@ork-orm/schema-parser

TypeScript-native Prisma Schema Language parser for Ork ORM.

Overview

This package provides a complete TypeScript-native implementation for parsing Prisma Schema Language (PSL) files, replacing the Rust-based schema engine with a lightweight, extensible TypeScript solution.

Features

• Pure TypeScript: No binary dependencies, runs anywhere Node.js runs
• AST Generation: Converts PSL into structured Abstract Syntax Trees
• Code Generation: Generates TypeScript interfaces and types from schemas
• Error Handling: Comprehensive error reporting with source locations
• ESM-only: Modern ES module architecture

Installation

pnpm add @ork-orm/schema-parser

Usage

import { parseSchema } from '@ork-orm/schema-parser'

const schema = `
datasource db {
  provider = "postgresql"
}

model User {
  id    Int     @id @default(autoincrement())
  email String  @unique
  name  String?
}
`

const result = parseSchema(schema)

if (result.errors.length === 0) {
  console.log('Schema parsed successfully:', result.ast)
} else {
  console.error('Parse errors:', result.errors)
}

API Reference

parseSchema(schema: string, options?: ParseOptions): ParseResult

Main parsing function that converts PSL source code into an AST.

Parameters

• schema: The Prisma schema source code as a string
• options: Optional parsing configuration

Returns

• ParseResult containing the AST and any parsing errors

Classes

• Lexer: Tokenizes PSL source code
• Parser: Converts tokens into AST nodes
• CodeGenerator: Generates TypeScript code from AST

Development

This package is part of the Ork ORM project and follows its development conventions:

# Build the package
pnpm build

# Run tests
pnpm test

# Development mode with watching
pnpm dev

Architecture

The parser consists of three main phases:

1. Lexical Analysis (Lexer): Converts source text into tokens
2. Parsing (Parser): Builds AST from tokens
3. Code Generation (CodeGenerator): Generates TypeScript from AST

License

Apache-2.0