feat: add random graph generation and improve metadata vectorization

Introduced `generateRandomGraph` method to create randomized graphs with configurable entity counts and types, supporting automated testing and experimentation. Enhanced metadata vectorization logic to handle various metadata formats reliably, ensuring compatibility and fallback mechanisms during embedding. Updated `README.md` with detailed feature descriptions and usage guides.
This commit is contained in:
David Snelling 2025-06-05 14:04:29 -07:00
parent 93eb73f471
commit e2f65c8146
8 changed files with 1373 additions and 278 deletions

654
README.md
View file

@ -1,211 +1,438 @@
# Soulcraft Brainy
# 🧠 Soulcraft Brainy
A combined Graph and Vector database that runs in browsers and Node.js environments. Brainy provides efficient vector
search capabilities with persistent storage options.
<div align="center">
## What is Brainy?
[![Version](https://img.shields.io/badge/version-0.7.1-blue.svg)](https://www.npmjs.com/package/@soulcraft/brainy)
[![License](https://img.shields.io/badge/license-MIT-green.svg)](LICENSE)
[![Node.js](https://img.shields.io/badge/node-%3E%3D18.0.0-brightgreen.svg)](https://nodejs.org/)
Brainy is a lightweight database that combines:
**A powerful, lightweight vector & graph database for browsers and Node.js**
- Vector search (for similarity-based retrieval)
- Graph relationships (for structured data connections)
- Persistent storage (works across sessions)
</div>
It's designed to work seamlessly in both browser and server environments.
## ✨ Overview
## How It Works
Brainy combines the power of vector search with graph relationships in a lightweight, cross-platform database. Whether you're building AI applications, recommendation systems, or knowledge graphs, Brainy provides the tools you need to store, connect, and retrieve your data intelligently.
Brainy combines three key technologies:
### 🚀 Key Features
1. **Vector Embeddings**: Converts data (text, images, etc.) into numerical vectors that capture semantic meaning
2. **HNSW Algorithm**: Enables fast similarity search through a hierarchical graph structure
3. **Persistent Storage**: Uses the best available storage option for your environment:
- Browser: Origin Private File System (OPFS)
- Node.js: File system
- Server: S3-compatible storage (optional)
- Fallback: In-memory storage
- **Vector Search** - Find semantically similar content using embeddings
- **Graph Relationships** - Connect data with meaningful relationships
- **Cross-Platform** - Works in browsers, Node.js, and server environments
- **Persistent Storage** - Data persists across sessions
- **TypeScript Support** - Fully typed API with generics
- **CLI Tools** - Powerful command-line interface for data management
## Data Model
## 📊 What Can You Build?
Brainy uses a graph-based data model with two primary concepts:
- **Semantic Search Engines** - Find content based on meaning, not just keywords
- **Recommendation Systems** - Suggest similar items based on vector similarity
- **Knowledge Graphs** - Build connected data structures with relationships
- **AI Applications** - Store and retrieve embeddings for machine learning models
- **Data Organization Tools** - Automatically categorize and connect related information
1. **Nouns**: The main entities in your data (nodes in the graph)
- Each noun has a unique ID, vector representation, and metadata
- Nouns can be categorized by type (Person, Place, Thing, Event, Concept, etc.)
- Nouns are automatically vectorized for similarity search
## 🔧 Installation
2. **Verbs**: Relationships between nouns (edges in the graph)
- Each verb connects a source noun to a target noun
- Verbs have types that define the relationship (RelatedTo, Controls, Contains, etc.)
- Verbs can have their own metadata to describe the relationship
## What Can You Use It For?
- **Semantic Search**: Find content based on meaning, not just keywords
- **Recommendation Systems**: Suggest similar items based on vector similarity
- **Knowledge Graphs**: Build connected data structures with relationships
- **Data Organization**: Automatically categorize and connect related information
- **AI Applications**: Store and retrieve embeddings for machine learning models
## Installation
Due to a dependency conflict between TensorFlow.js packages, you need to use the `--legacy-peer-deps` flag when
installing:
Due to a dependency conflict between TensorFlow.js packages, use the `--legacy-peer-deps` flag when installing:
```bash
npm install @soulcraft/brainy --legacy-peer-deps
```
If you encounter an error like this:
```
npm error ERESOLVE unable to resolve dependency tree
npm error Found: @tensorflow/tfjs-core@4.22.0
npm error Could not resolve dependency: peer @tensorflow/tfjs-core@"3.21.0" from @tensorflow/tfjs-converter@3.21.0
```
Use the `--legacy-peer-deps` flag as shown above to resolve it.
## Command Line Interface
Brainy includes a command-line interface (CLI) that allows you to experiment with the API and data directly from the
terminal.
### Using the CLI during development
```bash
# Run the CLI directly from the source
npm run cli help
# Initialize a database
npm run cli init
# Add some data
npm run cli add "Cats are independent pets" '{"noun":"Thing","category":"animal"}'
npm run cli add "Dogs are loyal companions" '{"noun":"Thing","category":"animal"}'
# Search for similar items
npm run cli search "feline pets" 5
# Add relationships between items
npm run cli addVerb <sourceId> <targetId> RelatedTo '{"description":"Both are pets"}'
# View database status
npm run cli status
```
### Installing the CLI globally
```bash
# Install the package globally (with --legacy-peer-deps to resolve TensorFlow.js dependency conflicts)
npm install -g @soulcraft/brainy --legacy-peer-deps
# Now you can use the 'brainy' command directly
brainy help
brainy init
brainy add "Some text" '{"noun":"Thing"}'
```
### Available Commands
- `init` - Initialize a new database
- `add <text> [metadata]` - Add a new noun with the given text and optional metadata
- `search <query> [limit]` - Search for nouns similar to the query
- `get <id>` - Get a noun by ID
- `delete <id>` - Delete a noun by ID
- `addVerb <sourceId> <targetId> <verbType> [metadata]` - Add a relationship between nouns
- `getVerbs <id>` - Get all relationships for a noun
- `status` - Show database status
- `completion-setup` - Setup shell autocomplete for the Brainy CLI
### Autocomplete
The CLI supports autocomplete for commands, noun types, and verb types. To enable autocomplete:
```bash
# Set up autocomplete for your shell
brainy completion-setup
# Restart your shell or source your shell configuration file
# For bash:
source ~/.bashrc
# For zsh:
source ~/.zshrc
```
Once enabled, you can use tab completion for:
- Commands (e.g., `brainy a<tab>` completes to `brainy add`)
- Noun types in metadata (e.g., when adding nouns)
- Verb types (e.g., when adding relationships)
## Basic Usage
## 🏁 Quick Start
```typescript
import {BrainyData, NounType, VerbType} from '@soulcraft/brainy';
import { BrainyData, NounType, VerbType } from '@soulcraft/brainy';
// Create and initialize the database
const db = new BrainyData();
await db.init();
// Add nouns (automatically converted to vectors)
// Add data (automatically converted to vectors)
const catId = await db.add("Cats are independent pets", {
noun: NounType.Thing,
category: 'animal'
noun: NounType.Thing,
category: 'animal'
});
const dogId = await db.add("Dogs are loyal companions", {
noun: NounType.Thing,
category: 'animal'
noun: NounType.Thing,
category: 'animal'
});
// Search for similar nouns
// Search for similar items
const results = await db.searchText("feline pets", 2);
console.log(results);
// Returns nouns similar to "feline pets" with their similarity scores
// Returns items similar to "feline pets" with similarity scores
// Add a verb (relationship) between nouns
