@devilsdev/rag-pipeline-utils 2.1.12 → 2.2.1

package/src/ai/orchestrator.js

@@ -0,0 +1,70 @@
+/**
+ * Model Training Orchestrator
+ * Provides model training coordination and job management
+ * Extracted from ai/index.js per CLAUDE.md decomposition requirements
+ */
+
+"use strict";
+
+// Import the existing model training implementation
+const modelTraining = require("./model-training");
+
+/**
+ * ModelTrainingOrchestrator class
+ * Coordinates and manages model training jobs across tenants
+ */
+class ModelTrainingOrchestrator {
+  constructor() {
+    this.jobs = new Map();
+    this.optimizations = new Map();
+    this.deployments = new Map();
+    this.jobCounter = 0;
+    this.optCounter = 0;
+    this.deployCounter = 0;
+  }
+
+  /**
+   * Create a training job
+   * @param {string} tenantId - Tenant identifier
+   * @param {object} trainingConfig - Training configuration
+   * @returns {Promise<string>} Job ID
+   */
+  async createTrainingJob(tenantId, trainingConfig) {
+    return await modelTraining.createTrainingJob(tenantId, trainingConfig);
+  }
+
+  /**
+   * Get training job status
+   * @param {string} jobId - Job identifier
+   * @returns {Promise<object>} Job status
+   */
+  async getJobStatus(jobId) {
+    return await modelTraining.getJobStatus(jobId);
+  }
+
+  /**
+   * Optimize model configuration
+   * @param {object} modelConfig - Model configuration
+   * @returns {Promise<object>} Optimized configuration
+   */
+  async optimizeModel(modelConfig) {
+    return await modelTraining.optimizeModel(modelConfig);
+  }
+
+  /**
+   * Deploy trained model
+   * @param {string} jobId - Training job identifier
+   * @param {object} deployConfig - Deployment configuration
+   * @returns {Promise<string>} Deployment ID
+   */
+  async deployModel(jobId, deployConfig) {
+    return await modelTraining.deployModel(jobId, deployConfig);
+  }
+}
+
+// Create and export singleton instance
+const orchestrator = new ModelTrainingOrchestrator();
+
+module.exports = orchestrator;
+module.exports.ModelTrainingOrchestrator = ModelTrainingOrchestrator;
+module.exports.default = module.exports;

package/src/ai/retrieval/adaptive-retrieval-engine.js

@@ -0,0 +1,398 @@
+/**
+ * Adaptive Retrieval Engine - Extracted from monolithic AI module
+ * Handles personalized content retrieval and user profile management
+ */
+
+const { EventEmitter } = require("events");
+const crypto = require("crypto");
+
+class AdaptiveRetrievalEngine extends EventEmitter {
+  constructor(options = {}) {
+    super();
+    this.config = {
+      learning: {
+        algorithm: options.algorithm || "contextual_bandit",
+        explorationRate: options.explorationRate || 0.1,
+        learningRate: options.learningRate || 0.01,
+      },
+      ...options,
+    };
+    this.userProfiles = new Map();
+    this.queryHistory = new Map();
+    this.feedbackHistory = [];
+  }
+
+  async initializeUserProfile(userId, preferences = {}) {
+    const profile = {
+      userId,
+      interests: preferences.interests || [],
+      preferences: preferences,
+      expertise: preferences.expertise || "beginner",
+      createdAt: Date.now(),
+      interactions: 0,
+      personalizedRankings: new Map(),
+      learningHistory: [],
+      adaptationWeights: {
+        contentSimilarity: 0.4,
+        userPreference: 0.3,
+        contextualRelevance: 0.2,
+        temporalDecay: 0.1,
+      },
+    };
+
+    this.userProfiles.set(userId, profile);
+    this.emit("userProfileInitialized", { userId, profile });
+    return profile;
+  }
+
+  async getUserProfile(userId) {
+    const profile = this.userProfiles.get(userId);
+    if (!profile) {
+      throw new Error(`User profile not found for ${userId}`);
+    }
+    return profile;
+  }
+
+  async adaptiveRetrieve(userId, query, options = {}) {
+    const profile = this.userProfiles.get(userId);
+    if (!profile) {
+      throw new Error(
+        `User profile not found for ${userId}. Please initialize profile first.`,
+      );
+    }
+
+    const maxResults = options.maxResults || 10;
+
+    // Simulate document retrieval with adaptive ranking
+    const baseDocuments = Array.from({ length: maxResults }, (_, i) => ({
+      id: `doc_${i + 1}`,
+      title: `Document ${i + 1} about ${query}`,
