Tuvima.Wikidata.AspNetCore 1.0.0

Tuvima.Wikidata

A .NET library for working with Wikidata and Wikipedia. It matches text to Wikidata entities, fetches structured data, retrieves Wikipedia content, and provides lightweight in-memory entity graph traversal — all with zero external dependencies.

In plain English: You have author names, book titles, or company names. This library figures out which Wikidata item each one refers to, gives you a confidence score, and then lets you enrich your data with everything Wikidata and Wikipedia know about those entities. It can also map relationships between entities — family trees, connection paths, and cross-media appearances — without needing a heavy graph database.

This is the first .NET Wikidata reconciliation library, filling a gap in the ecosystem where only Python and JavaScript implementations previously existed.

Who Is This For?

• Data engineers cleaning and linking datasets to structured identifiers
• App developers building search, autocomplete, or knowledge-powered features
• Library/archive systems matching catalog records to authority files (VIAF, ISNI, LoC)
• Research teams enriching study data with Wikidata's 100M+ items
• Content platforms pulling plot summaries, biographies, or descriptions from Wikipedia
• Media applications traversing entity relationships — family trees, character connections, cross-media appearances
• Anyone who needs to go from messy text to structured, linked data

What Can It Do?

Packages

Installation

dotnet add package Tuvima.Wikidata

Targets: .NET 8.0 (LTS) and .NET 10.0

Dependencies: None beyond System.Text.Json (built into .NET). AOT compatible and trimmable.

Quick Start

using Tuvima.Wikidata;

using var reconciler = new WikidataReconciler();

// Match text to a Wikidata entity
var results = await reconciler.ReconcileAsync("Douglas Adams");

Console.WriteLine(results[0].Id);          // "Q42"
Console.WriteLine(results[0].Name);        // "Douglas Adams"
Console.WriteLine(results[0].Description); // "English author and humourist (1952-2001)"
Console.WriteLine(results[0].Score);       // 100
Console.WriteLine(results[0].Match);       // true (confident auto-match)

using Tuvima.Wikidata.Graph;

// Build an entity graph from your data
var graph = new EntityGraph(nodes, edges);

// Find how two characters are connected
var paths = graph.FindPaths("Q937618", "Q312545");

// Build a family tree
var tree = graph.GetFamilyTree("Q937618", generations: 3);

// Find characters appearing in multiple works
var crossMedia = graph.FindCrossMediaEntities(minWorks: 2);

Features

Reconciliation

Match text to Wikidata entities with dual-search, fuzzy matching, type filtering, property constraints, and confidence scoring.

Reconciliation guide — type filtering, property constraints, batch processing, streaming, scoring breakdown

Entity Data & Wikipedia Content

Fetch structured entity data, Wikipedia summaries and sections, images, revision history, editions, and child entities.

Entity data guide — entity fetching, Wikipedia content, staleness detection, edition discovery, child entities

Graph Traversal

Lightweight in-memory entity graph for pathfinding, family trees, cross-media detection, and subgraph extraction. Pure C# with adjacency lists — no RDF or SPARQL dependencies.

Graph module guide — building graphs, pathfinding, family trees, cross-media entities, subgraphs

ASP.NET Core Integration

Host a W3C Reconciliation Service API compatible with OpenRefine and Google Sheets.

ASP.NET Core guide — DI registration, endpoint mapping, service manifest

Configuration

Tune scoring, concurrency, language, type hierarchy, and HTTP behavior.

Configuration guide — all options, caching, custom HttpClient, custom Wikibase instances

Architecture

The reconciliation pipeline has four stages: dual search, entity fetching, weighted scoring, and type filtering.

Architecture overview — pipeline stages, internal components, design decisions

What's New in v1.0.0

• Renamed from Tuvima.WikidataReconciliation — package name now reflects the library's full scope: reconciliation, entity data, Wikipedia content, and graph traversal
• Graph module — new Tuvima.Wikidata.Graph namespace with EntityGraph for in-memory entity graph traversal (pathfinding, family trees, cross-media detection, neighbor lookup, subgraph extraction). Zero dependencies, AOT compatible.
• v1.0.0 stability — signals production readiness after 10 minor versions of real-world use

See the changelog for the full version history.

Acknowledgements

The reconciliation algorithms in this library (dual-search strategy, scoring formula, fuzzy matching approach, type checking, and property matching) are based on openrefine-wikibase by Antonin Delpeuch, licensed under the MIT License.

Antonin Delpeuch. "A survey of OpenRefine reconciliation services." arXiv:1906.08092

The configurable Wikibase endpoint support was informed by the nfdi4culture fork.

The graph module's use cases (pathfinding, family trees, cross-media entity detection) were originally implemented using dotNetRDF in a consuming application. This module provides a lightweight, dependency-free reimplementation of those specific operations.

This is an independent C# implementation. No code was copied from the original projects. See the NOTICE file for full attribution details.

License

MIT. See LICENSE for the full text.