NoSQL - Document Module

A comprehensive MongoDB-compatible document database built for Cloudflare's edge infrastructure, featuring advanced vector integration, intelligent indexing, and real-time capabilities.

Features

🚀 MongoDB Compatibility

• Full MongoDB query language support
• CRUD operations (Create, Read, Update, Delete)
• Advanced aggregation pipelines
• Complex query operators ($and, $or, $in, $regex, etc.)
• GridFS-like large document support with R2 integration

🔍 Hybrid Search

• Text search with full-text indexing
• Vector similarity search
• Semantic search using embeddings
• Multi-modal search (text, image, audio)
• Personalized recommendations
• Similar document discovery

🔗 Relationships

• Define relationships between collections
• Automatic population of related documents
• One-to-one, one-to-many, many-to-many relationships
• Cascade operations with referential integrity
• Deep population with configurable depth

⚡ Performance

• Intelligent auto-indexing based on query patterns
• Dynamic field indexing for optimal performance
• Query plan optimization and explanation
• Caching with TTL support
• Connection pooling and batch operations

📊 Bulk Operations

• High-performance bulk writes
• Streaming inserts for large datasets
• Real-time document streams
• Parallel processing with configurable concurrency
• Error handling and recovery

🏷️ Smart Tagging

• Automatic content-based tagging
• Hierarchical tag systems
• Tag recommendations and suggestions
• Bulk tagging operations
• Tag analytics and cleanup

📈 Analytics & Monitoring

• Real-time performance metrics
• Query pattern analysis
• Index usage statistics
• Document lifecycle tracking
• Health monitoring and diagnostics

Quick Start

Installation

import { EdgeDocumentDB } from './src/document';

// Create database instance
const db = await EdgeDocumentDB.create({
  name: 'my_app_db',
  d1Database: env.DB,      // Cloudflare D1 binding
  kvStore: env.KV,         // Cloudflare KV binding
  r2Bucket: env.BUCKET,    // Cloudflare R2 binding
  options: {
    enableAutoIndexing: true,
    enableRelationships: true,
    vectorConfig: {
      enabled: true,
      autoEmbedding: true
    }
  }
});

Basic Operations

// Insert documents
const user = await db.insertOne('users', {
  name: 'John Doe',
  email: '[email protected]',
  tags: ['developer', 'javascript']
});

// Query documents
const users = await db.find('users', {
  tags: { $in: ['developer'] },
  createdAt: { $gte: new Date('2024-01-01') }
}, {
  sort: { name: 1 },
  limit: 10
});

// Update documents
await db.updateOne('users', 
  { email: '[email protected]' },
  { $set: { lastLogin: new Date() } }
);

// Aggregation pipeline
const stats = await db.aggregate('users', [
  { $match: { active: true } },
  { $group: { _id: '$department', count: { $sum: 1 } } },
  { $sort: { count: -1 } }
]);

Hybrid Search

// Text search
const articles = await db.textSearch('content', 'machine learning', {
  filters: { category: 'tech' },
  limit: 10
});

// Vector search
const similar = await db.vectorSearch('content', embeddings, {
  threshold: 0.7,
  limit: 5
});

// Semantic search
const semantic = await db.semanticSearch('content', 
  'artificial intelligence neural networks', {
    textWeight: 0.3,
    vectorWeight: 0.7
  }
);

// Hybrid search combining multiple signals
const hybrid = await db.hybridSearch('content', {
  text: 'react development',
  vector: queryEmbedding,
  filter: { publishedAt: { $gte: recentDate } },
  weights: { text: 0.4, vector: 0.6, metadata: 0.0 }
});

Relationships

// Define relationships
await db.defineRelationship('posts', 'users', {
  type: 'manyToOne',
  localField: 'authorId',
  foreignField: '_id',
  foreignCollection: 'users'
});

// Query with population
const posts = await db.findWithPopulate('posts', {}, [
  {
    path: 'authorId',
    select: 'name email avatar'
  },
  {
    path: 'comments',
    match: { approved: true },
    options: { sort: { createdAt: -1 }, limit: 5 }
  }
]);

Bulk Operations

// Bulk write operations
const bulkOps = [
  { insertOne: { document: { name: 'User 1' } } },
  { updateOne: { filter: { name: 'User 2' }, update: { $set: { active: true } } } },
  { deleteOne: { filter: { name: 'User 3' } } }
];

const result = await db.bulkWrite('users', bulkOps);

