Product Recommendation System API

A FastAPI-powered product recommendation system with PostgreSQL backend. The system provides text-based search functionality and generates product recommendations using both content-based (text similarity) and collaborative filtering approaches.

Tech Stack

• Programming Language: Python
• Frameworks: FastAPI/Flask/Django (depending on your implementation)
• Database: PostgreSQL/MySQL
• Libraries:
  • Sentence Transformers: For generating embeddings from textual data.
  • Scikit-learn: For calculating cosine similarity.
  • NLTK: For tokenizing and cleaning text data.
  • Levenshtein: For fuzzy string matching in search.
  • SQLAlchemy: For database ORM.
• Other Tools:
  • Docker: For containerized deployments (if applicable).
  • Cloud Storage: (Optional, for scalable deployments)

---

How It Works

Recommendation Service

The Recommendation Service combines collaborative filtering with semantic similarity using pre-trained text embeddings.

1. Collaborative Filtering:

   • Identifies products purchased by a customer.
   • Recommends similar products based on shared tags or attributes.

2. Semantic Similarity:

   • Preprocesses product descriptions and user inputs.
   • Converts texts into vector embeddings using a pre-trained model (paraphrase-MiniLM-L6-v2).
   • Measures cosine similarity between product vectors and user preferences.
   • Filters and ranks products based on similarity scores.

Example Flow:

• Input: User Query (e.g., "Wireless headphones")
• Output: Top N recommendations based on matching product descriptions, tags, and categories.

---

Search Service

The Search Service allows users to search for products using flexible filters and fuzzy matching.

1. Fuzzy Matching:

   • Compares the user's search term with product names, descriptions, tags, and brands.
   • Computes similarity scores using the Levenshtein distance.

2. Search Filtering:

   • Filters results by product attributes such as category, brand, and price range.

3. Sorting and Ranking:

   • Results are sorted based on fuzzy matching scores and presented to the user.

Example Flow:

• Input: Search Term (e.g., "Bluetooth earbuds"), Filters (e.g., Category: "Electronics", Price: "< 3000")
• Output: Sorted product list matching the query and filters.

---

Let me know if you’d like additional refinements!

Features

• Text-based product search with fuzzy matching
• Content-based recommendations using sentence transformers
• Collaborative filtering based on user purchase history
• Complete CRUD operations for products, customers, and transactions
• Sample data generation utility
• Detailed API documentation with OpenAPI/Swagger UI

Requirements

• Python (3.8 - 3.11)
• Note: Python 3.12 is not yet fully supported by all dependencies
• PostgreSQL 12+
• Poetry (recommended) or pip

Installation

1. Clone the repository:

git clone https://github.com/yourusername/product-recommendation-system.git
cd product-recommendation-system

2. Create and activate a virtual environment:

python -m venv venv
source venv/bin/activate  # Linux/Mac
# or
.\venv\Scripts\activate  # Windows

3. Install dependencies:

pip install -r requirements.txt

4. Set up PostgreSQL:

   • Create a new PostgreSQL database
   • Update the database connection settings in .env file (use .env.example as template)

5. Generate sample data:

python -m app.utils.data_generator

Usage

1. Start the API server:

uvicorn app.main:app --reload

2. Access the API documentation:
   • Swagger UI: http://localhost:8000/docs
   • ReDoc: http://localhost:8000/redoc

API Endpoints

Search and Recommendations

• GET /api/v1/search/: Search products with optional filters
• GET /api/v1/recommendations/similar/: Get similar products based on text similarity
• GET /api/v1/recommendations/collaborative/{customer_id}: Get recommendations based on purchase history

Products

• GET /api/v1/products/: List all products
• GET /api/v1/products/{product_id}: Get specific product

Customers

• GET /api/v1/customers/: List all customers

Transactions

• GET /api/v1/transactions/: List all transactions
• GET /api/v1/transactions/customer/{customer_id}: Get customer's transactions

Project Structure

product_recommendation_system/
├── app/
│   ├── api/
│   │   ├── __init__.py
│   │   └── endpoints.py
│   ├── core/
│   │   ├── __init__.py
│   │   └── config.py
│   ├── db/
│   │   ├── __init__.py
│   │   └── base.py
│   ├── models/
│   │   ├── __init__.py
│   │   ├── customer.py
│   │   ├── product.py
│   │   └── transaction.py
│   ├── schemas/
│   │   ├── __init__.py
│   │   ├── customer.py
│   │   ├── product.py
│   │   └── transaction.py
│   ├── services/
│   │   ├── __init__.py
│   │   ├── recommendation.py
│   │   └── search.py
│   ├── utils/
│   │   ├── __init__.py
│   │   └── data_generator.py
│   ├── __init__.py
│   └── main.py
├── requirements.txt
└── README.md

Environment Variables

Create a .env file in the project root with the following variables:

POSTGRES_SERVER=localhost
POSTGRES_USER=your_user
POSTGRES_PASSWORD=your_password
POSTGRES_DB=product_recommendation
POSTGRES_PORT=5432