Update YARRRML to achieve KG-Construct RML compliance
Promotors: Ben De Meester
Main contact: Ben De Meester
Problem
YARRRML has become one of the most widely used human-friendly syntaxes for writing RML mappings. In practice, many engineers start from YARRRML, convert to RML, and then execute mappings with engines such as RMLMapper. This makes YARRRML a critical gateway technology: if YARRRML lags behind the evolving RML specification, a large part of the ecosystem effectively lags behind as well.
RML itself is currently advancing through the W3C Knowledge Graph Construction Community Group, with clarified semantics, evolving vocabulary, and increasing focus on interoperability and conformance. These changes are essential for long-term stability, but they also create migration pressure on tooling. Today, parts of the YARRRML toolchain still reflect older RML assumptions, which can lead to generated mappings that are not fully aligned with the latest KG-Construct direction.
This is not only a feature gap; it is a standards adoption bottleneck. A standard is only useful if everyday tooling supports it. Without updated YARRRML support, users face friction when moving to newer RML versions, and implementers lose a practical path for validating and exercising new specification features at scale.
Goal
In this thesis, you will design and implement a compliance-focused update of the YARRRML toolchain so that YARRRML authoring can reliably target the latest KG-Construct-aligned RML model. Your work will connect specification evolution to developer reality: ensuring that mappings written in a concise DSL remain translatable into standards-compliant RML without manual patching.
Concretely, you will analyse differences between legacy RML assumptions in YARRRML and the current KG-Construct direction, define migration rules and compatibility behavior, and implement the required parser/translator updates. You will also develop conformance-oriented test coverage, including representative mapping patterns and edge cases, so compliance is measurable and regressions are detectable.
You will evaluate your implementation on correctness (does generated RML match expected semantics) and compatibility (how legacy YARRRML mappings behave under migration). The expected result is a robust, reference-quality YARRRML update that lowers adoption friction for the next generation of RML tooling and directly supports ongoing web standardization efforts.
The implementation scope focuses on the most impactful compliance and migration paths, not on exhaustively redesigning the entire YARRRML stack.