Experimentation around using an embedded Kuzu graph database to enhance a documentation site. This is leveraging work from DylanShang/kuzu-wasm. A WebAssembly version is not officially supported by the Kuzu team, as yet.
I'm most interested in getting this working in the context of quarto-live. The main experimentation document is docs/demo/kuzu-quarto-live.qmd, which gets rendered as docs/demo/kuzu-quarto-live.html.
This has a basic meta-data scraping and kuzu graph-loading script under scripts. These run as pre-rendering jobs in quarto, as specified here. The various component CSVs and resulting kuzu database get created in a data folder and are pushed as part of the rendered artifacts here. For that corresponding data folder to be available in pyodide, you need to include a resources reference in the kuzu-quarto-live.qmd file.
I was then trying to connect from a pyodide cell to the generated kuzu database, but it seems that these are alternate virtual file-systems that can't interact with eachother, so for now, I regenerated the simple kuzu database in the pyodide cell directly, per here.