Docker Compose · Wikibase · GPL v2

Auxiliary data services
for self-hosted Wikibase

A production-ready Docker Compose stack that adds SPARQL querying, full-text search, and batch editing to an existing Wikibase installation running on a native LAMP stack. No MediaWiki or MariaDB containers included — just the services you are missing.

View on GitHub → Read the docs

Quick start

# Clone and configure
git clone https://github.com/lucamauri/wikibase-data-services
cp template.env .env  # then fill in your values

# Start the stack
docker compose up -d
docker compose ps  # all services should be healthy

What's included

Six containerised services

Each service handles one concern. Together they fill the gap between a bare LAMP Wikibase and a fully featured wiki platform.

WDQS — Blazegraph

wikibase/wdqs:2

SPARQL query engine over your Wikibase entity data. Stores RDF triples and answers graph queries in real time.

WDQS Updater

wikibase/wdqs:2

Polls the Wikibase API for recent changes and pushes them into Blazegraph. Polling interval is configurable via UPDATER_DELAY.

WDQS Frontend

wikibase/wdqs-frontend:2

Browser-based SPARQL query editor. Serves a clean /sparql endpoint URL. Fully brandable via environment variables.

Elasticsearch

wikibase/elasticsearch:1

Full-text search backend for the MediaWiki CirrusSearch extension. Exposed on 127.0.0.1:9200 — not publicly accessible.

QuickStatements

custom build

Batch editing tool for Wikibase, patched for self-hosted operation. Fixes missing mysqli, adds a cron batch runner, and injects OAuth credentials securely.

Traefik

traefik:3

Internal HTTP router. Routes traffic from your host reverse proxy to the correct container by Host header. No TLS — terminated by Apache.

How it fits together

Architecture overview

Designed for deployments where Apache already handles TLS. Traefik stays on loopback and never touches the public internet.

Internet
  │
  ▼
Apache :443  TLS termination · certbot · host-managed
  │
  ▼
Anubis :8923  bot filtering · rate limiting (optional)
  │
  ▼
Apache :8080  internal backend · loopback only
  │
  ▼
Traefik :8880  routes by Host header · 127.0.0.1 only
  │
  ├──► wdqs-frontend   query.your-wiki.org
  └──► quickstatements qs.your-wiki.org

Prerequisites

What you need first

This stack assumes Wikibase is already running. It adds the services Wikibase doesn't include.

  • Docker 24.0 or later
  • Docker Compose V2 (docker compose)
  • Wikibase on a native LAMP stack
  • Apache, Nginx, or Caddy for TLS
  • ~3 GB available RAM for the full stack
  • A public hostname for SPARQL and QuickStatements

Documentation

Guides and references

Step-by-step guides for every part of the setup, from data loading to OAuth registration.

Data loading Initial Blazegraph population from a Wikibase RDF dump Data migration Moving Blazegraph and Elasticsearch to a new server QuickStatements OAuth Registering and approving an OAuth consumer on Wikibase QuickStatements database MariaDB setup for batch processing on the host Elasticsearch setup Connecting MediaWiki CirrusSearch to the container Apache examples VirtualHost configurations for WDQS and QuickStatements Architecture decisions ADRs explaining key design choices and trade-offs Contributing Issues, pull requests, and conventional commits guide