README

sndsabin/module-semanticsearch

• Main Functionality
• Installation
• Configuration
• Tested On

Main Functionality

This module uses product embeddings to enhance Magento 2 default search with natural language understanding. It depends on Embedding module to generate the product embedding.

Demo (Default Luma Store)

Model used for this demo: qllama/bge-small-en-v1.5 (384-dim embeddings). Results may vary depending on the embedding model used.

Query: "exercise dress"

Default Search	Semantic Search

Query: "pretty dress"

Default Search	Semantic Search

Installation

• Install the module via composer:

composer require sndsabin/module-semanticsearch

• Enable the module:

bin/magento module:enable SNDSABIN_Embedding SNDSABIN_SemanticSearch
bin/magento setup:upgrade

Configuration

⚠️ This module requires Embedding module, configure Embedding module first. Please refer to this for configuration.

⚠️ After configuration, generate product embeddings by following the instructions here.

This process may take some time depending on your catalog size. You can monitor progress by checking the document count in the configured OpenSearch index.

Go to Stores → Configuration → Semantic Search → Settings

General

• Enable
  • Set it to Yes to enable the semantic search functionality.

Tested On:

• Magento 2.4.8
• PHP 8.4
• RabbitMQ 4.1
• OpenSearch 3.3