SchemaEditor

The SchemaEditor component provides an interactive table-based interface for creating and editing JSON schemas.

SchemaEditor belongs to the Advanced React capability profile — a specialized, React-only surface with no Pixi dependency and no separate peer to install.

Purpose

SchemaEditor lets users define data structures by adding properties, choosing supported types and formats, and navigating nested object/array schemas in a table interface.

Key Features

Interactive property editing
Type selection from JSON schema types
Format specification for common patterns
required array tracking per object schema, kept consistent across renames and deletes (see Editing Properties — no in-table toggle control yet)
Array and object type support
Inline editing with validation
Add/remove properties
Read-only mode support

Quick Start

import { useState } from 'react';
import { SchemaEditor } from '@cratis/components/SchemaEditor';
import type { JsonSchema } from '@cratis/components/types';

export function MySchemaEditor() {
    const [schema, setSchema] = useState<JsonSchema>({
        type: 'object',
        properties: {
            name: { type: 'string' },
            age: { type: 'number' },
        },
        required: ['name'],
    });

    return <SchemaEditor schema={schema} onChange={setSchema} />;
}

Basic Schema Structure

The component works with JSON schemas following the JSON Schema specification:

{
    "type": "object",
    "properties": {
        "propertyName": {
            "type": "string",
            "format": "date"
        }
    },
    "required": ["propertyName"]
}

JSON Schema support

SchemaEditor and ObjectContentEditor share one JsonSchema/JsonSchemaProperty contract (@cratis/components/types), and it is a pragmatic authoring subset of JSON Schema, not a general-purpose validator. The supported shape is exactly what the types declare:

title, name, $id, $ref, type, format, description
properties and items (recursively, for object and array)
required (a string array on every object schema, including an object-valued property, naming required children in that same properties object)
definitions, for schemas reused via $ref

The type editor offers string, integer, number, and boolean as leaf types, plus array and object as container types, with the built-in format catalog (DEFAULT_TYPE_FORMATS, also exported from types): guid, date-time, date, and time for string; int16, int32, and int64 for integer; float and double for number.

What is deliberately not supported: JSON Schema composition keywords (oneOf, anyOf, allOf, not), enum/const, numeric or string constraints (minimum, maximum, minLength, maxLength, pattern), additionalProperties/patternProperties, and boolean schemas (true/false in place of a schema object). SchemaEditor does not display, edit, or validate these keywords. Unrelated unknown properties are generally preserved by its clone-and-spread mutation paths, but a targeted type change intentionally rewrites fields such as type, format, items, or properties and may leave richer external constraints inconsistent. Validate edited schemas with the application’s authoritative validator, or use a dedicated editor when richer vocabulary support is required.

Table Interface

Properties are displayed in a table:

Property	Type
name	string
email	email
age	number
createdAt	date-time

The Type column shows the property’s format when one is set, falling back to its JSON type otherwise. There is no separate Required column; see Editing Properties for how required is tracked.