Drive Materials to Ontology

Overview

Maps course documents that already live in Drive onto the ontology as CourseMaterial objects, each linked to its Course by material_of. It reads Drive, it does not write to it.

Nothing in Drive states which course a document serves — these documents sit at root with no per-course folder. The assignment is therefore a modelling choice, and it lives on the canvas in the "Material Rules" node so it can be read and argued with rather than buried in code.

Requires the seven ontology types and the Course objects (run the Odoo to Ontology Feeder first, or materials will upsert with nothing to link to).

Workflow Chain

"Material Rules"[value] ──► "Upsert Course Materials"[rules] ──[summary]──► "Sync Result"

Node Reference

• "Material Rules" (JSON Object) — course code → regex over the document NAME. Rules are evaluated in order, first match wins, so a document lands in exactly one course and the counts sum to the population. Put specific topics before broad ones. Configurable: value.
• "Upsert Course Materials" (Custom Code) — pages Drive, applies the rules, de-duplicates (01 - Foo.pdf and Foo are one material, preferring the row with a file extension), and mints CourseMaterial + material_of. Documents matching no rule are counted and reported, never silently dropped. Configurable: maxItems.
• "Sync Result" (Preview) — shows the per-course breakdown on the canvas.

Execution Strategy

You MUST complete both phases in order: Phase 1 → Phase 2. Do NOT skip either.

Writing to the ontology is a side effect. It is gated once, in Phase 1.

Phase 1: Confirm the mapping

1. applyCanvasData to commit the template, then nodeFinder for ids — preview ids are stale.
2. nodeFinder "Material Rules" and getNodeOutput, or read its value, and show the user the rules in full. They decide the taxonomy, not you.
3. Call requestUserDecision ONCE: "These rules assign each Drive document to a course, first match wins. Apply them?", allowFreeText: false, options proceed / edit / cancel. On edit, ask which rule to change, upsertNode the new value, and ask again. On cancel or cancelled: true, STOP and write nothing.

Phase 1 complete. Now proceed to Phase 2.

Phase 2: Load and report

4. nodeExecutor "Upsert Course Materials", then "Sync Result".
5. getNodeOutput and read summary and perCourse. Report the per-course counts, and say explicitly how many documents matched no rule — an unmatched document is a gap in the taxonomy, not a failure.
6. If linkedCount is lower than materialCount, say which courses were missing: materials were created but have nothing to hang off, which means the feeder has not run.
7. Call verification with the counts, then followUp offering to (a) refine the rules and re-run, (b) open the ontology graph view, (c) run the console.

Configuration Tips

Error Handling