Zotero RDF Server

This server loads multiple Zotero libraries into an RDF graph, exposes a local SPARQL endpoint, and allows exporting the graph. A visual query builder is found in /explorer to explore the graph or go to GitHub Pages.

Why this Tool?

While Zotero offers robust functionality for storing and collaboratively managing cloud-hosted libraries, it lacks support for federated access and cross-library exploration or search. This Zotero RDF Server is an initial attempt to fill that gap. It implements basic entity mapping (e.g., tags, creators), but remains tightly constrained by Zotero’s inherently textual data model and API structure. A logical next step would be to implement a knowledge base mapping layer to enable richer semantic interoperability.

One Thing You Need: Zotero!

Features

This app provides a web API for working with RDF data, based on your Zotero libraries.

• 🔁 Export RDF data from the store or individual named graphs (TRiG, Turtle, N-Triples, JSON-LD...)
• 💾 Create and restore full store backups
• 📥 Import/export RDF from/to CSV (with smart triple mapping)
• 📝 Convert RDF blocks in Zotero Notes ↔ RDF graphs
• 🧩 Parse semantic notes written in enhanced HTML as Zotero notes
• 🔐 API config supports multiple libraries with sync settings
• 🧪 Query and inspect your store via OpenAPI (see /docs)
• ⚙️ Generate static RDF exports via GitHub Actions — see local_z2rdf.py for how it works

Built with FastAPI and Oxigraph

Mapping

Zotero only provides strings — but some fields deserve more: Creators, Places, Tags, Publishers, etc. are better modeled as entities, not just literals. This app tries to detect when identical or similar values already exist and links them accordingly.

Parse Notes

As a plugin, you can parse your HTML Zotero notes with the Semantic-HTML package (Docs). It is only loaded if the trigger is set in the config.yaml or called via /parse_notes in the API. The results are parsed as RDF and loaded to the store. A mapping example for the RDF parsing is defined in app/parser/mapping.json and can be specified in config.yaml for each library

Configuration

Place both YAML filenames in your .env (example in .env.backup), not in the code or Dockerfile. Docker-Compose will mount these files into /app (or any other directory set in .env) and your Python code loads them via os.getenv(...) with sensible defaults (config.yaml and zotero.yaml).

config.yaml

Defines server and storage settings, see app/config.yaml as an example with comments.

zotero.yaml

Contains the Zotero-specific settings, see app/zotero.yaml as an example with comments.

Running

Locally

bash pip install -r requirements.txt python src/zotero_rdf_server/main.py

You can even run it as a GitHub Action for RDF export, see local/local_z2rdf.py and .github/workflows/rdf_export.yml

Docker

bash docker-compose up --build -d

API

Visit / when the app is running for the Swagger UI or /redoc for alternative OpenAPI documentation.

A static documentation (OpenAPI) is found in docs/openapi.json and HTML

License

MIT License