il2cpp-dump-analyzer-mcp/dist/embeddings/embeddings.js

Agentic RAG system for analyzing IL2CPP dump.cs files from Unity games

"use strict";
var __importDefault = (this && this.__importDefault) || function (mod) {
    return (mod && mod.__esModule) ? mod : { "default": mod };
};
Object.defineProperty(exports, "__esModule", { value: true });
exports.IL2CPPEmbeddingManager = void 0;
const documents_1 = require("@langchain/core/documents");
const xenova_embeddings_1 = require("./xenova-embeddings");
const dotenv_1 = __importDefault(require("dotenv"));
// Load environment variables
dotenv_1.default.config();
/**
 * Manages the creation and storage of embeddings for IL2CPP code chunks
 */
class IL2CPPEmbeddingManager {
    constructor() {
        // Initialize the embeddings model
        const modelName = process.env.EMBEDDING_MODEL || 'Xenova/all-MiniLM-L6-v2';
        this.embeddings = new xenova_embeddings_1.XenovaEmbeddings(modelName);
    }
    /**
     * Convert code chunks to LangChain documents with embeddings
     * @param chunks Array of code chunks
     * @returns Array of documents with embeddings
     */
    async createEmbeddings(chunks) {
        // Convert chunks to documents
        const documents = chunks.map(chunk => new documents_1.Document({
            pageContent: chunk.text,
            metadata: chunk.metadata
        }));
        // Create embeddings for all documents
        await this.embeddings.embedDocuments(documents.map(doc => doc.pageContent));
        return documents;
    }
    /**
     * Create an embedding for a query string
     * @param query Query string
     * @returns Embedding vector
     */
    async createQueryEmbedding(query) {
        return this.embeddings.embedQuery(query);
    }
    /**
     * Find similar documents to a query
     * @param query Query string
     * @param documents Array of documents to search
     * @param k Number of results to return
     * @returns Array of documents with similarity scores
     */
    async findSimilarDocuments(query, documents, k = 5) {
        const queryEmbedding = await this.createQueryEmbedding(query);
        const documentEmbeddings = await this.embeddings.embedDocuments(documents.map(doc => doc.pageContent));
        // Calculate similarity scores
        const similarities = [];
        for (let i = 0; i < documents.length; i++) {
            const similarity = this.cosineSimilarity(queryEmbedding, documentEmbeddings[i]);
            similarities.push([documents[i], similarity]);
        }
        // Sort by similarity (descending)
        similarities.sort((a, b) => b[1] - a[1]);
        // Return top k results
        return similarities.slice(0, k);
    }
    /**
     * Calculate cosine similarity between two vectors
     * @param a First vector
     * @param b Second vector
     * @returns Cosine similarity score
     */
    cosineSimilarity(a, b) {
        let dotProduct = 0;
        let normA = 0;
        let normB = 0;
        for (let i = 0; i < a.length; i++) {
            dotProduct += a[i] * b[i];
            normA += a[i] * a[i];
            normB += b[i] * b[i];
        }
        if (normA === 0 || normB === 0) {
            return 0;
        }
        return dotProduct / (Math.sqrt(normA) * Math.sqrt(normB));
    }
    /**
     * Get the dimensionality of the embeddings
     * @returns The number of dimensions in the embedding vectors
     */
    getDimension() {
        return this.embeddings.getDimension();
    }
}
exports.IL2CPPEmbeddingManager = IL2CPPEmbeddingManager;
//# sourceMappingURL=embeddings.js.map