generate-docx

name	generate-docx
description	Generate a Word document from ingested source documents using the agnostic Document tree (CommonMark/Pandoc-aligned)
model	deepseek/deepseek-v4-pro
tier	pro
inputs	[{"name":"source_document_ids","type":"list[str]","required":true,"description":"Prisma Document.id values to ground generation in"},{"name":"brief","type":"str","required":true,"description":"What the user wants generated"},{"name":"style","type":"str","required":false,"description":"Tone/style hint (e.g. \"clinical\", \"casual\", \"technical\")"},{"name":"audience","type":"str","required":false,"description":"Target reader description"},{"name":"target_length","type":"str","required":false,"description":"short | medium | long (default: medium)"}]
outputs	[{"name":"file_path","type":"str"},{"name":"file_url","type":"str"},{"name":"title","type":"str"},{"name":"node_count","type":"int"},{"name":"model","type":"str"},{"name":"source_document_ids","type":"list[str]"}]

generate-docx

Generates a Microsoft Word (.docx) document from one or more ingested source documents stored in the OfficePlane database.

What it does

1. Loads the requested source documents from Postgres via Prisma (chapters, sections, summaries).
2. Prompts DeepSeek v4-flash to produce a agnostic Document JSON tree aligned to the CommonMark / Pandoc AST vocabulary — NOT the legacy modules → lessons → blocks schema.
3. Parses the JSON into a Document dataclass tree via parse_document.
4. Renders the tree to .docx bytes via render_docx.
5. Writes the result to /data/workspaces/<job_id>/output.docx.

Expected JSON shape

The LLM must return a strict JSON object conforming to the agnostic Document schema:

{
  "type": "document",
  "schema_version": "1.0",
  "meta": {"title": "My Document", "language": "en"},
  "children": [
    {
      "type": "section",
      "id": "s1",
      "level": 1,
      "heading": "Introduction",
      "children": [
        {"type": "paragraph", "text": "Opening prose here."},
        {
          "type": "list",
          "ordered": false,
          "items": [
            {"type": "paragraph", "text": "First bullet"},
            {"type": "paragraph", "text": "Second bullet"}
          ]
        }
      ]
    }
  ],
  "attributions": [
    {
      "node_id": "s1",
      "document_id": "<source-doc-uuid>",
      "section_id": "<source-section-id>"
    }
  ]
}

Key rules:

• Sections nest recursively. Use level 1..6 to indicate depth.
• Block types: heading, paragraph, list, table, figure, code, callout, quote, divider.
• Never emit "modules" or "lessons" — use nested sections instead.
• Every non-trivial node should have an id (short string).
• Provide at least one attribution per major section pointing back to a source document_id.

Minimal output example

{
  "type": "document",
  "meta": {"title": "BP Primer"},
  "children": [
    {
      "type": "section",
      "id": "intro",
      "level": 1,
      "heading": "Why Measure Blood Pressure",
      "children": [
        {"type": "paragraph", "text": "Early detection of hypertension saves lives."}
      ]
    }
  ],
  "attributions": [
    {"node_id": "intro", "document_id": "doc-abc123"}
  ]
}