+      content: `Content related to ${query} with relevance score`,
+      baseScore: 0.9 - i * 0.1,
+      metadata: {
+        source: "knowledge_base",
+        timestamp: Date.now() - Math.random() * 86400000,
+        tags: profile.interests.slice(0, 2),
+      },
+    }));
+
+    // Apply adaptive ranking based on user profile
+    const adaptedDocuments = baseDocuments
+      .map((doc) => {
+        const adaptiveScore = this._calculateAdaptiveScore(doc, profile, query);
+        return {
+          ...doc,
+          adaptiveScore,
+          finalScore: doc.baseScore * 0.6 + adaptiveScore * 0.4,
+          adaptationFactors: {
+            userInterests: profile.interests.some((interest) =>
+              doc.title.toLowerCase().includes(interest.toLowerCase()),
+            )
+              ? 0.2
+              : 0,
+            expertise: profile.expertise === "advanced" ? 0.1 : 0,
+            historicalPreference: Math.random() * 0.1,
+          },
+        };
+      })
+      .sort((a, b) => b.finalScore - a.finalScore);
+
+    const result = {
+      documents: adaptedDocuments,
+      adaptationMetadata: {
+        userId,
+        query,
+        adaptationStrategy: this.config.learning.algorithm,
+        profileFactors: {
+          interests: profile.interests,
+          expertise: profile.expertise,
+          interactionCount: profile.interactions,
+        },
+        retrievalTimestamp: Date.now(),
+      },
+    };
+
+    // Update interaction count
+    profile.interactions++;
+
+    this.emit("adaptiveRetrievalCompleted", {
+      userId,
+      query,
+      resultsCount: adaptedDocuments.length,
+      adaptationApplied: true,
+    });
+
+    return result;
+  }
+
+  async processFeedback(userId, feedback) {
+    const profile = this.userProfiles.get(userId);
+    if (!profile) {
+      throw new Error(`User profile not found for ${userId}`);
+    }
+
+    // Store feedback in learning history
+    const feedbackEntry = {
+      id: crypto.randomUUID(),
+      timestamp: Date.now(),
+      query: feedback.query,
+      results: feedback.results,
+      ratings: feedback.ratings || [],
+      clickedResults: feedback.clickedResults || [],
+      dwellTime: feedback.dwellTime || [],
+      feedbackType: "explicit",
+    };
+
+    profile.learningHistory.push(feedbackEntry);
+
+    // Initialize preferences.queryPatterns if not exists
+    if (!profile.preferences.queryPatterns) {
+      profile.preferences.queryPatterns = [];
+    }
+
+    // Update query patterns based on feedback
+    if (feedback.query) {
+      const existingPattern = profile.preferences.queryPatterns.find(
+        (p) => p.query === feedback.query,
+      );
+      if (existingPattern) {
+        existingPattern.frequency++;
+        existingPattern.lastUsed = Date.now();
+        if (feedback.ratings && feedback.ratings.length > 0) {
+          const avgRating =
+            feedback.ratings.reduce((sum, rating) => sum + rating, 0) /
+            feedback.ratings.length;
+          existingPattern.avgRating =
+            (existingPattern.avgRating + avgRating) / 2;
+        }
+      } else {
+        const avgRating =
+          feedback.ratings && feedback.ratings.length > 0
+            ? feedback.ratings.reduce((sum, rating) => sum + rating, 0) /
+              feedback.ratings.length
+            : 3;
+        profile.preferences.queryPatterns.push({
+          query: feedback.query,
+          frequency: 1,
+          avgRating,
+          lastUsed: Date.now(),
+          category: this._categorizeQuery(feedback.query),
+        });
+      }
+    }
+
+    // Update user preferences based on feedback
+    if (feedback.ratings && feedback.ratings.length > 0) {
+      const avgRating =
+        feedback.ratings.reduce((sum, rating) => sum + rating, 0) /
+        feedback.ratings.length;
+
+      // Extract topics from highly rated results
+      if (avgRating >= 4) {
+        feedback.results.forEach((result, index) => {
+          if (feedback.ratings[index] >= 4) {
+            // Extract keywords and add to interests if not already present
+            const keywords = result.title.toLowerCase().split(" ");
+            keywords.forEach((keyword) => {
