2. Components

The software stack for a typical FAIR Data Point deployment consists of five components, as depicted in the following image:

Structure Overview

Important

All FAIR Data Point components adhere to Semantic Versioning, using the MAJOR.MINOR.PATCH format. It is recommended to use matching MAJOR.MINOR versions of the FAIRDataPoint and FAIRDataPoint-client components. The PATCH versions do not need to match.

2.1. FAIRDataPoint

The FAIRDataPoint component (fdp) is the core back-end component that handles all the business logic and operations on semantic data. All functionality of this server-side application is exposed through a REST API. Documentation for this API is generated by the fdp itself, and is available at the path /swagger-ui/index.html.

The fdp component can be deployed without a client (API only), but it is usually combined with the FAIRDataPoint-client component described below. It is also possible to combine the fdp component with a custom client.

The FAIRDataPoint component is available as Docker image fairdata/fairdatapoint.

2.2. FAIRDataPoint-client

The FAIRDataPoint-client component (fdp-client) provides a user interface for humans, including a convenient admin interface. The fdp-client is a client-side application that runs in a web browser and communicates with the fdp back-end using the API.

Note that the fdp-client component uses a built-in nginx instance for request routing based on content negotiation:

  • Requests coming from a web browser (Accept: text/html) are assumed to be for human consumption, and are routed to the client-side application, which is served in the form of a static site.

  • Requests with other content-types, such as text/turtle, or those with an explicit format in the url query, such as ?format=ttl, are assumed to be for machine-readable data, and are routed directly to the API.

The FAIRDataPoint-client component is available as Docker image fairdata/fairdatapoint-client.

2.3. Third-party components

The FAIR Data Point relies on several third-party components

MongoDB

The FAIRDataPoint component depends on a MongoDB database to store application data, i.e. data required for the application to function properly, such as settings, user accounts, roles, metadata schemas, and so on.

This component is available as Docker image mongo.

Triple Store

The FAIRDataPoint component depends on a triple store to store its semantic data. A variety of triple stores are supported, as described in Triple Stores configuration. If no external triple store is configured, the FDP will fall back on an in-memory triple store.

For example, GraphDB is available as Docker image ontotext/graphdb.

Reverse Proxy

In a production deployment, there is usually a reverse proxy that handles HTTPS certificates (SSL/TLS), so the connection to the FAIR Data Point is secured. See production deployment to learn how to configure one.

For example, nginx is available as Docker image nginx.