Get your own local copy of scholia

[WolfgangFahl]

Setting Up a Full Local Copy of Scholia with Wikidata

This issue collects the necessary components for setting up a complete local Scholia installation including a Wikidata mirror.
The text has been assembled with the help of Claude Sonnet 4.5

see #2760 for a concrete installation example

Prerequisites

Recommended Hardware

• CPU: AMD Ryzen (12+ cores recommended)
• RAM: 128 GB minimum
• Storage: 4 TB SSD minimum
• Cost: Approximately 1,700 EUR (including assembly)

Recommended Software

• Operating System: Linux (Ubuntu/Debian)
• Version Control: Git
• Web Server: Apache or Caddy (for reverse proxy)
• Docker: for running services

Needed Libraries

• QLever dependencies
• Python dependencies for Scholia
• WDQS update tools

---

Components

Local Wikidata Query Service

WDQS Updater Setup:

Requirements:

• data.jnl file
• ./launch script

Update Mechanism: WDQS updating server with data.jnl seeding approach

References:

• Phabricator discussion: https://phabricator.wikimedia.org/T403627#11457722
• Implementation: https://github.com/WolfgangFahl/get-your-own-wdqs
• Based on James Hare's solution
• See detailed instructions: https://wiki.bitplan.com/index.php/Wikidata_Import_2025-12-13

Alternative Triple Store - QLever

Repositories:

• QLever Core: https://github.com/ad-freiburg/qlever
• QLever Control: https://github.com/ad-freiburg/qlever-control
• QLever UI: https://github.com/ad-freiburg/qlever-ui

Build and Installation:

• See detailed instructions: https://wiki.bitplan.com/index.php/Wikidata_Import_2025-12-10

** scholia.ini**

[query-server]
sparql_endpoint = https://qlever-api.wikidata.dbis.rwth-aachen.de
sparql_editurl = https://qlever.wikidata.dbis.rwth-aachen.de/wikidata/?query=
sparql_embedurl = 
sparql_endpoint_name = RWTH Aachen QLever instance

3. Scholia Application

Installation:

git clone https://github.com/WDscholia/scholia
cd scholia
git checkout qlever  # Use qlever branch for QLever version

Required Patch:

• Currently requires --port patch from issue --config command line option and port configurable in configuration #2746

Backend Configuration:

The backend needs to be configured to point to your local QLever endpoint. Example configuration (based on RWTH setup):

{
  "sparql_editurl": "http://localhost:7001/wikidata/?query=",
  "sparql_embedurl": "",
  "sparql_endpoint": "http://localhost:7001",
  "sparql_endpoint_name": "Local QLever instance",
  "text_to_topic_q_text_enabled": false,
  "third_parties_enabled": true
}

4. Web Server Configuration

Apache Configuration Example:

File: /etc/apache2/sites-available/scholia.conf

# /etc/apache2/sites-available/scholia.conf
# Apache Configuration for scholia.wikidata.dbis.rwth-aachen.de
# scholia Version 0.3 of 2025-12-12
# HTTP Port: 8100
# SSL Port: 443

<VirtualHost *:443>
    ServerName scholia.wikidata.dbis.rwth-aachen.de
    ServerAdmin [email protected]

    # Your shared SSL settings and certificates
    Include ssl.conf

    # Reverse proxy to the app on 8100
    ProxyPreserveHost On
    RequestHeader set X-Forwarded-Proto "https"
    ProxyPass / http://127.0.0.1:8100/
    ProxyPassReverse / http://127.0.0.1:8100/

    ErrorLog ${APACHE_LOG_DIR}/scholia_error_ssl.log
    CustomLog ${APACHE_LOG_DIR}/scholia_ssl.log combined
</VirtualHost>

<VirtualHost *:80>
    ServerName scholia.wikidata.dbis.rwth-aachen.de
    ServerAdmin [email protected]

    Redirect permanent / https://scholia.wikidata.dbis.rwth-aachen.de/

    ErrorLog ${APACHE_LOG_DIR}/scholia_error.log
    CustomLog ${APACHE_LOG_DIR}/scholia.log combined
</VirtualHost>

Enable the site:

sudo a2ensite scholia.conf
sudo systemctl reload apache2

---

Reference Implementation

The RWTH Aachen setup provides a working reference:

WDQS based

• Scholia: https://scholia.wikidata.dbis.rwth-aachen.de/
• Backend Config: https://scholia.wikidata.dbis.rwth-aachen.de/backend
• ** WDQS ** : https://wdqs.wikidata.dbis.rwth-aachen.de
  JSON version of Backend Configuration:

curl -L -H "Accept: application/json" https://scholia.wikidata.dbis.rwth-aachen.de/backend

Qlever based

• Scholia: https://scholia-qlever.wikidata.dbis.rwth-aachen.de/
• Backend Config: https://scholia-qlever.wikidata.dbis.rwth-aachen.de/backend
• QLever Endpoint: https://qlever-api.wikidata.dbis.rwth-aachen.de

JSON version of Backend Configuration:

curl -L -H "Accept: application/json" https://scholia-qlever.wikidata.dbis.rwth-aachen.de/backend

Summary of Steps

1. Set up WDQS updater with data.jnl file seeding see https://wiki.bitplan.com/index.php/Wikidata_Import_2025-12-13
2. Install and configure QLever using instructions from https://wiki.bitplan.com/index.php/Wikidata_Import_2025-12-10
3. Clone Scholia repository twice - ones with main and once with checkout qlever branch
4. Apply --port patch from issue --config command line option and port configurable in configuration #2746 to run the copies on two differen ports
5. Configure backend to point to local WDQs/QLever endpoint
6. Set up Apache reverse proxy
7. Configure SSL certificates
8. Start services and verify functionality