Add support for YARS-PG serialization (export + optional import)

[domel]

Summary

Please consider adding YARS-PG as a first-class serialization format in graphfaker (initially as an export target; optionally as an import/loader later). YARS-PG is a concise, text-based format designed specifically for property graphs, with a formal grammar and reference implementations. Supporting it would immediately improve interoperability, reproducibility, and downstream tooling integration for datasets generated by graphfaker.

---

Why YARS-PG would benefit graphfaker

1. Interoperability across PG tools and research

   • YARS-PG was designed for publication and exchange of property graphs. Exporting graphfaker outputs to YARS-PG makes it straightforward to share synthetic benchmarks and teaching datasets across tools, languages, and databases that already speak or validate YARS-PG.

2. Bridging PG and RDF/Linked Data ecosystems*

   • There are existing bridges that parse YARS-PG into RDF* datasets. This unlocks RDF/Linked-Data tooling (validation, transformation, indexing) on top of graphfaker outputs without bespoke converters.

3. Plain-text, diff-friendly artifacts (great for Git)

   • YARS-PG files are human-readable and line-diff-friendly, which improves code review, dataset versioning, and reproducible experiments—key for academic and industrial workflows.

4. Feature coverage aligned with property-graph practice

   • The format and grammar capture the core PG concepts (nodes, edges, properties, labels/types, edge properties/annotations), which map naturally from graphfaker’s internal NetworkX graphs.

5. Ready-to-use grammar and parsers

   • A formal ANTLR4 grammar with Python and Java parsers exists. This lowers implementation risk and provides a built-in way to validate exported files as part of CI.

6. Licensing compatibility

   • graphfaker is MIT-licensed; YARS-PG grammar/impl. is BSD-3-Clause and/or CC-BY for docs—compatible for optional dependency or dev-time validator.

---

What is YARS-PG (at a glance)

• A text-based serialization purpose-built for property graphs.
• Comes with a formal grammar (ANTLR4 + EBNF) and reference parsers (Python/Java).
• Designed to be simple, extensible, and platform-independent.
• Validated in a peer-reviewed venue and used in example datasets/benchmarks.

References

• Grammar & implementation (ANTLR4, EBNF, examples, Python/Java parsers): https://github.com/lszeremeta/yarspg
• Academic paper (IEEE Access, 2024): https://ieeexplore.ieee.org/document/10536116/
• Release record (v5.0.0, DOI): https://zenodo.org/records/10676695
• YARS-PG → RDF* bridge (example implementation): https://github.com/gkellogg/yarspg

[denironyx]

Hi @domel ,

Thanks for creating this issue. I will be reviewing sugggestion.. However, do you think you have sometime in the coming weeks to hop on call?

That way we can discuss and understand lift to incorporate this into a networkx graph.. Although, I am currently thinking on supporting graph database engine.. YARS-PG looks like a good option too.

[domel]

Hi @denironyx

Thanks for getting back to me. At the moment, we have quite a lot on our plate, but in a little while we should be able to set up a call to discuss this further. We're also considering contributing directly and preparing a PR to help with the integration.

/CC @lszeremeta

[denironyx]

I understand.

Yes, if you can, do create a PR. And we can take it up from there.

Thanks again @domel