// Add a relationship between items
await db.addVerb(catId, dogId, {
verb: VerbType.RelatedTo,
description: 'Both are common household pets'
verb: VerbType.RelatedTo,
description: 'Both are common household pets'
});
```
## 🧩 How It Works
Brainy combines three key technologies:
1. **Vector Embeddings** - Converts data (text, images, etc.) into numerical vectors that capture semantic meaning
2. **HNSW Algorithm** - Enables fast similarity search through a hierarchical graph structure
3. **Persistent Storage** - Uses the best available storage option for your environment:
- Browser: Origin Private File System (OPFS)
- Node.js: File system
- Server: S3-compatible storage (optional)
- Fallback: In-memory storage
## 🚀 The Brainy Pipeline
Brainy's data processing pipeline transforms raw data into searchable, connected knowledge. Here's how the magic happens:
```
Raw Data → Embedding → Vector Storage → Graph Connections → Query & Retrieval
```
### 🔄 Pipeline Stages
1. **Data Ingestion** 🍽️
- Raw text or pre-computed vectors enter the pipeline
- Data is validated and prepared for processing
2. **Embedding Generation** 🧠
- Text is transformed into numerical vectors using embedding models
- Choose between TensorFlow Universal Sentence Encoder (high quality) or Simple Embedding (faster)
- Custom embedding functions can be plugged in for specialized domains
3. **Vector Indexing** 🔍
- Vectors are indexed using the HNSW algorithm
- Hierarchical structure enables lightning-fast similarity search
- Configurable parameters for precision vs. performance tradeoffs
4. **Graph Construction** 🕸️
- Nouns (entities) become nodes in the knowledge graph
- Verbs (relationships) connect related entities
- Typed relationships add semantic meaning to connections
5. **Persistent Storage** 💾
- Data is saved using the optimal storage for your environment
- Automatic selection between OPFS, filesystem, S3, or memory
- Configurable storage adapters for custom persistence needs
### 🧩 Augmentation Types
Brainy uses a powerful augmentation system to extend functionality. Augmentations are processed in the following order:
1. **SENSE** 👁️
- Ingests and processes raw, unstructured data into nouns and verbs
- Handles text, images, audio streams, and other input formats
- Example: Converting raw text into structured entities
2. **MEMORY** 💾
- Provides storage capabilities for data in different formats
- Manages persistence across sessions
- Example: Storing vectors in OPFS or filesystem
3. **COGNITION** 🧠
- Enables advanced reasoning, inference, and logical operations
- Analyzes relationships between entities
- Example: Inferring new connections between existing data
4. **CONDUIT** 🔌
- Establishes high-bandwidth channels for structured data exchange
- Connects with external systems
- Example: Integrating with third-party APIs
5. **ACTIVATION**
- Initiates actions, responses, or data manipulations
- Triggers events based on data changes
- Example: Sending notifications when new data is processed
6. **PERCEPTION** 🔍
- Interprets, contextualizes, and visualizes identified nouns and verbs
- Creates meaningful representations of data
- Example: Generating visualizations of graph relationships
7. **DIALOG** 💬
- Facilitates natural language understanding and generation
- Enables conversational interactions
- Example: Processing user queries and generating responses
8. **WEBSOCKET** 🌐
- Enables real-time communication via WebSockets
- Can be combined with other augmentation types
- Example: Streaming data processing in real-time
### 🌊 Streaming Data Support
Brainy's pipeline is designed to handle streaming data efficiently:
1. **WebSocket Integration** 🔄
- Built-in support for WebSocket connections
- Process data as it arrives without blocking
- Example: `setupWebSocketPipeline(url, dataType, options)`
2. **Asynchronous Processing**
- Non-blocking architecture for real-time data handling
- Parallel processing of incoming streams
- Example: `createWebSocketHandler(connection, dataType, options)`
3. **Event-Based Architecture** 📡
- Augmentations can listen to data feeds and streams
- Real-time updates propagate through the pipeline
- Example: `listenToFeed(feedUrl, callback)`
4. **Threaded Execution** 🧵
- Optional multi-threading for high-performance streaming
- Configurable execution modes (SEQUENTIAL, PARALLEL, THREADED)
- Example: `executeTypedPipeline(augmentations, method, args, { mode: ExecutionMode.THREADED })`
### 🏃‍♀️ Running the Pipeline
The pipeline runs automatically when you:
```typescript
// Add data (runs embedding → indexing → storage)
const id = await db.add("Your text data here", { metadata });
// Search (runs embedding → similarity search)
const results = await db.searchText("Your query here", 5);
// Connect entities (runs graph construction → storage)
await db.addVerb(sourceId, targetId, { verb: VerbType.RelatedTo });
```
Using the CLI:
```bash
# Add data through the CLI pipeline
brainy add "Your text data here" '{"noun":"Thing"}'
# Search through the CLI pipeline
brainy search "Your query here" --limit 5
# Connect entities through the CLI
brainy addVerb <sourceId> <targetId> RelatedTo
```
### 🔧 Extending the Pipeline
Brainy's pipeline is designed for extensibility at every stage:
1. **Custom Embedding** 🧩
```typescript
// Create your own embedding function
const myEmbedder = async (text) => {
// Your custom embedding logic here
return [0.1, 0.2, 0.3, ...]; // Return a vector
};
// Use it in Brainy
const db = new BrainyData({
embeddingFunction: myEmbedder
});
```
2. **Custom Distance Functions** 📏
```typescript
// Define your own distance function
const myDistance = (a, b) => {
// Your custom distance calculation
return Math.sqrt(a.reduce((sum, val, i) => sum + Math.pow(val - b[i], 2), 0));
};
// Use it in Brainy
const db = new BrainyData({
distanceFunction: myDistance
});
```
3. **Custom Storage Adapters** 📦
```typescript
// Implement the StorageAdapter interface
class MyStorage implements StorageAdapter {
// Your storage implementation
}
// Use it in Brainy
const db = new BrainyData({
storageAdapter: new MyStorage()
});
```
4. **Augmentations System** 🧠
```typescript
// Create custom augmentations to extend functionality
const myAugmentation = {
type: 'memory',
name: 'my-custom-storage',
// Implementation details
};
// Register with Brainy
db.registerAugmentation(myAugmentation);
```
## 📝 Data Model
Brainy uses a graph-based data model with two primary concepts:
### Nouns (Entities)
The main entities in your data (nodes in the graph):
- Each noun has a unique ID, vector representation, and metadata
- Nouns can be categorized by type (Person, Place, Thing, Event, Concept, etc.)
- Nouns are automatically vectorized for similarity search
### Verbs (Relationships)
Connections between nouns (edges in the graph):
- Each verb connects a source noun to a target noun
- Verbs have types that define the relationship (RelatedTo, Controls, Contains, etc.)
- Verbs can have their own metadata to describe the relationship
## 🖥️ Command Line Interface
Brainy includes a powerful CLI for managing your data:
```bash
# Install globally
npm install -g @soulcraft/brainy --legacy-peer-deps
# Initialize a database
brainy init
# Add some data
brainy add "Cats are independent pets" '{"noun":"Thing","category":"animal"}'
brainy add "Dogs are loyal companions" '{"noun":"Thing","category":"animal"}'
# Search for similar items
brainy search "feline pets" 5
# Add relationships between items
brainy addVerb <sourceId> <targetId> RelatedTo '{"description":"Both are pets"}'
# Visualize the graph structure
brainy visualize
brainy visualize --root <id> --depth 3
```
### 🔄 Using During Development
```bash
# Run the CLI directly from the source
npm run cli help
# Generate a random graph for testing
npm run cli generate-random-graph --noun-count 20 --verb-count 40
```
