AiSemanticSearch

AI-powered Semantic Search module using:

• Laravel (Package-based architecture)
• Qdrant Vector Database
• Python Flask AI Service
• Sentence Transformers (MiniLM)
• Docker (for Qdrant)

---

Installation Guide

Follow the steps carefully.

clone this repo and move "{base_unopim_directory}/package"

---

Add PSR-4 Autoload

Open composer.json and add:

"autoload": {
    "psr-4": {
        "AiSemanticSearch\\": "packages/AiSemanticSearch/src/",
        "AiSemanticSearch\\Database\\": "packages/AiSemanticSearch/Database/"
    }
}

Then run:

composer dump-autoload

---

Register Service Provider

Open:

config/app.php

Add inside providers array:

AiSemanticSearch\Providers\AiSemanticSearchServiceProvider::class,

---

Setup Qdrant (Vector Database)

Make sure Docker is installed.

Run:

docker run -d \
  --name qdrant \
  -p 6333:6333 \
  -v $(pwd)/qdrant_storage:/qdrant/storage \
  qdrant/qdrant

Verify:

docker ps

Qdrant should be running on:

http://localhost:6333

---

Fix Permissions

Give write permission to setup script:

chmod +x packages/AiSemanticSearch/setup_ai_search.sh

If needed (full permission):

chmod -R 775 packages/AiSemanticSearch

---

Run Setup Script

bash packages/AiSemanticSearch/setup_ai_search.sh

This script will:

• Create Qdrant collection
• Prepare environment
• Validate connection

---

Start Python AI Service

Go to AI service folder:

cd ai_service

Start Flask server:

python app.py

Default runs on:

http://localhost:5000

Keep this running.

---

Dummy Data for Testing

Insert test products into database:

INSERT INTO `products` (`sku`, `status`, `type`, `parent_id`, `attribute_family_id`, `values`, `additional`, `avg_completeness_score`, `created_at`, `updated_at`) VALUES
('veg-001', 1, 'simple', NULL, 1, '{"common": {"sku": "veg-001"}, "channel_specific": {"default": {"cost": {"USD": "3.8"}}}, "channel_locale_specific": {"default": {"en_US": {"name": "Paneer Butter Masala", "price": {"USD": "13.99"}, "meta_title": "Paneer Butter Masala", "description": "<p>Soft cottage cheese cubes in a rich, creamy tomato and butter gravy with kasuri methi.</p>", "meta_keywords": "paneer butter masala, cottage cheese, curry, vegetarian", "meta_description": "Creamy and rich paneer curry", "short_description": "<p>Cottage cheese in creamy tomato gravy.</p>", "url_key": "paneer-butter-masala", "brand": "Punjab Grill", "product_number": "VEG001", "categories": ["Indian Vegetarian", "Curries", "Paneer"]}}}}', NULL, NULL, NOW(), NOW());

(You can insert remaining entries similarly.)

---

System Architecture

Laravel → Calls Flask API
Flask → Generates Embeddings
Embeddings → Stored in Qdrant
Search Query → Vector Search → Ranked Results

---

Services Overview

Service	Port	Description
Laravel	8000	Main App
Qdrant	6333	Vector DB
Flask AI	5000	Embedding Generator

---

Troubleshooting

Qdrant not running?

docker start qdrant

Port already in use?

Change port mapping:

-p 6334:6333

---

Final Checklist

• PSR-4 added
• Service Provider registered
• Composer dumped
• Qdrant running
• Setup script executed
• Flask running
• Dummy data inserted

---

Demo :

If everything is green ✅

Your AI Semantic Search is LIVE 🚀