+              if (keyword.length > 3 && !profile.interests.includes(keyword)) {
+                profile.interests.push(keyword);
+              }
+            });
+          }
+        });
+      }
+    }
+
+    // Update adaptation weights based on feedback patterns
+    this._updateAdaptationWeights(profile, feedbackEntry);
+
+    this.emit("feedbackProcessed", {
+      userId,
+      feedbackId: feedbackEntry.id,
+      learningHistorySize: profile.learningHistory.length,
+      updatedInterests: profile.interests,
+    });
+
+    return {
+      processed: true,
+      feedbackId: feedbackEntry.id,
+      profileUpdated: true,
+      newInterestsCount: profile.interests.length,
+    };
+  }
+
+  async generatePersonalizedRankings(userId, results, context = {}) {
+    const profile = this.userProfiles.get(userId);
+    if (!profile) {
+      throw new Error(`User profile not found for ${userId}`);
+    }
+
+    // Simulate personalized ranking based on user interests
+    const rankedResults = results
+      .map((result) => ({
+        ...result,
+        personalizedScore: Math.random() * 0.5 + 0.5,
+        relevanceFactors: profile.interests.slice(0, 2),
+      }))
+      .sort((a, b) => b.personalizedScore - a.personalizedScore);
+
+    this.emit("personalizedRankingsGenerated", {
+      userId,
+      resultsCount: rankedResults.length,
+    });
+    return rankedResults;
+  }
+
+  async personalizeRanking(userId, documents, context = {}) {
+    const profile = this.userProfiles.get(userId);
+    if (!profile) {
+      throw new Error(`User profile not found for ${userId}`);
+    }
+
+    // Apply personalized ranking to documents
+    const personalizedResults = documents
+      .map((doc) => {
+        let personalizedScore = doc.score || 0.5;
+        const rankingFactors = {
+          contentRelevance: 0,
+          userInterests: 0,
+          historicalPreference: 0,
+          contextualMatch: 0,
+        };
+
+        // Boost score based on user interests
+        if (profile.interests && profile.interests.length > 0) {
+          const interestMatch = profile.interests.some((interest) =>
+            doc.content.toLowerCase().includes(interest.toLowerCase()),
+          );
+          if (interestMatch) {
+            rankingFactors.userInterests = 0.2;
+            personalizedScore += 0.2;
+          }
+        }
+
+        // Apply historical preferences
+        if (profile.learningHistory && profile.learningHistory.length > 0) {
+          rankingFactors.historicalPreference = Math.random() * 0.1;
+          personalizedScore += rankingFactors.historicalPreference;
+        }
+
+        // Contextual matching with query
+        if (context.query) {
+          const queryWords = context.query.toLowerCase().split(" ");
+          const contentWords = doc.content.toLowerCase().split(" ");
+          const overlap = queryWords.filter((word) =>
+            contentWords.includes(word),
+          ).length;
+          rankingFactors.contextualMatch = (overlap / queryWords.length) * 0.15;
+          personalizedScore += rankingFactors.contextualMatch;
+        }
+
+        rankingFactors.contentRelevance = doc.score || 0.5;
+
+        return {
+          ...doc,
+          personalizedScore: Math.min(personalizedScore, 1.0),
+          rankingFactors,
+        };
+      })
+      .sort((a, b) => b.personalizedScore - a.personalizedScore);
+
+    this.emit("personalizedRankingCompleted", {
+      userId,
+      documentsCount: documents.length,
+      query: context.query,
+    });
+
+    return personalizedResults;
+  }
+
+  async optimizeRetrieval(query, context = {}) {
+    const optimizedQuery = `optimized: ${query}`;
+    const retrievalStrategy = "adaptive";
+
+    this.emit("retrievalOptimized", {
+      originalQuery: query,
+      optimizedQuery,
+      strategy: retrievalStrategy,
+      context,
+    });
+
+    return {
+      query: optimizedQuery,
+      strategy: retrievalStrategy,
+      confidence: Math.random() * 0.3 + 0.7,
+    };
+  }
+
+  _categorizeQuery(query) {
+    const lowerQuery = query.toLowerCase();
+    if (lowerQuery.includes("machine learning") || lowerQuery.includes("ml"))
+      return "machine_learning";
+    if (lowerQuery.includes("deep learning") || lowerQuery.includes("neural"))