// Streaming inserts
async function* generateData() {
  for (let i = 0; i < 100000; i++) {
    yield { id: i, value: Math.random() };
  }
}

const streamResult = await db.streamInsert('analytics', generateData(), {
  batchSize: 1000,
  onProgress: (inserted, total) => console.log(`${inserted}/${total}`)
});

Indexing

// Create indexes
await db.createIndex('products', {
  key: { category: 1, price: -1 },
  options: { name: 'category_price_idx' }
});

// Text index
await db.createIndex('articles', {
  key: { title: 'text', content: 'text' },
  options: { weights: { title: 10, content: 5 } }
});

// Vector index
await db.createIndex('embeddings', {
  key: { vector: 'vector' },
  options: {
    vectorOptions: {
      dimensions: 1536,
      similarity: 'cosine',
      type: 'hnsw'
    }
  }
});

// Auto-indexing
await db.autoCreateIndexes('products');

// Get recommendations
const recommendations = await db.getIndexRecommendations('products');

Tagging System

// Auto-tag documents
const tags = await db.autoTag('articles', document, {
  tagSources: ['content', 'metadata'],
  customTagger: (doc) => {
    const tags = [];
    if (doc.content?.includes('React')) tags.push('react');
    return tags;
  }
});

// Apply tags
await db.tagDocument('articles', documentId, tags);

// Find by tags
const taggedDocs = await db.findByTags('articles', ['react', 'typescript'], {
  operator: 'and',
  includeHierarchy: true
});

// Tag statistics
const tagStats = await db.getTagStats('articles', {
  sortBy: 'count',
  limit: 20
});

Advanced Features

Vector Integration

The document module seamlessly integrates with NoSQL's vector capabilities:

// Documents with embedded vectors
const document = {
  title: 'AI Research Paper',
  content: 'Latest advances in machine learning...',
  _vector: {
    id: 'doc1',
    data: new Float32Array([0.1, 0.2, 0.3, ...]), // 1536 dimensions
    metadata: { model: 'text-embedding-ada-002' }
  }
};

// Automatic embedding generation
const db = await EdgeDocumentDB.create({
  name: 'ai_db',
  d1Database: env.DB,
  options: {
    vectorConfig: {
      enabled: true,
      autoEmbedding: true,
      embeddingFields: ['content', 'title'],
      defaultModel: 'text-embedding-ada-002'
    }
  }
});

Real-time Streams

// Create real-time document stream
const stream = db.createDocumentStream('events', {
  batchSize: 100,
  flushInterval: 5000,
  compression: true,
  transform: (doc) => ({
    ...doc,
    processedAt: new Date()
  }),
  errorHandler: (error, batch) => {
    console.error('Stream error:', error);
    // Implement retry logic or dead letter queue
  }
});

// Write to stream
await stream.write({
  event: 'user_action',
  userId: 'user123',
  action: 'click',
  timestamp: new Date()
});

// Stop stream
await stream.stop();

Performance Monitoring

// Database statistics
const stats = await db.stats();
console.log('Total documents:', stats.totalDocuments);
console.log('Index count:', stats.indexCount);

// Query performance
const queryMetrics = db.queryEngine.getQueryMetrics();
const slowQueries = queryMetrics.filter(m => m.latency > 100);

// Index usage
const indexStats = await db.indexManager.getIndexStats('mydb', 'collection');
console.log('Unused indexes:', indexStats.recommendations);

Configuration Options

interface DocumentDatabaseConfig {
  name: string;
  d1Database: any;          // Cloudflare D1 instance
  kvStore?: any;            // Cloudflare KV store
  r2Bucket?: any;           // Cloudflare R2 bucket
  
  // Performance settings
  maxDocumentSize?: number;       // Default: 16MB
  queryTimeout?: number;          // Default: 30s
  batchSize?: number;             // Default: 100
  
  // Caching
  enableQueryCache?: boolean;     // Default: true
  queryCacheTTL?: number;        // Default: 300s
  cacheSize?: number;            // Default: 100MB
  
  // Indexing
  enableAutoIndexing?: boolean;   // Default: true
  autoIndexThreshold?: number;    // Default: 1000
  maxIndexedFields?: number;      // Default: 20
  
  // Vector integration
  vectorConfig?: {
    enabled?: boolean;             // Default: true
    defaultDimensions?: number;    // Default: 1536
    defaultModel?: string;         // Default: 'text-embedding-ada-002'
    autoEmbedding?: boolean;       // Default: false
    embeddingFields?: string[];    // Default: ['content', 'text']
  };
  