### 🔍 Available Commands
#### Basic Database Operations:
- `init` - Initialize a new database
- `add <text> [metadata]` - Add a new noun with text and optional metadata
- `search <query> [limit]` - Search for nouns similar to the query
- `get <id>` - Get a noun by ID
- `delete <id>` - Delete a noun by ID
- `addVerb <sourceId> <targetId> <verbType> [metadata]` - Add a relationship
- `getVerbs <id>` - Get all relationships for a noun
- `status` - Show database status
- `clear` - Clear all data from the database
- `generate-random-graph` - Generate test data
- `visualize` - Visualize the graph structure
- `completion-setup` - Setup shell autocomplete
#### Pipeline and Augmentation Commands:
- `list-augmentations` - List all available augmentation types and registered augmentations
- `augmentation-info <type>` - Get detailed information about a specific augmentation type
- `test-pipeline [text]` - Test the sequential pipeline with sample data
- `-t, --data-type <type>` - Type of data to process (default: 'text')
- `-m, --mode <mode>` - Execution mode: sequential, parallel, threaded (default: 'sequential')
- `-s, --stop-on-error` - Stop execution if an error occurs
- `-v, --verbose` - Show detailed output
- `stream-test` - Test streaming data through the pipeline (simulated)
- `-c, --count <number>` - Number of data items to stream (default: 5)
- `-i, --interval <ms>` - Interval between data items in milliseconds (default: 1000)
- `-t, --data-type <type>` - Type of data to process (default: 'text')
- `-v, --verbose` - Show detailed output
## 🔌 API Reference
### Database Management
```typescript
// Initialize the database
await db.init();
// Clear all data
await db.clear();
// Get database status
const status = await db.status();
```
### Working with Nouns (Entities)
```typescript
// Add a noun (automatically vectorized)
const id = await db.add(textOrVector, {
noun: NounType.Thing,
// other metadata...
});
// Retrieve a noun
const cat = await db.get(catId);
console.log(cat);
// Get all relationships for a noun
const catRelationships = await db.getVerbsBySource(catId);
console.log(catRelationships);
const noun = await db.get(id);
// Update noun metadata
await db.updateMetadata(catId, {
noun: NounType.Thing,
category: 'animal',
size: 'small'
await db.updateMetadata(id, {
noun: NounType.Thing,
// updated metadata...
});
// Delete a noun
await db.delete(dogId);
await db.delete(id);
// Search for similar nouns
const results = await db.search(vectorOrText, numResults);
const textResults = await db.searchText("query text", numResults);
// Search by noun type
const thingNouns = await db.searchByNounTypes([NounType.Thing], numResults);
```
## Key Features
### Working with Verbs (Relationships)
- **Automatic Vectorization**: Converts text and data to vector embeddings
- **Fast Similarity Search**: Uses HNSW algorithm for efficient retrieval
- **Noun-Based Data Model**: Organize entities with typed categorization (Person, Place, Thing, etc.)
- **Verb-Based Relationships**: Create meaningful connections between nouns with typed relationships
- **Cross-Platform**: Works in browsers, Node.js, and server environments
- **Persistent Storage**: Multiple storage options (OPFS, filesystem, cloud)
- **TypeScript Support**: Fully typed API with generics
- **Flexible Configuration**: Customize distance functions, embedding models, and more
```typescript
// Add a relationship between nouns
await db.addVerb(sourceId, targetId, {
verb: VerbType.RelatedTo,
// other metadata...
});
## Advanced Usage
// Get all relationships
const verbs = await db.getAllVerbs();
// Get relationships by source noun
const outgoingVerbs = await db.getVerbsBySource(sourceId);
// Get relationships by target noun
const incomingVerbs = await db.getVerbsByTarget(targetId);
// Get relationships by type
const containsVerbs = await db.getVerbsByType(VerbType.Contains);
// Get a specific relationship
const verb = await db.getVerb(verbId);
// Delete a relationship
await db.deleteVerb(verbId);
```
## ⚙️ Advanced Configuration
### Custom Embedding
```typescript
import {BrainyData, createSimpleEmbeddingFunction} from '@soulcraft/brainy';
import { BrainyData, createSimpleEmbeddingFunction } from '@soulcraft/brainy';
// Use a custom embedding function (faster but less accurate)
const db = new BrainyData({
@ -217,10 +444,10 @@ await db.init();
const vector = await db.embed("Some text to convert to a vector");
```
### Configuration Options
### Performance Tuning
```typescript
import {BrainyData, euclideanDistance} from '@soulcraft/brainy';
import { BrainyData, euclideanDistance } from '@soulcraft/brainy';
// Configure with custom options
const db = new BrainyData({
@ -254,90 +481,19 @@ const db = new BrainyData({
});
```
## API Reference
### Database Methods
```typescript
// Initialize the database
await db.init();
// Database management
await db.clear();
const size = db.size();
const status = await db.status();
```
### Noun Methods (Entities)
```typescript
// Add a noun (automatically vectorized)
const id = await db.add(textOrVector, {
noun: NounType.Thing,
// other metadata...
});
// Retrieve a noun
const noun = await db.get(id);
// Update noun metadata
await db.updateMetadata(id, {
noun: NounType.Thing,
// updated metadata...
});
// Delete a noun
await db.delete(id);
// Search for similar nouns
const results = await db.search(vectorOrText, numResults);
const textResults = await db.searchText("query text", numResults);
// Search by noun type
const thingNouns = await db.searchByNounTypes([NounType.Thing], numResults);
```
### Verb Methods (Relationships)
```typescript
// Add a relationship between nouns
await db.addVerb(sourceId, targetId, {
verb: VerbType.RelatedTo,
// other metadata...
});
// Get all relationships
const verbs = await db.getAllVerbs();
// Get relationships by source noun
const outgoingVerbs = await db.getVerbsBySource(sourceId);
// Get relationships by target noun
const incomingVerbs = await db.getVerbsByTarget(targetId);
// Get relationships by type
const containsVerbs = await db.getVerbsByType(VerbType.Contains);
// Get a specific relationship
const verb = await db.getVerb(verbId);
// Delete a relationship
await db.deleteVerb(verbId);
```
### Distance Functions
## 🧪 Distance Functions
- `cosineDistance` (default)
- `euclideanDistance`
- `manhattanDistance`
- `dotProductDistance`
### Embedding Options
## 🔋 Embedding Options
- Default: TensorFlow Universal Sentence Encoder (high quality)
- Alternative: Simple character-based embedding (faster)
## Extensions
## 🧰 Extensions
Brainy includes an augmentation system for extending functionality:
@ -345,21 +501,10 @@ Brainy includes an augmentation system for extending functionality:
- **Sense Augmentations**: Process raw data
- **Cognition Augmentations**: Reasoning and inference
- **Dialog Augmentations**: Natural language processing
- **Perception Augmentations**: Data interpretation and/or visualization
- **Perception Augmentations**: Data interpretation and visualization
- **Activation Augmentations**: Trigger actions
For detailed documentation on extensions, see the [API docs](https://github.com/soulcraft-labs/brainy/docs).
## Examples
The repository includes several examples:
- Web demo: `examples/demo.html`
- Basic usage: `examples/basicUsage.js`
- Custom storage: `examples/customStorage.js`
- Memory augmentations: `examples/memoryAugmentationExample.js`
## Browser Compatibility
## 🌐 Browser Compatibility
Works in all modern browsers:
@ -370,10 +515,19 @@ Works in all modern browsers:
For browsers without OPFS support, falls back to in-memory storage.
## Requirements
## 📚 Examples
The repository includes several examples:
- Web demo: `examples/demo.html`
- Basic usage: `examples/basicUsage.js`
- Custom storage: `examples/customStorage.js`
- Memory augmentations: `examples/memoryAugmentationExample.js`
## 📋 Requirements
- Node.js >= 18.0.0
## License
## 📄 License
MIT
[MIT](LICENSE)

