieve this, as we do here.
You could aso use Conda, which additionaly allows for installing non-python packages like postgres, rabbitmq and even quantumespresso!
conda create -n aiida -c conda-forge aiida-core=1.6 aiida-core.services
python -m venv .venvs/aiida
source .venvs/aiida/bin/activate
pip list
pip install -U pip setuptools wheel
pip install "aiida-core~=1.6.4"
pip list
pip check
When first calling verdi you may note that it may error.
verdi
This is because we first need to call reentry, to update all the plugins.
reentry scan
verdi
verdi status
From the status, you may note that the work path defaults to .aiida in your home directory.
When working on multiple AiiDA projects, it may be helpful to change this, using the AIIDA_PATH environmental variable.
export AIIDA_PATH="/workspaces/aiida-install/.aiida"
verdi status
To add tab completion for the verdi CLI, run:
# for mac users
# autoload -Uz compinit && compinit
eval "$(_VERDI_COMPLETE=source verdi)"
verdi <TAB>
To automate the path and completion setup, we can add these two commands to our environment activation script:
echo "export AIIDA_PATH=\"/workspaces/aiida-install/.aiida\"" >> .venvs/aiida/bin/activate
echo "$(_VERDI_COMPLETE=source verdi)" >> .venvs/aiida/bin/activate
(Note in in VS Code, you can also set "terminal.integrated.env.linux" or in Conda see: https://conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html#setting-environment-variables)
AiiDA uses a number of external services to run. Note these services are not actually required to be on your local system, they are accessed by AiiDA using web-based protocols. For this demonstration though, we shall set them up locally.
To read/write all the data AiiDA uses, the principle service is a "backend" database. AiiDA principally uses PostgreSQL for this purpose.
sudo apt update
sudo apt install postgresql postgresql-server-dev-all postgresql-client rabbitmq-server
# container only command
sudo chown -R vscode:vscode /var/run/postgresql
/usr/lib/postgresql/11/bin/initdb .aiida/database
/usr/lib/postgresql/11/bin/pg_ctl -D .aiida/database -l .aiida/postgres.log start
(set PGDATA environmental variable, to use pg_ctl without -D)
We can now set up our AiiDA profile.
verdi quicksetup can automatically search for the running PostgreSQL service, store the connections details and set up the initial database schema:
verdi quicksetup --help
verdi quicksetup
verdi status
verdi profile show
verdi config list
verdi database summary -v
Alternatively, you can use a YAML configuration file:
verdi quicksetup --config quicksetup.yml
verdi profile list
We now have multiple profiles set up and you can switch between them.
Note YAML configuration files can also be used for setting up computers and codes. See https://github.com/aiidateam/aiida-code-registry for some examples.
This service is only required when you want to run processes (not to access data). RabbitMQ handles storing what processes are running and enables them to persist e.g. if you turn off your computer.
sudo rabbitmq-server -detached start
(see https://aiida.readthedocs.io/projects/aiida-core/en/latest/intro/troubleshooting.html#rabbitmq-installation-unix if you have issues with this.)
verdi status
To handle the actual running of the processes, we need to start one or more "daemon" (a.k.a background) processes. You can start as many as you want, but should limit to the number of actua CPUs on your computer.
verdi daemon start 2
verdi status
verdi daemon status
You can import archives from local files or remote URLs:
verdi archive import "https://archive.materialscloud.org/record/file?filename=HiCond_bands_calculations.aiida&file_id=ee287780-ac04-4b8b-b03c-3cea6e446f9d&record_id=477"
verdi database summary -v
You can create an archive from one or more computers/codes/groups/nodes:
verdi group list -a -A
verdi archive create -G 2 -- myarchive.aiida
verdi archive inspect myarchive.aiida
pip install jupyterlab
mkdir -p ~/.ipython/profile_default/startup/
cp aiida_magic_register.py ~/.ipython/profile_default/startup/
jupyter lab
Now open Example.ipynb and trying running AiiDA.
Note, you can also use Jupyter within VS Code: https://code.visualstudio.com/docs/datascience/jupyter-notebooks