aiida-optimade 0.3.0

Expose an AiiDA database according to the OPTiMaDe specification.

Optimade API implementation for AiiDA

Latest release	Build status	Activity

This is a RESTful API server created with FastAPI that exposes an AiiDA database according to the OPTiMaDe specification.

It is mainly used by Materials Cloud to expose access to archived AiiDA databases through the OPTiMaDe API. But it may be freely implemented by any to fulfill a similar purpose.

The server is based on the test server "template" used in the optimade-python-tools package. Indeed, the filter grammar and parser and pydantic models from optimade-python-tools are used directly here.

Lastly, the server utilizes the FastAPI concept of routers, which means each endpoint can be "setup" several times, allowing multiple base URLs and more flexibility.

Prerequisites

Environment where AiiDA is installed.
AiiDA database containing StructureData nodes, since these are the only AiiDA nodes that are currently exposed with this API (under the /structures endpoint).

Note: At the moment, aiida-optimade works most optimally with an AiiDA database using the SQLAlchemy backend.

Installation

git clone https://github.com/Materials-Consortia/optimade-python-tools
pip install -e optimade-python-tools/
git clone https://github.com/aiidateam/aiida-optimade
pip install -e aiida-optimade/

Running the server locally

# specify AiiDA profile (will use default otherwise)
export AIIDA_PROFILE=optimade
sh run.sh

Navigate to http://127.0.0.1:5000/optimade/info

Running via docker

Adapt profiles/quicksetup.json and profiles/docker-compose.yml appropriately.

docker-compose -f profiles/docker-compose.yml up --build

Navigate to http://127.0.0.1:3253/optimade/info

Stop by using

docker-compose -f profiles/docker-compose.yml down

Design choices

Q: Why create an individual config.json file instead of just mounting an existing .aiida directory and using that directly?
A: This, currently, wouldn't work because the REPOSITORY_URI needs to point to the right path inside the container, not on the host. Furthermore, storing all configurations in the same file can be fragile.