Skip to main content

@an5/orm

The Schema, Code Generator, and Migration Toolkit for the AN5 ecosystem.

Overview

@an5/orm provides core schema parsing, multi-language code generation (TypeScript, Python, .NET C#, Golang, Rust), database introspection (pull), schema deployment (push), and automated database migrations.

Runtime database connections, query building, transactions, and multi-database drivers (MSSQL, PostgreSQL, MySQL, SQLite, Google Sheets) are powered by @an5/adapters.

Features

  • Schema Definition: Define models and relationships using concise .an5 syntax with native database types.
  • Multi-Language Generator: Generate typed clients and entity models for TypeScript, Python, .NET (C#), Golang, Rust, Java, Kotlin, and Swift.
  • Database Introspection (db:pull): Inspect an existing database schema and generate matching .an5 models.
  • Direct Schema Push (db:push): Synchronize .an5 schema definitions directly to the database.
  • Automated Migrations (db:migrate): Diff schema against database, generate up/down migration SQL scripts, apply, rollback, and check migration status.
  • Seeding & Cleanup: Generic runner for project seed scripts (seed.ts/seed.js) and database cleanup.
  • Seamless Adapter Integration: Built-in connection support via @an5/adapters.

Quick Start

Installation

npm install @an5/orm @an5/adapters

Configuration

Create an5Orm.config.js in your project root:

module.exports = {
  // Schema directory (default: 'an5Schema')
  schemaDir: 'an5Schema',

  // Generated client outputs
  outputs: {
    typescript: {
      outputDir: 'an5Client/typescript',
      metadataFile: 'an5Client/typescript/an5Metadata.ts',
    },
    python: {
      outputDir: 'an5Client/python',
      metadataFile: 'an5Client/python/an5_metadata.py',
    },
    dotnet: {
      outputDir: 'an5Client/dotnet',
    },
    golang: {
      outputDir: 'an5Client/golang',
    },
  },

  // Database pull options
  pull: {
    exclude: ['^__', '^sys\\.', '^migrations'],
    preserveRelations: true,
  },
};

Configure your environment variables in .env:

DATABASE_URL=sqlserver://localhost:1433;database=mydb;user=sa;password=Secret123!;encrypt=false

CLI Commands

You can run AN5 commands directly using npx @an5/orm (or npx an5):

# Generate client code from .an5 schemas
npx @an5/orm generate
# or: npx an5 generate

# Push schema directly to database
npx @an5/orm db:push

# Pull database schema into .an5 files
npx @an5/orm db:pull

# Check schema diff
npx @an5/orm db:migrate diff

# Generate new migration script
npx @an5/orm db:migrate generate

# Apply pending migrations
npx @an5/orm db:migrate apply

# Rollback last migration
npx @an5/orm db:migrate rollback

# View migration status
npx @an5/orm db:migrate status

# Run database seed script
npx @an5/orm db:seed

# Clean up extra tables and views
npx @an5/orm db:cleanup

You can also add these commands to your project's package.json scripts:

{
  "scripts": {
    "generate": "an5 generate",
    "db:push": "an5 db:push",
    "db:pull": "an5 db:pull",
    "db:migrate": "an5 db:migrate",
    "db:seed": "an5 db:seed",
    "db:cleanup": "an5 db:cleanup"
  }
}

Then run them with npm run generate, npm run db:push, etc.

Runtime Database Usage (via @an5/adapters)

To connect to your database and perform queries, use @an5/adapters:

import { createAn5Adapter } from '@an5/adapters';

const db = createAn5Adapter({
  connectionString: process.env.DATABASE_URL!,
});

// Connect to database
await db.$connect();

// Model operations
const users = await db.user.findMany({
  where: { email: { contains: '@example.com' } },
  orderBy: { createdAt: 'desc' },
  take: 10,
});

// Create record
const newUser = await db.user.create({
  data: {
    email: 'alex@example.com',
    name: 'Alex',
  },
});

// Transactions
await db.$transaction(async (tx) => {
  await tx.exec('UPDATE Users SET active = 1 WHERE id = @p_0', { p_0: 'u1' });
});

// Disconnect
await db.$disconnect();

