Showcase

Data visualizer

Tables that live inside the document: authored as plain SemText (sem-data-visualizer, sem-data-source, sem-data-join, sem-data-query — see the tag reference ), rendered and wired by the reading bundle. Sort any column, switch views, edit the query — everything below is the bundle's own chrome over authored markup, served from this site at /dataviz/semtext-reading.js.

Inline source, named views

One visualizer, one inline Markdown table: per-column typed sort on every header, an aria-pressed view switch for the declared views (all, operators), and the expand / copy-link chrome — the container id makes it deep-linkable (#dv=dv-fleet&view=…) and the modal opens the same rendered view full-page.

Files, a declarative join, the editable query

One visualizer, the whole story: two buckets fetched from files next to this page (CSV and YAML), a sem-data-join composing them on the named key column, and the editable default query reading the joined bucket — the results re-run as you type (debounced), and the edited statement deep-links too (#dv=dv-ops&q=…). The authored query text is never rewritten: edits are session-only.

With scripts off both sources read as pre-wrapped authored text — the document is the record; the chrome is presentation.