56
cli-wrapper.js Executable file
View file

@ -0,0 +1,56 @@
#!/usr/bin/env node
/**
* CLI Wrapper Script
*
* This script serves as a wrapper for the Brainy CLI, ensuring that command-line arguments
* are properly passed to the CLI when invoked through npm scripts.
*/
import { spawn } from 'child_process'
import { fileURLToPath } from 'url'
import { dirname, join } from 'path'
import fs from 'fs'
// Get the directory of the current module
const __filename = fileURLToPath(import.meta.url)
const __dirname = dirname(__filename)
// Path to the actual CLI script
const cliPath = join(__dirname, 'dist', 'cli.js')
// Check if the CLI script exists
if (!fs.existsSync(cliPath)) {
console.error(`Error: CLI script not found at ${cliPath}`)
console.error('Make sure you have built the project with "npm run build"')
process.exit(1)
}
// Special handling for version flags
if (process.argv.includes('--version') || process.argv.includes('-V')) {
// Read version directly from package.json to ensure it's always correct
try {
const packageJsonPath = join(__dirname, 'package.json')
const packageJson = JSON.parse(fs.readFileSync(packageJsonPath, 'utf8'))
console.log(packageJson.version)
process.exit(0)
} catch (error) {
console.error('Error loading version information:', error.message)
process.exit(1)
}
}
// Forward all arguments to the CLI script
const args = process.argv.slice(2)
// Check if npm is passing --force flag
// When npm runs with --force, it sets the npm_config_force environment variable
if (process.env.npm_config_force === 'true' && args.includes('clear') && !args.includes('--force') && !args.includes('-f')) {
args.push('--force')
}
const cli = spawn('node', [cliPath, ...args], { stdio: 'inherit' })
cli.on('close', (code) => {
process.exit(code)
})

36
package-lock.json generated
View file

@ -16,12 +16,13 @@
"@tensorflow/tfjs-backend-cpu": "^4.22.0",
"@tensorflow/tfjs-core": "^4.22.0",
"@tensorflow/tfjs-layers": "^4.22.0",
"ascii-tree": "^0.3.0",
"commander": "^14.0.0",
"omelette": "^0.4.17",
"uuid": "^9.0.0"
},
"bin": {
"brainy": "dist/cli.js"
"brainy": "cli-wrapper.js"
},
"devDependencies": {
"@types/jest": "^29.5.3",
@ -2876,12 +2877,13 @@
}
},
"node_modules/@tensorflow/tfjs-converter": {
"version": "4.22.0",
"resolved": "https://registry.npmjs.org/@tensorflow/tfjs-converter/-/tfjs-converter-4.22.0.tgz",
"integrity": "sha512-PT43MGlnzIo+YfbsjM79Lxk9lOq6uUwZuCc8rrp0hfpLjF6Jv8jS84u2jFb+WpUeuF4K33ZDNx8CjiYrGQ2trQ==",
"version": "3.21.0",
"resolved": "https://registry.npmjs.org/@tensorflow/tfjs-converter/-/tfjs-converter-3.21.0.tgz",
"integrity": "sha512-12Y4zVDq3yW+wSjSDpSv4HnpL2sDZrNiGSg8XNiDE4HQBdjdA+a+Q3sZF/8NV9y2yoBhL5L7V4mMLDdbZBd9/Q==",
"license": "Apache-2.0",
"peer": true,
"peerDependencies": {
"@tensorflow/tfjs-core": "4.22.0"
"@tensorflow/tfjs-core": "3.21.0"
}
},
"node_modules/@tensorflow/tfjs-core": {
@ -2932,6 +2934,15 @@
"@tensorflow/tfjs-core": "4.22.0"
}
},
"node_modules/@tensorflow/tfjs/node_modules/@tensorflow/tfjs-converter": {
"version": "4.22.0",
"resolved": "https://registry.npmjs.org/@tensorflow/tfjs-converter/-/tfjs-converter-4.22.0.tgz",
"integrity": "sha512-PT43MGlnzIo+YfbsjM79Lxk9lOq6uUwZuCc8rrp0hfpLjF6Jv8jS84u2jFb+WpUeuF4K33ZDNx8CjiYrGQ2trQ==",
"license": "Apache-2.0",
"peerDependencies": {
"@tensorflow/tfjs-core": "4.22.0"
}
},
"node_modules/@types/babel__core": {
"version": "7.20.5",
"resolved": "https://registry.npmjs.org/@types/babel__core/-/babel__core-7.20.5.tgz",
@ -3451,6 +3462,15 @@
"node": ">=8"
}
},
"node_modules/ascii-tree": {
"version": "0.3.0",
"resolved": "https://registry.npmjs.org/ascii-tree/-/ascii-tree-0.3.0.tgz",
"integrity": "sha512-eC3TvbRQLw+SY0HCVKPXv56JlZ6EeqLOIHT5oHGd0j98ngLDCjTEYbyHYHxrJcb4OwhqhiEo0zYr/FQpj1+Rpw==",
"license": "MIT",
"dependencies": {
"freetree": "0.2.2"
}
},
"node_modules/async": {
"version": "3.2.6",
"resolved": "https://registry.npmjs.org/async/-/async-3.2.6.tgz",
@ -4596,6 +4616,12 @@
"node": ">= 6"
}
},
"node_modules/freetree": {
"version": "0.2.2",
"resolved": "https://registry.npmjs.org/freetree/-/freetree-0.2.2.tgz",
"integrity": "sha512-bZ0U0Qv2THYrnkrD7B4UMAhWnDm3+va6y2mNfReke9HY7PRdvLgEf5Vl1j70P5KffoioFvV0RFRfIBQMWOLpIw==",
"license": "MIT"
},
"node_modules/fs.realpath": {
"version": "1.0.0",
"resolved": "https://registry.npmjs.org/fs.realpath/-/fs.realpath-1.0.0.tgz",

View file

@ -25,10 +25,13 @@
"node": ">=18.0.0"
},
"scripts": {
"prebuild": "node scripts/generate-version.js",
"build": "tsc",
"test": "jest",
"start": "node dist/index.js",
"cli": "node dist/cli.js",
"cli": "node ./cli-wrapper.js",
"cli:version": "node ./cli-wrapper.js --version",
"cli:v": "node ./cli-wrapper.js -V",
"version:patch": "npm version patch",
"version:minor": "npm version minor",
"version:major": "npm version major",
@ -36,7 +39,7 @@
"postinstall": "echo 'Note: If you encounter dependency conflicts with TensorFlow.js packages, please use: npm install --legacy-peer-deps'"
},
"bin": {
"brainy": "dist/cli.js"
"brainy": "cli-wrapper.js"
},
"keywords": [
"vector-database",

