🗄 DataHub Implementation Guide

Overview

DataHub is the RABS platform's flexible data management system, providing a structured workspace → board → table → row/column hierarchy for organising operational data. It supports multiple layout modes, inline editing, threaded comments, and integration with the Reggie context system.

Conceptual Model

Workspaces
    └── Boards
          └── Tables
                ├── Columns (shared schema per board)
                └── Rows
                      └── Cells (column + row intersection)

Workspaces

Workspaces are the highest-level grouping, representing major operational domains (e.g., Participants, Operations, HR). Each workspace is rendered by datahub_workspace.html and serves as a navigation container for its boards.

Boards

Each workspace contains multiple boards. Boards define the shared column schema, context bucket (used for Reggie integration), and layout mode.

Attribute	Description
tables	Collection of tables within the board
columns	Shared column schema — all tables in a board share the same columns
context bucket	Reggie context integration identifier

Board Layouts

Three layout modes are available:

Layout	Description
Standard tables	Full table view with all columns visible
Compact tables	Condensed rows, reduced column width
Featured cards	Card-based layout highlighting key fields

Tables

Tables hold rows and inherit the board's column schema. Individual tables can override the board's context bucket for Reggie integration. Columns within a table are reorderable, hideable, and support width presets.

Column and Cell Types

Type	Description
text	Free-text input
number	Numeric value
bool	Boolean toggle
date	Date picker
json	Structured JSON data
dropdown	Single-select dropdown
dropdown_multi	Multi-select dropdown
status	Status badge with colour coding
file	File attachment

Row Editing Workflow

Slide-Over Panel

Clicking a row opens a slide-over panel with two sections:

Section	Content
Top	Editable fields for all row cells
Bottom	Threaded comments

When a row has more than 4 comments, the panel switches to split mode — side-by-side editor and comments view.

Inline Quick Edit

Status and dropdown badges support inline quick editing directly from the table view. The edit writes via the updateRow endpoint with the following payload shape:

{
  cells: {
    columnKey: {
      value: "single value"      // for single-select types
      // OR
      values: ["val1", "val2"]   // for multi-select types
    }
  }
}

Comments

Row comments are loaded via authenticated API call:

GET /api/v1/datahub/rows/:rowId/comments
Authorization: Bearer <token>

Comment avatars are resolved via:

GET /api/v1/profile-photo/serve/resolved

Database Schema

All DataHub tables reside in the datahub schema:

Table	Purpose
datahub.workspaces	Workspace definitions and metadata
datahub.boards	Board configuration, layout mode, context bucket
datahub.tables	Table definitions within boards
datahub.columns	Column schema — type, order, visibility, width
datahub.rows	Row data containers
datahub.cells	Cell values at column × row intersections
datahub.row_comments	Threaded comments attached to rows

Key Files

File	Purpose
admin/src/js/pages/datahub_board.js	Primary board logic — rendering, editing, CRUD
admin/src/html/datahub_board_*.html	Three layout wrapper templates (standard, compact, featured)
admin/src/html/datahub_workspace.html	Workspace landing page

UX Design Standards

The DataHub interface follows these design conventions:

Element	Specification
Visual aesthetic	Frosted participant style
Status chips	Squared corners, 0.35rem border radius
Inline edit	Quick-edit menus for status and dropdown badges
Row editor	Slide-over panel with split mode at >4 comments
Context labels	Display "Reggie's Context" badge when context bucket is active

Build

DataHub is part of the admin frontend and is built with Vite:

npm --prefix admin run build