Retrieve Moodle PDFs in Natural Language

[ziyad-m97]

Description

Use existing Moodle connector to fulfill natural language requests for PDFs, rendering them inline in an embedded PDF viewer (no external deeplink).

Tasks

Intent handling:

• Detect “fetch/download PDF” intents; extract course/resource name/time context (“yesterday’s slides”).

Backend:

• Leverage Moodle connector to locate and download the PDF; handle auth/permission/404; stream or signed URL.
• Provide metadata (title, size, last modified) and error codes for UX.

Frontend:

• Inline embedded PDF viewer with loading/progress; download fallback if embed fails.

• Show context (course, resource name, date), error states, and retry.

• Respect Markdown rendering around the viewer if mixed content is shown.

Acceptance Criteria

• Natural language requests fetch the correct PDF and render inline in the app’s viewer.

• No forced external deeplinks; download fallback available on failures.

• Metadata (title/date/size) is displayed alongside the viewer.