Qdrant Cookbook

This repository contains a set of scripts and utilities for working with Qdrant, OpenAI, and embeddings. The scripts allow you to create embeddings for product data, upload the embeddings to a Qdrant collection, and perform search queries on the embeddings.

Installation

To use the scripts in this repository, follow these steps:

1. Clone the repository to your local machine:

   git clone https://github.com/PappaPaj/qdrant-cookbook.git
   

2. Install the required Python packages by running the following command in the project directory:

   pip install -r requirements.txt
   

3. Set up your environment variables by creating a .env file in the project directory and adding the following lines:

   OPENAI_API_KEY=<your-openai-api-key>
   QDRANT_URL=<your-qdrant-url>
   QDRANT_API_KEY=<your-qdrant-api-key>
   

   Replace <your-openai-api-key>, <your-qdrant-url>, and <your-qdrant-api-key> with your respective API keys and URLs.

Usage

make_embeddings.py

This script creates embeddings for product names and descriptions using the OpenAI API.

Usage:

python make_embeddings.py

This script reads product data from the data/products.csv file, creates embeddings for each product name and description using the OpenAI API, and saves the embeddings into a new CSV file (data/products_with_embeddings.csv). The embeddings are added as new columns (name_vector and description_vector) in the CSV file.

upload_embeddings.py

This script uploads the embeddings from the data/products_with_embeddings.csv file to a Qdrant collection.

Usage:

python upload_embeddings.py

This script reads the embeddings data from the data/products_with_embeddings.csv file, initializes the Qdrant client, creates a collection in Qdrant with the specified vector configuration, and inserts the embeddings data into the collection.

search_embeddings.py

This script allows you to perform search queries on the Qdrant collection using user input.

Usage:

python search_embeddings.py

This script prompts the user to enter a query string and performs a search query on the Qdrant collection. It retrieves the top results based on the query and prints them to the console.

metadata_filtered_search.py

This script demonstrates how to perform similarity search with metadata filtering on the Qdrant collection.

Usage:

python metadata_filtered_search.py

This script performs similarity search on the Qdrant collection for a set of queries. It applies a filter condition to further filter the results based on specific brands. The script prints the matched persona, metadata, and score for each query.

Data

The data folder contains the following files:

• products.csv: CSV file containing the product data.
• products_with_embeddings.csv: CSV file containing the product data with embeddings.

Feel free to replace the products.csv file with your own data file in the same format.

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

License

This project is licensed under the MIT License.