# Data Model

## Canonical Input

The canonical input is one Reanimator hardware JSON snapshot.

The viewer does not own the schema. The Reanimator JSON contract remains external to this repository.

## Viewer Data Model

The viewer uses a presentation model derived directly from the input JSON:

- top-level snapshot metadata
- ordered top-level sections
- table definitions per section
- raw field values

## Rendering Rule

The viewer is schema-preserving, not schema-owning.

- Known sections have a preferred order.
- Known important fields may have a preferred column order.
- Every field present in the input must remain visible.
- Unknown fields must not be dropped.

## Section Model

Section types expected initially:

- singleton object section, for example `board`
- array-of-object section, for example `cpus`
- nested grouped section, for example `sensors.{fans,power,temperatures,other}`

## Column Rules

For array sections:

1. collect the union of keys present across all rows
2. place known fields first
3. append unknown fields after known fields
4. render absent values as empty cells

## Special Field Handling

- `status` is rendered as a colored badge
- arrays such as `mac_addresses` may be rendered as line-separated values or badges inside one cell
- nested values such as `status_history` may be rendered in expandable detail blocks inside one cell

## Persistence

There is no database and no canonical stored state in v1.

The viewer operates on the in-memory snapshot provided by the embedding application or local file input.