Generate Doclific ERD JSON

Overview

This skill helps you generate the JSON for ERD (Entity Relationship Diagram) components in Doclific MDX documentation.

Instructions

When the user asks you to create an ERD diagram:

1. Gather requirements: Ask the user about their database schema - tables, columns, and relationships.

1. Generate the JSON: Create properly formatted JSON for the tables and relationships attributes.

1. Output the MDX: Provide the complete <ERD> component with the generated JSON.

ERD Component Structure

<ERD tables="[JSON array of tables]" relationships="[JSON array of relationships]"></ERD>

Tables JSON Structure

Each table object must have:

{
	"id": "unique-table-id",
	"type": "tableNode",
	"data": {
		"name": "table_name",
		"columns": [
			{
				"name": "column_name",
				"dataType": "varchar",
				"isPrimaryKey": false,
				"isNullable": true
			}
		]
	},
	"position": {
		"x": 50,
		"y": 50
	}
}

Table fields:

• id (string): Unique identifier for the table (e.g., "t1", "users-table")
• type (string): The type of node, must be "tableNode"
• data (object): The data for the table
• data.name (string): The actual table name in the database
• data.columns (array): Array of column definitions
• position (object): The position of the table
• position.x (number): Horizontal position for diagram layout (start at 50, increment by 300 for each table), maximum 1000, minimum -1000
• position.y (number): Vertical position for diagram layout (typically 50-100), maximum 1000, minimum -1000

Columns JSON Structure

Each column object must have:

{
	"name": "column_name",
	"type": "varchar",
	"primaryKey": false,
	"nullable": true
}

Column fields:

• id (string): Unique identifier for the column (e.g., "c1", "users-id")
• name (string): The column name
• type (string): PostgreSQL data type (see list below)
• primaryKey (boolean): true if this is the primary key
• nullable (boolean): true if the column allows NULL values

Supported data types:

• Numeric: smallint, integer, bigint, decimal, numeric, real, double precision, smallserial, serial, bigserial
• Monetary: money
• Character: char, varchar, text
• Binary: bytea
• Date/Time: date, time, timetz, timestamp, timestamptz, interval
• Boolean: boolean
• Network: inet, cidr, macaddr, macaddr8
• Bit strings: bit, bit varying
• UUID: uuid
• JSON: json, jsonb
• Arrays: array

Relationships JSON Structure

Each relationship object must have:

{
	"table1": "table-name",
	"column1": "column-name",
	"table2": "table-name",
	"column2": "column-name",
	"cardinality": "1:N"
}

Relationship fields:

• table1 (string): Name of the source table (the "one" side)
• column1 (string): Name of the source column (usually the primary key)
• table2 (string): Name of the target table (the "many" side)
• column2 (string): Name of the target column (the foreign key)
• cardinality (string): Cardinality of the relationship (one of "1:1", "1:N", "N:N", "N:1")

Example

See the example.json file in this directory for a complete example.

Workflow

1. User describes their database schema
2. Identify all tables and their columns
3. Identify relationships between tables
4. Generate unique IDs for tables, columns, and relationships
5. Calculate x/y positions (space tables horizontally)
6. Write JSON to temporary file: Create a temporary JSON file containing the tables and relationships arrays in the format described above
7. Validate and encode: Run the validate-and-encode.js script with the temporary file path as an argument:

``bash node skills/generate-doclific-erd-json/validate-and-encode.js <temp-file-path> ` The script will: - Validate the JSON against the zod schema - HTML encode both the tables and relationships JSON arrays - Return a JSON object with tables and relationships` properties containing the encoded strings

1. Output the MDX: Use the returned encoded strings as the tables and relationships props in the <ERD> component:

``mdx <ERD tables="{encodedtablesstring}" relationships="{encodedrelationshipsstring}"></ERD> ``

1. Clean up: Delete the temporary JSON file after use