View file

@ -628,7 +628,23 @@ export class BrainyData<T = any> {
// If metadata is provided and no vector is provided or forceEmbed is true, vectorize the metadata
if (options.metadata && (!vector || options.forceEmbed)) {
try {
verbVector = await this.embeddingFunction(options.metadata)
// Extract a string representation from metadata for embedding
let textToEmbed: string;
if (typeof options.metadata === 'string') {
textToEmbed = options.metadata;
} else if (options.metadata.description && typeof options.metadata.description === 'string') {
textToEmbed = options.metadata.description;
} else {
// Convert to JSON string as fallback
textToEmbed = JSON.stringify(options.metadata);
}
// Ensure textToEmbed is a string
if (typeof textToEmbed !== 'string') {
textToEmbed = String(textToEmbed);
}
verbVector = await this.embeddingFunction(textToEmbed)
} catch (embedError) {
throw new Error(`Failed to vectorize verb metadata: ${embedError}`)
}
@ -964,6 +980,144 @@ export class BrainyData<T = any> {
}
}
}
/**
* Generate a random graph of data with typed nouns and verbs for testing and experimentation
* @param options Configuration options for the random graph
* @returns Object containing the IDs of the generated nouns and verbs
*/
public async generateRandomGraph(options: {
nounCount?: number; // Number of nouns to generate (default: 10)
verbCount?: number; // Number of verbs to generate (default: 20)
nounTypes?: NounType[]; // Types of nouns to generate (default: all types)
verbTypes?: VerbType[]; // Types of verbs to generate (default: all types)
clearExisting?: boolean; // Whether to clear existing data before generating (default: false)
seed?: string; // Seed for random generation (default: random)
} = {}): Promise<{
nounIds: string[];
verbIds: string[];
}> {
await this.ensureInitialized();
// Check if database is in read-only mode
this.checkReadOnly();
// Set default options
const nounCount = options.nounCount || 10;
const verbCount = options.verbCount || 20;
const nounTypes = options.nounTypes || Object.values(NounType);
const verbTypes = options.verbTypes || Object.values(VerbType);
const clearExisting = options.clearExisting || false;
// Clear existing data if requested
if (clearExisting) {
await this.clear();
}
try {
// Generate random nouns
const nounIds: string[] = [];
const nounDescriptions: Record<string, string> = {
[NounType.Person]: "A person with unique characteristics",
[NounType.Place]: "A location with specific attributes",
[NounType.Thing]: "An object with distinct properties",
[NounType.Event]: "An occurrence with temporal aspects",
[NounType.Concept]: "An abstract idea or notion",
[NounType.Content]: "A piece of content or information",
[NounType.Group]: "A collection of related entities",
[NounType.List]: "An ordered sequence of items",
[NounType.Category]: "A classification or grouping"
};
for (let i = 0; i < nounCount; i++) {
// Select a random noun type
const nounType = nounTypes[Math.floor(Math.random() * nounTypes.length)];
// Generate a random label
const label = `Random ${nounType} ${i + 1}`;
// Create metadata
const metadata = {
noun: nounType,
label,
description: nounDescriptions[nounType] || `A random ${nounType}`,
randomAttributes: {
value: Math.random() * 100,
priority: Math.floor(Math.random() * 5) + 1,
tags: [`tag-${i % 5}`, `category-${i % 3}`]
}
};
// Add the noun
const id = await this.add(metadata.description, metadata as T);
nounIds.push(id);
}
// Generate random verbs between nouns
const verbIds: string[] = [];
const verbDescriptions: Record<string, string> = {
[VerbType.AttributedTo]: "Attribution relationship",
[VerbType.Controls]: "Control relationship",
[VerbType.Created]: "Creation relationship",
[VerbType.Earned]: "Achievement relationship",
[VerbType.Owns]: "Ownership relationship",
[VerbType.MemberOf]: "Membership relationship",
[VerbType.RelatedTo]: "General relationship",
[VerbType.WorksWith]: "Collaboration relationship",
[VerbType.FriendOf]: "Friendship relationship",
[VerbType.ReportsTo]: "Reporting relationship",
[VerbType.Supervises]: "Supervision relationship",
[VerbType.Mentors]: "Mentorship relationship"
};
for (let i = 0; i < verbCount; i++) {
// Select random source and target nouns
const sourceIndex = Math.floor(Math.random() * nounIds.length);
let targetIndex = Math.floor(Math.random() * nounIds.length);
// Ensure source and target are different
while (targetIndex === sourceIndex && nounIds.length > 1) {
targetIndex = Math.floor(Math.random() * nounIds.length);
}
const sourceId = nounIds[sourceIndex];
const targetId = nounIds[targetIndex];
// Select a random verb type
const verbType = verbTypes[Math.floor(Math.random() * verbTypes.length)];
// Create metadata
const metadata = {
verb: verbType,
description: verbDescriptions[verbType] || `A random ${verbType} relationship`,
weight: Math.random(),
confidence: Math.random(),
randomAttributes: {
strength: Math.random() * 100,
duration: Math.floor(Math.random() * 365) + 1,
tags: [`relation-${i % 5}`, `strength-${i % 3}`]
}
};
// Add the verb
const id = await this.addVerb(sourceId, targetId, undefined, {
type: verbType,
weight: metadata.weight,
metadata
});
verbIds.push(id);
}
return {
nounIds,
verbIds
};
} catch (error) {
console.error('Failed to generate random graph:', error);
throw new Error(`Failed to generate random graph: ${error}`);
}
}
}
// Export distance functions for convenience

