brainy/docs/vfs/VFS_INITIALIZATION.md

VFS Initialization Guide

Quick Start

The Brainy VFS is automatically initialized during brain.init(). No separate initialization needed!

import { Brainy } from '@soulcraft/brainy'

// Create and initialize Brainy
const brain = new Brainy({
  storage: { type: 'filesystem', path: './data' }
})
await brain.init()  // VFS is auto-initialized here!

// Use VFS immediately - it's a property, not a method!
await brain.vfs.writeFile('/test.txt', 'Hello World')
const files = await brain.vfs.readdir('/')

What Changed in v5.1.0?

Before (v4.x and early v5.0.0):

const brain = new Brainy(...)
await brain.init()

const vfs = brain.vfs()     // ❌ Method call
await vfs.init()             // ❌ Separate initialization
await vfs.writeFile(...)

After:

const brain = new Brainy(...)
await brain.init()           // VFS auto-initialized!

await brain.vfs.writeFile(...)  // ✅ Property access, just works!

Key Changes:

1. vfs() → vfs: Method call becomes property access
2. Auto-initialization: VFS initialized during brain.init()
3. Zero complexity: No separate vfs.init() call needed
4. Consistent pattern: VFS treated like any other brain API

Migration from v4.x/v5.0.0

Old Pattern (DEPRECATED):

const vfs = brain.vfs()
await vfs.init()
await vfs.writeFile('/test.txt', 'data')

New Pattern:

// Just remove the () and init() call
await brain.vfs.writeFile('/test.txt', 'data')

Why Auto-Initialization?

VFS stores files as entities and relationships in the same graph as everything else. There's no reason to treat it differently! Auto-initialization:

• ✅ Simpler API: One less step to remember
• ✅ Fewer errors: Can't forget to initialize
• ✅ More intuitive: Property access feels natural
• ✅ Consistent: Matches how other brain APIs work

Complete Example

import { Brainy } from '@soulcraft/brainy'

async function useVFS() {
  // Initialize Brainy
  const brain = new Brainy({
    storage: {
      type: 'filesystem',  // or 'memory', 's3', 'r2'
      path: './brainy-data'
    }
  })
  await brain.init()  // VFS ready after this!

  // Use VFS immediately
  await brain.vfs.writeFile('/readme.txt', 'Welcome to VFS!')
  await brain.vfs.mkdir('/documents')

  const files = await brain.vfs.readdir('/')
  console.log('Files in root:', files.map(f => f.name))

  const content = await brain.vfs.readFile('/readme.txt')
  console.log('File content:', content.toString())
}

useVFS().catch(console.error)

TypeScript Usage

import { Brainy, VirtualFileSystem } from '@soulcraft/brainy'

class FileManager {
  private brain: Brainy

  async initialize(): Promise<void> {
    this.brain = new Brainy({
      storage: { type: 'filesystem' }
    })
    await this.brain.init()
    // VFS is ready! No separate initialization needed
  }

  async writeFile(path: string, content: string): Promise<void> {
    if (!this.brain) {
      throw new Error('FileManager not initialized. Call initialize() first.')
    }
    // Use VFS as property
    await this.brain.vfs.writeFile(path, content)
  }

  async listFiles(path: string): Promise<string[]> {
    const entries = await this.brain.vfs.readdir(path)
    return entries.map(e => e.name)
  }
}

Error Messages

If you see this error:

Brainy not initialized. Call init() first.

It means you tried to use VFS before calling brain.init(). Always initialize Brainy first:

await brain.init()  // Required!
await brain.vfs.writeFile(...)  // Now this works

Snapshot Support

VFS files are ordinary entities, so they participate fully in the 8.0 Db API: a persisted snapshot contains every VFS file, and a snapshot opened with Brainy.load() serves VFS reads at the snapshot's state:

const brain = new Brainy({ storage: { type: 'filesystem', rootDirectory: './data' } })
await brain.init()

// Create files
await brain.vfs.writeFile('/config.json', '{"version": 1}')

// Snapshot the whole store (VFS files included)
const pin = brain.now()
await pin.persist('/backups/with-vfs')
await pin.release()

// Later changes never touch the snapshot
await brain.vfs.writeFile('/config.json', '{"version": 2}')

See Snapshots & Time Travel.

FAQ

Q: Do I need to call vfs.init() anymore?

A: No! VFS is automatically initialized during brain.init() in v5.1.0+.

Q: Why did the API change from vfs() to vfs?

A: VFS uses the same entity/relationship graph as everything else. There's no reason to treat it differently from other brain APIs.

Q: Will my old code break?

A: If you're using brain.vfs() or await vfs.init(), you'll need to update to the new pattern. The migration is simple - just remove the () and init() calls.

Q: Can I still configure VFS?

A: Yes, VFS configuration is passed through brain.init() config. The VFS-specific options are applied during auto-initialization.

Q: Does this work with all storage adapters?

A: Yes! VFS auto-initialization works with both shipped adapters (FileSystem and Memory) and any plugin-provided storage adapter.

Q: What if I need multiple VFS instances?

A: Each Brainy instance has its own VFS. Create multiple Brainy instances if you need multiple VFS instances.

Related Documentation

• VFS Quick Start - 5-minute setup guide
• VFS API Guide - Complete API reference
• Common Patterns - Best practices and patterns
• Snapshots & Time Travel - Backups and point-in-time reads