Schema Definition (.an5)

Field types are validated against the database you are generating for. The provider comes from the connection string (DATABASE_URL first, then connectionString in an5Orm.config.js): sqlserver:// (and the fallback) is SQL Server, postgres:// PostgreSQL, mysql:// MySQL, sqlite:// SQLite, googlesheets:// Google Sheets. A type the provider does not have stops generation with every offending field listed at once — see PROVIDER_FIELD_TYPES in generator/src/field-types.ts.

model User {
  id        NVARCHAR(1000) @id @default(uuid())
  email     NVARCHAR(255)  @unique
  name      NVARCHAR(255)?
  createdAt DATETIME2      @default(now())
  orders    Order[]

  @@map("users")
}

model Order {
  id        NVARCHAR(1000) @id @default(uuid())
  userId    NVARCHAR(1000)
  total     INT            @default(0)
  user      User           @relation(fields: [userId], references: [id])

  @@map("orders")
}

Views are declared with the same field syntax, so generated clients can type their rows, plus the SELECT that defines them. A view is read-only: writes are not generated, and db:push/db:migrate create the view after the tables it reads — and after any view it reads. db:pull writes a database view back as a view block, with its columns and the SQL body from the catalog. db:cleanup drops views that are no longer in the schema, before the tables they might read.

@@sql holds the SELECT body; a full CREATE VIEW ... AS ... statement pasted from a catalog is reduced to that body, so its header and object name are not repeated in the generated DDL.

Generated clients expose a view through a read-only client — ViewClient in TypeScript, Python, .NET, Java and Kotlin, ViewModelClient in Swift, and ViewClient[T] in Go and Rust. Those types have the read methods only, so a write to a view is rejected at compile time rather than by the database.

view ActiveUser {
  id    NVARCHAR(1000)
  email NVARCHAR(255)

  @@map("active_users")
  @@sql("SELECT id, email FROM users WHERE active = 1")
}

A longer query can be written over several lines with triple quotes; the indentation shared by the SQL lines is removed before it reaches the database:

view ActiveUser {
  id    NVARCHAR(1000)
  email NVARCHAR(255)

  @@map("active_users")
  @@sql("""
    SELECT id, email
    FROM users
    WHERE active = 1
  """)
}

License

MIT

Schema-grounded code requests

The @an5/orm/generator export prepareCodeRequest({ request, projectRoot, schemaPath, language: 'auto' }) returns context_ready with schema models and real generated API references for the caller's AI model. It generates reference files in temporary storage and cleans them afterward. detectCodeLanguage reads language manifests in the selected project directory and rejects ambiguous or missing markers. This API does not call an LLM, execute queries, or write application code.

Metadata

Release files for an5-orm 1.3.6

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for an5-orm 1.3.6
File Size Uploaded Provenance Actions
an5_orm-1.3.6.tar.gz 5.0 kB Without provenance Details for an5_orm-1.3.6.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for an5-orm 1.3.6
File Interpreter ABI Platform Actions
an5_orm-1.3.6-py3-none-any.whl Python 3 none any Details for an5_orm-1.3.6-py3-none-any.whl

Total release size: 9.6 kB

Release files / an5_orm-1.3.6.tar.gz

Download URL an5_orm-1.3.6.tar.gz
Size 5.0 kB
Tags Source
SHA-256 checksum
How to use checksums
e930119cd72d4754a7bfa311c6da8e3b52bde4baeb03c2bb8daa9234b939229e
BLAKE2b-256 checksum
How to use checksums
3a20840b0260611f4b29ef0d3b346ab647203f4c337c26fb315d20f067c94b93
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.15.2

Release files / an5_orm-1.3.6-py3-none-any.whl

Download URL an5_orm-1.3.6-py3-none-any.whl
Size 4.7 kB
Tags Python 3
SHA-256 checksum
How to use checksums
b7773d41d78107048a1d50279bd09089db17eeeb18a3658ac3967993ae5daa73
BLAKE2b-256 checksum
How to use checksums
ef45dce8512c3d282ddf4c45b8194dcaffef10dcdf11dfa11dc03591b2ba9003
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.15.2

Security for an5-orm 1.3.6

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page