feature/recommendation-model

[andres15alvarez]

This pull request introduces a recommendation system for products using machine learning and integrates it into the existing application. It also includes updates to dependencies, enhancements to DTOs, and improvements to the ProductsService. Below is a summary of the key changes:

Recommendation System Integration:

• Added a new RecommendationModule with a RecommendationService that uses a nearest neighbors algorithm to recommend products based on user purchase history. The service fetches user-specific product data and calculates recommendations using TensorFlow (src/recommendation/recommendation.module.ts, src/recommendation/recommendation.service.ts, src/recommendation/model.ts, src/recommendation/data.ts) [1] [2] [3] [4].
• Integrated the recommendation system into the ProductsController with a new GET /product/recommendations endpoint, guarded by authentication (src/products/products.controller.ts).

Dependency Updates:

• Added @tensorflow/tfjs-node for machine learning functionality and updated typeorm to version ^0.3.22 (package.json) [1] [2].

Enhancements to DTOs:

• Added the @Expose decorator to multiple DTOs to ensure proper serialization and used class-transformer's @Type decorator for nested objects (src/category/dto/category.dto.ts, src/products/dto/manufacturer.dto.ts, src/products/dto/product.dto.ts) [1] [2] [3].

Improvements to Products Service:

• Refactored the ProductsService by deprecating the old getProducts method and introducing a new, more flexible version that supports additional filters like categoryIds, manufacturerIds, and priceRange (src/products/products.service.ts) [1] [2].
• Removed unused repositories from the service for cleaner code (src/products/products.service.ts).

Database Enhancements:

• Added a VirtualColumn to the ProductPresentation entity to calculate stock quantity dynamically using a SQL query (src/products/entities/product-presentation.entity.ts).