DraftDocumentService

Kind: Service

Source: atloria-monorepo/apps/api/src/document/services/draft-document.service.ts

Single reusable draft→publish backbone entry point: create a state:'DRAFT', version:1 Document with a unique urlId + slug. Every draft producer (the recorder's flow→manual path, grounded AI page assist, and the Wave-3 ticket→content loop) funnels through here so they all land in the SAME review/publish flow (POST documents/:id/publish) — no forked publish paths.

DraftDocumentService is the shared backend entry point for creating new draft documents with state: 'DRAFT', version: 1, and unique urlId and slug values. Recorder flows, grounded AI assistance, and ticket-to-content generation all use this service so every document follows the same review and POST /documents/:id/publish publishing path.

Methods

Dependencies

• PrismaService

Where it refuses work

• DraftDocumentService stops the work with an early return when !audienceSlug.
• DraftDocumentService stops the work with an early return when !clash.

Diagram

mermaid
sequenceDiagram
  participant Producer as Draft Producer
  participant Service as DraftDocumentService
  participant DB as Document Repository
  participant Review as Review/Publish Flow

  Producer->>Service: createDraft(input)
  Service->>Service: Generate unique urlId and slug
  Service->>DB: Create Document<br/>state: DRAFT, version: 1
  DB-->>Service: Draft document
  Service-->>Producer: Draft document ID and metadata

  Producer->>Review: POST /documents/:id/publish
  Review->>DB: Validate and publish draft
  DB-->>Review: Published document

Usage

ts
import { Injectable } from '@nestjs/common';
import { DraftDocumentService } from './services/draft-document.service';

@Injectable()
export class TicketContentService {
  constructor(
    private readonly draftDocumentService: DraftDocumentService,
  ) {}

  async createDraftFromTicket(ticket: {
    title: string;
    content: string;
    workspaceId: string;
    authorId: string;
  }) {
    const draft = await this.draftDocumentService.createDraft({
      title: ticket.title,
      content: ticket.content,
      workspaceId: ticket.workspaceId,
      authorId: ticket.authorId,
    });

    // Return the draft for the standard review/publish workflow.
    return {
      documentId: draft.id,
      state: draft.state,
      publishEndpoint: `/documents/${draft.id}/publish`,
    };
  }
}

AI Coding Instructions

• Route every new draft-producing workflow through DraftDocumentService; do not create documents directly from recorder, AI, or ticket feature code.
• Preserve the initial document invariants: state: 'DRAFT' and version: 1.
• Let the service own unique urlId and slug generation rather than accepting caller-generated values unless the service contract explicitly supports them.
• Do not add alternate publishing paths for generated content; published transitions must continue through POST /documents/:id/publish.
• When adding producer-specific metadata, keep it additive and avoid changing the shared draft/review lifecycle.

Relationships

• DEPENDS_ON → PrismaService

Referenced By

• CaptureService (DEPENDS_ON)
• DraftDocumentModule (MODULE_PROVIDES)
• DraftDocumentModule (MODULE_EXPORTS)
• ManualsInsightsService (DEPENDS_ON)
• TechnicalDocsMcpController (DEPENDS_ON)