/**
 * MongoDB Connection Helper
 * Generated by vai v{{vaiVersion}} on {{generatedAt}}
 * 
 * Database: {{db}}
 * Collection: {{collection}}
 */

const { MongoClient } = require('mongodb');

const MONGODB_URI = process.env.MONGODB_URI;

if (!MONGODB_URI) {
  throw new Error('MONGODB_URI environment variable is required');
}

let client = null;
let db = null;

/**
 * Connect to MongoDB and return the database instance.
 * Reuses existing connection if available.
 * @returns {Promise<import('mongodb').Db>}
 */
async function connect() {
  if (db) return db;

  client = new MongoClient(MONGODB_URI);
  await client.connect();
  db = client.db('{{db}}');
  
  console.log('Connected to MongoDB: {{db}}');
  return db;
}

/**
 * Get the documents collection.
 * @returns {Promise<import('mongodb').Collection>}
 */
async function getCollection() {
  const database = await connect();
  return database.collection('{{collection}}');
}

/**
 * Close the MongoDB connection.
 */
async function close() {
  if (client) {
    await client.close();
    client = null;
    db = null;
  }
}

/**
 * Perform a vector search on the collection.
 * @param {number[]} embedding - Query embedding vector
 * @param {object} options - Search options
 * @param {number} options.limit - Number of results (default: 10)
 * @param {number} options.numCandidates - Candidates to consider (default: limit * 10)
 * @param {object} options.filter - Optional pre-filter
 * @returns {Promise<Array<{document: object, score: number}>>}
 */
async function vectorSearch(embedding, options = {}) {
  const collection = await getCollection();
  const limit = options.limit || 10;
  const numCandidates = options.numCandidates || limit * 10;

  const pipeline = [
    {
      $vectorSearch: {
        index: '{{index}}',
        path: '{{field}}',
        queryVector: embedding,
        numCandidates,
        limit,
        {{#if filter}}
        filter: options.filter,
        {{/if}}
      },
    },
    {
      $project: {
        _id: 1,
        text: 1,
        metadata: 1,
        score: { $meta: 'vectorSearchScore' },
      },
    },
  ];

  const results = await collection.aggregate(pipeline).toArray();
  return results.map(doc => ({
    document: doc,
    score: doc.score,
  }));
}

/**
 * Insert documents with embeddings.
 * @param {Array<{text: string, embedding: number[], metadata?: object}>} docs
 * @returns {Promise<{insertedCount: number}>}
 */
async function insertDocuments(docs) {
  const collection = await getCollection();
  
  const documents = docs.map(doc => ({
    text: doc.text,
    {{field}}: doc.embedding,
    metadata: doc.metadata || {},
    _embeddedAt: new Date(),
    _model: '{{model}}',
  }));

  const result = await collection.insertMany(documents);
  return { insertedCount: result.insertedCount };
}

module.exports = {
  connect,
  getCollection,
  close,
  vectorSearch,
  insertDocuments,
};