  // Features
  enableValidation?: boolean;     // Default: true
  enableSchemaEvolution?: boolean; // Default: true
  enableChangeStreams?: boolean;  // Default: true
  enableRelationships?: boolean;  // Default: true
  enableQueryLogging?: boolean;   // Default: false
  enablePerformanceMetrics?: boolean; // Default: true
  
  // Bulk operations
  bulkWriteBatchSize?: number;    // Default: 1000
  bulkWriteParallelism?: number;  // Default: 4
}

Architecture

The document module is built with a modular architecture:

src/document/
├── edge-document-db.ts          # Main database class
├── types.ts                     # Type definitions
├── storage/
│   └── document-storage.ts      # Core storage engine
├── operations/
│   ├── query-engine.ts          # MongoDB query processing
│   ├── hybrid-search.ts         # Hybrid search engine
│   └── bulk-operations.ts       # Bulk and streaming operations
├── indexes/
│   └── index-manager.ts         # Intelligent indexing
├── relationships/
│   └── relationship-manager.ts  # Document relationships
├── metadata/
│   └── tagging-system.ts        # Smart tagging system
└── examples/
    └── basic-usage.ts           # Comprehensive examples

Performance Characteristics

Latency Targets

• Simple queries: < 10ms p99
• Complex aggregations: < 100ms p99
• Bulk operations: 10,000+ docs/second
• Vector similarity: < 50ms p99

Scalability

• Documents: Unlimited (distributed across D1 + R2)
• Collections: Unlimited
• Indexes: 20 dynamic indexes per collection
• Concurrent operations: 1000+ per database

Storage Efficiency

• Automatic compression for large documents
• Intelligent caching with LRU eviction
• Vector quantization for storage optimization
• Tiered storage (D1 for metadata, R2 for large docs)

Best Practices

Query Optimization

// Use indexes effectively
await db.find('products', {
  category: 'electronics',  // Indexed field
  price: { $gte: 100 }     // Indexed field
});

// Limit results
await db.find('products', filter, { limit: 20 });

// Use projection to reduce data transfer
await db.find('products', filter, {
  projection: { name: 1, price: 1 }
});

// Explain queries for optimization
const explanation = await db.explain('products', filter);

Memory Management

// Use streaming for large datasets
const stream = db.createDocumentStream('logs', {
  batchSize: 1000,
  flushInterval: 5000
});

// Clear caches periodically
db.hybridSearchEngine.clearSearchCache();
db.relationshipManager.clearPopulateCache();

Error Handling

try {
  await db.insertOne('users', document);
} catch (error) {
  if (error instanceof DuplicateKeyError) {
    // Handle duplicate key
  } else if (error instanceof ValidationError) {
    // Handle validation error
  } else {
    // Handle other errors
  }
}

Integration with NoSQL

The document module is designed to work seamlessly with other NoSQL modules:

// Unified database instance
import { NoSQLDB } from '../index';

const vectorDB = new NoSQLDB({
  name: 'unified_db',
  d1Database: env.DB,
  kvStore: env.KV,
  r2Bucket: env.BUCKET
});

// Access document operations
const documentDB = vectorDB.documents();
await documentDB.insertOne('content', document);

// Access vector operations
const vectorStore = vectorDB.vectors();
await vectorStore.addVectors(vectors);

// Hybrid operations
const results = await documentDB.hybridSearch('content', {
  text: 'search query',
  vector: queryVector
});

Migration from MongoDB

The document module provides a migration-friendly API:

// MongoDB equivalent operations
const collection = db.collection('users');

// NoSQL
const users = await db.find('users', filter, options);
const user = await db.findOne('users', filter);
const result = await db.insertOne('users', document);
const updateResult = await db.updateMany('users', filter, update);
const deleteResult = await db.deleteOne('users', filter);

// Aggregation pipelines work identically
const pipeline = [
  { $match: { active: true } },
  { $group: { _id: '$department', count: { $sum: 1 } } }
];
const results = await db.aggregate('users', pipeline);

Monitoring and Observability

// Performance metrics
const metrics = db.queryEngine.getQueryMetrics();
const slowQueries = metrics.filter(m => m.latency > 100);

// Index recommendations
const recommendations = await db.getIndexRecommendations('collection');

// Database health
const isHealthy = await db.ping();

// Resource usage
const cacheStats = db.hybridSearchEngine.getSearchCacheStats();
const populateStats = db.relationshipManager.getPopulateCacheStats();

License

Part of NoSQL - distributed under the same license as the main project.