View file

@ -11,18 +11,15 @@ import { dirname, join } from 'path'
import fs from 'fs'
import { Command } from 'commander'
import omelette from 'omelette'
import { VERSION } from './utils/version.js'
import { sequentialPipeline } from './sequentialPipeline.js'
import { augmentationPipeline, ExecutionMode } from './augmentationPipeline.js'
import { AugmentationType } from './types/augmentations.js'
// Get the directory of the current module
const __filename = fileURLToPath(import.meta.url)
const __dirname = dirname(__filename)
// Import package.json for version info
const packageJson = {
name: '@soulcraft/brainy',
version: '0.6.0',
description: 'A vector database using HNSW indexing with Origin Private File System storage'
}
// Helper function to parse JSON safely
function parseJSON(str: string): any {
try {
@ -70,9 +67,9 @@ const program = new Command()
// Configure the program
program
.name(packageJson.name)
.description(packageJson.description)
.version(packageJson.version)
.name('@soulcraft/brainy')
.description('A vector database using HNSW indexing with Origin Private File System storage')
.version(VERSION, '-V, --version', 'Output the current version')
// Create data directory if it doesn't exist
const dataDir = join(__dirname, '..', 'data')
@ -279,6 +276,323 @@ program
}
})
program
.command('clear')
.description('Clear all data from the database')
.option('-f, --force', 'Skip confirmation prompt', false)
.action(async (options) => {
try {
// Confirm unless --force is used
if (!options.force) {
console.log('WARNING: This will permanently delete ALL data in the database.')
console.log('To proceed without confirmation, use the --force option.')
// Exit without doing anything
console.log('Operation cancelled. No data was deleted.')
console.log('To clear all data, use: brainy clear --force')
return
}
const db = createDb()
await db.init()
await db.clear()
console.log('Database cleared successfully. All data has been removed.')
} catch (error) {
console.error('Error:', (error as Error).message)
process.exit(1)
}
})
program
.command('visualize')
.description('Visualize the graph structure in ASCII format')
.option('-r, --root <id>', 'ID of the root noun to start visualization from')
.option('-d, --depth <number>', 'Maximum depth of the graph to visualize', '2')
.option('-t, --type <type>', 'Filter by noun type')
.option('-l, --limit <number>', 'Maximum number of nodes to display per level', '10')
.action(async (options) => {
try {
const db = createDb()
await db.init()
// Parse options
const depth = parseInt(options.depth, 10)
const limit = parseInt(options.limit, 10)
const rootId = options.root
const nounType = options.type ? resolveNounType(options.type) : undefined
// Get all nouns if no root is specified
if (!rootId && !nounType) {
// Get all nouns (limited by the limit option)
const allNouns = []
let count = 0
// Since there's no direct method to get all nouns, we'll use search with a high limit
const searchResults = await db.search("", 1000, {
forceEmbed: true
})
for (const result of searchResults) {
if (count >= limit) break
allNouns.push(result)
count++
}
if (allNouns.length === 0) {
console.log('No nouns found in the database.')
return
}
console.log(`Graph Overview (showing ${allNouns.length} nouns):\n`)
for (const noun of allNouns) {
// Get outgoing verbs
const outgoingVerbs = await db.getVerbsBySource(noun.id)
// Get incoming verbs
const incomingVerbs = await db.getVerbsByTarget(noun.id)
const nounType = noun.metadata?.noun || 'Unknown'
const label = noun.metadata?.label || noun.id.substring(0, 8)
console.log(`[${nounType}] ${label} (${noun.id})`)
if (outgoingVerbs.length > 0) {
console.log(' Outgoing:')
for (const verb of outgoingVerbs.slice(0, limit)) {
const targetNoun = await db.get(verb.targetId)
const targetLabel = targetNoun?.metadata?.label || verb.targetId.substring(0, 8)
console.log(` --(${verb.metadata?.verb || 'relates to'})--→ [${targetNoun?.metadata?.noun || 'Unknown'}] ${targetLabel}`)
}
if (outgoingVerbs.length > limit) {
console.log(` ... and ${outgoingVerbs.length - limit} more`)
}
}
if (incomingVerbs.length > 0) {
console.log(' Incoming:')
for (const verb of incomingVerbs.slice(0, limit)) {
const sourceNoun = await db.get(verb.sourceId)
const sourceLabel = sourceNoun?.metadata?.label || verb.sourceId.substring(0, 8)
console.log(` ←--(${verb.metadata?.verb || 'relates to'})-- [${sourceNoun?.metadata?.noun || 'Unknown'}] ${sourceLabel}`)
}
if (incomingVerbs.length > limit) {
console.log(` ... and ${incomingVerbs.length - limit} more`)
}
}
console.log('')
}
return
}
// If noun type is specified but no root, show all nouns of that type
if (!rootId && nounType) {
console.log(`Visualizing nouns of type: ${nounType}\n`)
// Search for nouns of the specified type
const searchResults = await db.search("", 1000, {
nounTypes: [nounType],
forceEmbed: true
})
const filteredNouns = searchResults.slice(0, limit)
if (filteredNouns.length === 0) {
console.log(`No nouns found with type: ${nounType}`)
return
}
for (const noun of filteredNouns) {
const label = noun.metadata?.label || noun.id.substring(0, 8)
console.log(`[${nounType}] ${label} (${noun.id})`)
// Get outgoing verbs
const outgoingVerbs = await db.getVerbsBySource(noun.id)
if (outgoingVerbs.length > 0) {
console.log(' Outgoing:')
for (const verb of outgoingVerbs.slice(0, limit)) {
const targetNoun = await db.get(verb.targetId)
const targetLabel = targetNoun?.metadata?.label || verb.targetId.substring(0, 8)
console.log(` --(${verb.metadata?.verb || 'relates to'})--→ [${targetNoun?.metadata?.noun || 'Unknown'}] ${targetLabel}`)
}
if (outgoingVerbs.length > limit) {
console.log(` ... and ${outgoingVerbs.length - limit} more`)
}
}
console.log('')
}
return
}
// If root is specified, visualize the graph starting from that root
if (rootId) {
const rootNoun = await db.get(rootId)
if (!rootNoun) {
console.error(`Root noun with ID ${rootId} not found`)
return
}
console.log(`Visualizing graph from root: ${rootNoun.metadata?.label || rootId}\n`)
// Use a breadth-first search to visualize the graph
const visited = new Set<string>()
const queue: Array<{ id: string; level: number; path: string }> = [
{ id: rootId, level: 0, path: '' }
]
while (queue.length > 0) {
const { id, level, path } = queue.shift()!
if (visited.has(id) || level > depth) {
continue
}
visited.add(id)
const noun = await db.get(id)
if (!noun) {
console.warn(`Noun with ID ${id} not found`)
continue
}
const nounType = noun.metadata?.noun || 'Unknown'
const label = noun.metadata?.label || id.substring(0, 8)
// Print the current noun with proper indentation
console.log(`${' '.repeat(level * 2)}${path}[${nounType}] ${label} (${id})`)
// Get outgoing verbs
const outgoingVerbs = await db.getVerbsBySource(id)
// Add target nouns to the queue for the next level
let verbCount = 0
for (const verb of outgoingVerbs) {
if (verbCount >= limit) {
console.log(`${' '.repeat((level + 1) * 2)}... and ${outgoingVerbs.length - limit} more`)
break
}
const verbType = verb.metadata?.verb || 'relates to'
queue.push({
id: verb.targetId,
level: level + 1,
path: `--(${verbType})--→ `
})
verbCount++
}
}
}
} catch (error) {
console.error('Error:', (error as Error).message)
process.exit(1)
}
})
program
.command('generate-random-graph')
.description('Generate a random graph of data with typed nouns and verbs for testing')
.option('-n, --noun-count <number>', 'Number of nouns to generate', '10')
.option('-v, --verb-count <number>', 'Number of verbs to generate', '20')
.option('-c, --clear', 'Clear existing data before generating', false)
.option('-t, --noun-types <types>', 'Comma-separated list of noun types to use')
.option('-r, --verb-types <types>', 'Comma-separated list of verb types to use')
.action(async (options) => {
try {
const db = createDb()
await db.init()
// Parse options
const nounCount = parseInt(options.nounCount, 10)
const verbCount = parseInt(options.verbCount, 10)
const clearExisting = options.clear
// Parse noun types if provided
let nounTypes: NounType[] | undefined
if (options.nounTypes) {
const typeNames = options.nounTypes.split(',').map((t: string) => t.trim())
nounTypes = typeNames.map((name: string) => {
// Try to match by key name (case insensitive)
const key = Object.keys(NounType).find(
k => k.toLowerCase() === name.toLowerCase()
)
if (key) return NounType[key as keyof typeof NounType]
// If not found by key, check if it's a valid value
if (Object.values(NounType).includes(name as NounType)) {
return name as NounType
}
console.warn(`Warning: Unknown noun type "${name}", ignoring`)
return null
}).filter(Boolean) as NounType[]
}
// Parse verb types if provided
let verbTypes: VerbType[] | undefined
if (options.verbTypes) {
const typeNames = options.verbTypes.split(',').map((t: string) => t.trim())
verbTypes = typeNames.map((name: string) => {
// Try to match by key name (case insensitive)
const key = Object.keys(VerbType).find(
k => k.toLowerCase() === name.toLowerCase()
)
if (key) return VerbType[key as keyof typeof VerbType]
// If not found by key, check if it's a valid value
if (Object.values(VerbType).includes(name as VerbType)) {
return name as VerbType
}
console.warn(`Warning: Unknown verb type "${name}", ignoring`)
return null
}).filter(Boolean) as VerbType[]
}
console.log(`Generating random graph with ${nounCount} nouns and ${verbCount} verbs...`)
if (clearExisting) {
console.log('Clearing existing data first...')
}
const result = await db.generateRandomGraph({
nounCount,
verbCount,
nounTypes,
verbTypes,
clearExisting
})
console.log('Random graph generated successfully!')
console.log(`Created ${result.nounIds.length} nouns and ${result.verbIds.length} verbs`)
// Print some sample IDs
if (result.nounIds.length > 0) {
console.log('\nSample noun IDs:')
result.nounIds.slice(0, 3).forEach(id => console.log(`- ${id}`))
if (result.nounIds.length > 3) {
console.log(`... and ${result.nounIds.length - 3} more`)
}
}
if (result.verbIds.length > 0) {
console.log('\nSample verb IDs:')
result.verbIds.slice(0, 3).forEach(id => console.log(`- ${id}`))
if (result.verbIds.length > 3) {
console.log(`... and ${result.verbIds.length - 3} more`)
}
}
console.log('\nUse the search, get, or visualize commands to explore the generated graph')
} catch (error) {
console.error('Error:', (error as Error).message)
process.exit(1)
}
})
// Add examples to help text
program.addHelpText('after', `
Examples:
@ -286,6 +600,11 @@ Examples:
$ brainy add "Cats are independent pets" '{"noun":"Thing","category":"animal"}'
$ brainy search "feline pets" --limit 5
$ brainy addVerb id1 id2 RelatedTo '{"description":"Both are pets"}'
$ brainy clear --force
$ brainy generate-random-graph --noun-count 20 --verb-count 30 --clear
$ brainy generate-random-graph --noun-types Person,Thing --verb-types RelatedTo,Owns
$ brainy visualize --type Thing --limit 10