+      return "deep_learning";
+    if (
+      lowerQuery.includes("ai") ||
+      lowerQuery.includes("artificial intelligence")
+    )
+      return "artificial_intelligence";
+    if (lowerQuery.includes("data") || lowerQuery.includes("analytics"))
+      return "data_science";
+    return "general";
+  }
+
+  _calculateAdaptiveScore(document, profile, query) {
+    let score = 0;
+
+    // Interest matching
+    const interestMatch = profile.interests.some(
+      (interest) =>
+        document.title.toLowerCase().includes(interest.toLowerCase()) ||
+        document.content.toLowerCase().includes(interest.toLowerCase()),
+    );
+    if (interestMatch) score += 0.3;
+
+    // Expertise level matching
+    if (
+      profile.expertise === "advanced" &&
+      document.metadata.tags.includes("advanced")
+    ) {
+      score += 0.2;
+    } else if (
+      profile.expertise === "beginner" &&
+      document.metadata.tags.includes("basic")
+    ) {
+      score += 0.2;
+    }
+
+    // Historical preference (simplified)
+    score += Math.random() * 0.2;
+
+    return Math.min(score, 1.0);
+  }
+
+  _updateAdaptationWeights(profile, feedbackEntry) {
+    // Simple learning rate adjustment based on feedback quality
+    const avgRating =
+      feedbackEntry.ratings.length > 0
+        ? feedbackEntry.ratings.reduce((sum, r) => sum + r, 0) /
+          feedbackEntry.ratings.length
+        : 3;
+
+    if (avgRating >= 4) {
+      // Positive feedback - slightly increase user preference weight
+      profile.adaptationWeights.userPreference = Math.min(
+        0.5,
+        profile.adaptationWeights.userPreference + 0.01,
+      );
+    } else if (avgRating <= 2) {
+      // Negative feedback - increase content similarity weight
+      profile.adaptationWeights.contentSimilarity = Math.min(
+        0.6,
+        profile.adaptationWeights.contentSimilarity + 0.01,
+      );
+    }
+  }
+}
+
+module.exports = { AdaptiveRetrievalEngine };

package/src/ai/retrieval-engine.js

@@ -0,0 +1,69 @@
+/**
+ * Adaptive Retrieval Engine
+ * Provides intelligent document retrieval with adaptive algorithms
+ * Extracted from ai/index.js per CLAUDE.md decomposition requirements
+ */
+
+"use strict";
+
+// Import the existing adaptive retrieval implementation
+const adaptiveRetrieval = require("./adaptive-retrieval");
+
+/**
+ * AdaptiveRetrievalEngine class
+ * Manages intelligent document retrieval with performance optimization
+ */
+class AdaptiveRetrievalEngine {
+  constructor() {
+    this.algorithms = new Map();
+    this.performanceMetrics = new Map();
+    this.adaptationThreshold = 0.1;
+  }
+
+  /**
+   * Process retrieval query with adaptive algorithms
+   * @param {string} tenantId - Tenant identifier
+   * @param {string} query - Search query
+   * @param {object} options - Retrieval options
+   * @returns {Promise<object>} Retrieval results
+   */
+  async retrieveDocuments(tenantId, query, options = {}) {
+    return await adaptiveRetrieval.retrieveDocuments(tenantId, query, options);
+  }
+
+  /**
+   * Optimize retrieval algorithm based on performance data
+   * @param {string} tenantId - Tenant identifier
+   * @param {object} performanceData - Performance metrics
+   * @returns {Promise<object>} Optimization results
+   */
+  async optimizeRetrieval(tenantId, performanceData) {
+    return await adaptiveRetrieval.optimizeRetrieval(tenantId, performanceData);
+  }
+
+  /**
+   * Get current algorithm configuration
+   * @param {string} tenantId - Tenant identifier
+   * @returns {Promise<object>} Current algorithm configuration
+   */
+  async getAlgorithmConfig(tenantId) {
+    return await adaptiveRetrieval.getAlgorithmConfig(tenantId);
+  }
+
+  /**
+   * Update retrieval strategy
+   * @param {string} tenantId - Tenant identifier
+   * @param {string} strategy - New retrieval strategy
+   * @returns {Promise<boolean>} Success status
+   */
+  async updateStrategy(tenantId, strategy) {
+    return await adaptiveRetrieval.updateStrategy(tenantId, strategy);
+  }
+}
+
+// Create and export singleton instance
+const engine = new AdaptiveRetrievalEngine();
+
+module.exports = engine;
+module.exports.AdaptiveRetrievalEngine = AdaptiveRetrievalEngine;
+module.exports.default = module.exports;