$ brainy visualize --root id1 --depth 3
`)
// Setup autocomplete
@ -308,16 +627,23 @@ completion.tree({
'delete',
'getVerbs',
'status',
'clear',
'visualize',
'generate-random-graph',
'completion-setup',
'init',
'help'
'help',
'list-augmentations',
'augmentation-info',
'test-pipeline',
'stream-test'
],
// Command-specific completions
add: {
// For the second argument of 'add' command (metadata)
_: () => {
// Generate templates for each noun type
return getNounTypes().map(type =>
return getNounTypes().map(type =>
`{"noun":"${type}","category":"example"}`
)
}
@ -339,6 +665,51 @@ completion.tree({
delete: {},
getVerbs: {},
status: {},
clear: {
_: () => [
'--force'
]
},
'generate-random-graph': {
_: () => [
'--noun-count 10',
'--verb-count 20',
'--clear',
`--noun-types ${getNounTypes().join(',')}`,
`--verb-types ${getVerbTypes().join(',')}`
]
},
'list-augmentations': {},
'augmentation-info': {
_: () => [
'sense',
'memory',
'cognition',
'conduit',
'activation',
'perception',
'dialog',
'websocket'
]
},
'test-pipeline': {
_: () => [
'--data-type text',
'--mode sequential',
'--mode parallel',
'--mode threaded',
'--stop-on-error',
'--verbose'
]
},
'stream-test': {
_: () => [
'--count 5',
'--interval 1000',
'--data-type text',
'--verbose'
]
},
'completion-setup': {},
init: {},
help: {}
@ -354,6 +725,326 @@ if (process.argv.includes('--completion-setup')) {
process.exit(0)
}
// Pipeline and Augmentation Commands
program
.command('list-augmentations')
.description('List all available augmentation types and registered augmentations')
.action(async () => {
try {
// Initialize the pipeline
await augmentationPipeline.initialize()
// Get available augmentation types
const availableTypes = augmentationPipeline.getAvailableAugmentationTypes()
console.log('Available Augmentation Types:')
if (availableTypes.length === 0) {
console.log(' No augmentation types available')
} else {
availableTypes.forEach(type => {
const augmentations = augmentationPipeline.getAugmentationsByType(type)
console.log(`\n${type.toUpperCase()} (${augmentations.length} registered):`)
if (augmentations.length === 0) {
console.log(' No augmentations registered for this type')
} else {
augmentations.forEach(aug => {
console.log(` - ${aug.name}: ${aug.description}`)
console.log(` Status: ${aug.enabled ? 'Enabled' : 'Disabled'}`)
})
}
})
}
// Show WebSocket augmentations separately
const webSocketAugs = augmentationPipeline.getWebSocketAugmentations()
console.log('\nWebSocket-Enabled Augmentations:')
if (webSocketAugs.length === 0) {
console.log(' No WebSocket-enabled augmentations available')
} else {
webSocketAugs.forEach(aug => {
console.log(` - ${aug.name}: ${aug.description}`)
console.log(` Status: ${aug.enabled ? 'Enabled' : 'Disabled'}`)
})
}
} catch (error) {
console.error('Error:', (error as Error).message)
process.exit(1)
}
})
program
.command('test-pipeline')
.description('Test the sequential pipeline with sample data')
.argument('[text]', 'Sample text to process through the pipeline', 'This is a test of the Brainy pipeline')
.option('-t, --data-type <type>', 'Type of data to process', 'text')
.option('-m, --mode <mode>', 'Execution mode (sequential, parallel, threaded)', 'sequential')
.option('-s, --stop-on-error', 'Stop execution if an error occurs', false)
.option('-v, --verbose', 'Show detailed output', false)
.action(async (text, options) => {
try {
// Initialize the pipeline
await sequentialPipeline.initialize()
console.log(`Processing data: "${text}"`)
console.log(`Data type: ${options.dataType}`)
console.log(`Execution mode: ${options.mode}`)
console.log(`Stop on error: ${options.stopOnError}`)
console.log()
// Set execution mode
let executionMode = ExecutionMode.SEQUENTIAL
switch (options.mode.toLowerCase()) {
case 'parallel':
executionMode = ExecutionMode.PARALLEL
break
case 'threaded':
executionMode = ExecutionMode.THREADED
break
default:
executionMode = ExecutionMode.SEQUENTIAL
}
// Process the data
const result = await sequentialPipeline.processData(
text,
options.dataType,
{
stopOnError: options.stopOnError,
timeout: 30000
}
)
console.log('Pipeline Execution Result:')
console.log(`Success: ${result.success}`)
if (result.error) {
console.log(`Error: ${result.error}`)
}
console.log('\nStage Results:')
// Display stage results
Object.entries(result.stageResults).forEach(([stage, stageResult]) => {
console.log(`\n${stage.toUpperCase()}:`)
console.log(` Success: ${stageResult?.success}`)
if (stageResult?.error) {
console.log(` Error: ${stageResult.error}`)
}
if (stageResult?.data && options.verbose) {
console.log(' Data:')
console.log(JSON.stringify(stageResult.data, null, 2)
.split('\n')
.map(line => ` ${line}`)
.join('\n')
)
}
})
console.log('\nFinal Result Data:')
console.log(JSON.stringify(result.data, null, 2)
.split('\n')
.map(line => ` ${line}`)
.join('\n')
)
} catch (error) {
console.error('Error:', (error as Error).message)
process.exit(1)
}
})
program
.command('stream-test')
.description('Test streaming data through the pipeline (simulated)')
.option('-c, --count <number>', 'Number of data items to stream', '5')
.option('-i, --interval <ms>', 'Interval between data items in milliseconds', '1000')
.option('-t, --data-type <type>', 'Type of data to process', 'text')
.option('-v, --verbose', 'Show detailed output', false)
.action(async (options) => {
try {
// Initialize the pipeline
await sequentialPipeline.initialize()
const count = parseInt(options.count, 10)
const interval = parseInt(options.interval, 10)
console.log(`Simulating stream of ${count} data items at ${interval}ms intervals`)
console.log(`Data type: ${options.dataType}`)
console.log()
// Create a handler function similar to what would be used with WebSockets
const handler = (data: string) => {
// Process the data asynchronously without blocking
sequentialPipeline.processData(
data,
options.dataType,
{ stopOnError: false }
).then(result => {
console.log(`\nProcessed: "${data}"`)
console.log(`Success: ${result.success}`)
if (options.verbose) {
console.log('Stage Results:')
Object.entries(result.stageResults).forEach(([stage, stageResult]) => {
if (stageResult?.success) {
console.log(` ${stage}: Success`)
} else {
console.log(` ${stage}: Failed - ${stageResult?.error || 'Unknown error'}`)
}
})
}
if (result.data) {
console.log('Result Data:')
console.log(JSON.stringify(result.data, null, 2)
.split('\n')
.map(line => ` ${line}`)
.join('\n')
)
}
}).catch(error => {
console.error(`Error processing "${data}":`, error.message)
})
}
// Generate sample data items
const sampleTexts = [
"The quick brown fox jumps over the lazy dog",
"Artificial intelligence is transforming how we interact with data",
"Vector databases enable semantic search capabilities",
"Graph relationships connect entities in meaningful ways",
"Streaming data requires efficient real-time processing",
"WebSockets provide bidirectional communication channels",
"Augmentations extend the functionality of the core system",
"Sequential pipelines process data in defined stages",
"Parallel execution improves throughput for large datasets",
"Threaded operations utilize multiple CPU cores efficiently"
]
// Simulate streaming data
console.log('Starting simulated data stream...')
for (let i = 0; i < count; i++) {
// Use modulo to cycle through sample texts if count > samples
const text = sampleTexts[i % sampleTexts.length]
// Wait for the specified interval
if (i > 0) {
await new Promise(resolve => setTimeout(resolve, interval))
}
console.log(`\nStreaming item ${i + 1}/${count}: "${text}"`)
// Process the data
handler(text)
}
console.log('\nSimulated stream complete. Some processing may still be ongoing.')
console.log('In a real WebSocket scenario, the connection would remain open for continuous data.')
} catch (error) {
console.error('Error:', (error as Error).message)
process.exit(1)
}
})
program
.command('augmentation-info')
.description('Get detailed information about a specific augmentation type')
.argument('<type>', 'Augmentation type (sense, memory, cognition, conduit, activation, perception, dialog, websocket)')
.action(async (typeArg) => {
try {
// Initialize the pipeline
await augmentationPipeline.initialize()
// Resolve the augmentation type
let augType: AugmentationType | undefined
// Convert input to proper enum value
const normalizedType = typeArg.toLowerCase()
switch (normalizedType) {
case 'sense':
augType = AugmentationType.SENSE
break
case 'memory':
augType = AugmentationType.MEMORY
break
case 'cognition':
augType = AugmentationType.COGNITION
break
case 'conduit':
augType = AugmentationType.CONDUIT
break
case 'activation':
augType = AugmentationType.ACTIVATION
break
case 'perception':
augType = AugmentationType.PERCEPTION
break
case 'dialog':
augType = AugmentationType.DIALOG
break
case 'websocket':
augType = AugmentationType.WEBSOCKET
break
default:
console.error(`Unknown augmentation type: ${typeArg}`)
console.log('Available types: sense, memory, cognition, conduit, activation, perception, dialog, websocket')
process.exit(1)
}
// Get augmentations of the specified type
const augmentations = augmentationPipeline.getAugmentationsByType(augType)
console.log(`\n${augType.toUpperCase()} Augmentation Details:`)
if (augmentations.length === 0) {
console.log(' No augmentations registered for this type')
} else {
// Display information about each augmentation
augmentations.forEach((aug, index) => {
console.log(`\n${index + 1}. ${aug.name}`)
console.log(` Description: ${aug.description}`)
console.log(` Status: ${aug.enabled ? 'Enabled' : 'Disabled'}`)
// List available methods
console.log(' Available Methods:')
// Get all methods that aren't from Object.prototype
const methods = Object.getOwnPropertyNames(Object.getPrototypeOf(aug))
.filter(method =>
method !== 'constructor' &&
typeof (aug as any)[method] === 'function' &&
!['initialize', 'shutDown', 'getStatus'].includes(method)
)
if (methods.length === 0) {
console.log(' No custom methods available')
} else {
methods.forEach(method => {
console.log(` - ${method}`)
})
}
})
}
// Show pipeline order information
console.log('\nPipeline Execution Order:')
console.log(' 1. SENSE - Process raw data into structured nouns and verbs')
console.log(' 2. MEMORY - Store and retrieve data')
console.log(' 3. COGNITION - Analyze and reason about data')
console.log(' 4. CONDUIT - Exchange data with external systems')
console.log(' 5. ACTIVATION - Trigger actions based on data')
console.log(' 6. PERCEPTION - Interpret and visualize data')
console.log(' 7. DIALOG - Process natural language interactions')
console.log(' * WEBSOCKET - Enable real-time communication (can be combined with other types)')
} catch (error) {
console.error('Error:', (error as Error).message)
process.exit(1)
}
})
// Add a command for setting up autocomplete
program
.command('completion-setup')

View file

@ -114,7 +114,7 @@ export class HNSWIndex {
for (const neighborId of connections) {
const neighbor = this.nouns.get(neighborId)
if (!neighbor) {
console.error(`Neighbor with ID ${neighborId} not found in addItem traversal`)
// Skip neighbors that don't exist (expected during rapid additions/deletions)
continue
}
const distToNeighbor = this.distanceFunction(vector, neighbor.vector)
@ -149,7 +149,7 @@ export class HNSWIndex {
for (const [neighborId, _] of neighbors) {
const neighbor = this.nouns.get(neighborId)
if (!neighbor) {
console.error(`Neighbor with ID ${neighborId} not found`)
// Skip neighbors that don't exist (expected during rapid additions/deletions)
continue
}
@ -240,7 +240,7 @@ export class HNSWIndex {
for (const neighborId of connections) {
const neighbor = this.nouns.get(neighborId)
if (!neighbor) {
console.error(`Neighbor with ID ${neighborId} not found in search`)
// Skip neighbors that don't exist (expected during rapid additions/deletions)
continue
}
const distToNeighbor = this.distanceFunction(
@ -284,7 +284,7 @@ export class HNSWIndex {
for (const neighborId of connections) {
const neighbor = this.nouns.get(neighborId)
if (!neighbor) {
console.error(`Neighbor with ID ${neighborId} not found in removeItem`)
// Skip neighbors that don't exist (expected during rapid additions/deletions)
continue
}
if (neighbor.connections.has(level)) {
@ -432,7 +432,7 @@ export class HNSWIndex {
const neighbor = this.nouns.get(neighborId)
if (!neighbor) {
console.error(`Neighbor with ID ${neighborId} not found in searchLayer`)
// Skip neighbors that don't exist (expected during rapid additions/deletions)
continue
}
const distToNeighbor = this.distanceFunction(
@ -501,7 +501,7 @@ export class HNSWIndex {
for (const neighborId of connections) {
const neighbor = this.nouns.get(neighborId)
if (!neighbor) {
console.error(`Neighbor with ID ${neighborId} not found in pruneConnections`)
// Skip neighbors that don't exist (expected during rapid additions/deletions)
continue
}

View file

@ -669,7 +669,7 @@ export class FileSystemStorage implements StorageAdapter {
}
/**
* Delete a directory and all its contents
* Delete a directory and all its contents recursively
*/
private async deleteDirectory(dirPath: string): Promise<void> {
try {
@ -677,8 +677,19 @@ export class FileSystemStorage implements StorageAdapter {
for (const file of files) {
const filePath = path.join(dirPath, file)
await fs.promises.unlink(filePath)
const stats = await fs.promises.stat(filePath)
if (stats.isDirectory()) {
// Recursively delete subdirectories
await this.deleteDirectory(filePath)
} else {
// Delete files
await fs.promises.unlink(filePath)
}
}
// After all contents are deleted, remove the directory itself
await fs.promises.rmdir(dirPath)
} catch (error) {
// If the directory doesn't exist, that's fine
if ((error as NodeJS.ErrnoException).code !== 'ENOENT') {