Photo by Brooke Lark on Unsplash

AI Nutrition Assistant: a RAG Application

In today’s fast-paced world, maintaining healthy eating habits can feel like an uphill battle. For individuals navigating dietary restrictions or striving to achieve specific health goals, the task becomes even more daunting. The sheer volume of meal options available, coupled with the complexity of nutritional information, can lead to confusion and frustration. Many people often find themselves overwhelmed, unsure of how to make informed dietary choices that align with their needs and preferences. Traditional meal planning methods typically lack the personalization and accessibility required to empower users in their nutritional journeys.

This is where the AI Nutrition Assistant steps in — offering a transformative approach to meal planning and nutrition guidance. Our assistant leverages advanced conversational AI technology to simplify the process, providing users with tailored meal suggestions, easy-to-follow recipes, and comprehensive cooking instructions — all based on their unique dietary preferences.

Introducing the AI Nutrition Assistant

The AI Nutrition Assistant is not just another recipe app; it is a Retrieval-Augmented Generation (RAG) application designed with the primary goal of helping users develop healthier eating habits while simultaneously making nutrition management more engaging and accessible. By integrating cutting-edge technology with a user-friendly interface, the assistant bridges the gap between healthy eating and convenience.

Key Features of the AI Nutrition Assistant

Our assistant boasts an array of powerful features designed to cater to the diverse needs of users:

• Personalized Meal Recommendations: Suggest meals based on dietary preferences (e.g., vegan, gluten-free).
• Nutritional Guidance: Provide nutritional information for selected meals (calories).
• Recipe Retrieval: Filter recipes by meal type (breakfast, lunch, dinner) and cuisine.
• Interactive Q&A: Answer user questions about nutrition and healthy eating in real time.
• Cooking Instructions: Offer step-by-step cooking instructions for selected recipes

Technologies Used

• Python 3.11.9
• Docker and Docker Compose for containerization
• Minsearch for full-text search
• Flask as the API interface
• Grafana for monitoring and PostgreSQL as the backend
• OpenAI for language model integration

Running the Application

Database Configuration

Before the application starts for the first time, the database needs to be initialized.

First, run postgres:

docker-compose up postgres

Then execute the database initialization script:

pipenv shell

cd nutrition_assistant

export POSTGRES_HOST=localhost
python db_start.py

Running with Docker Compose

For the easiest setup, run the app with Docker Compose:

docker-compose up

Running with Docker (without Compose)

If you prefer to run the application in Docker without Docker Compose (e.g., for debugging purposes), you can do the following:

Build the image:

docker build -t nutrition_assistant .

Run it using docker

docker run -it --rm \
    --network="nutrition-assistant_default" \
    --env-file=".env" \
    -e OPENAI_API_KEY=${OPENAI_API_KEY} \
    -e DATA_PATH="data/data.csv" \
    -p 5000:5000 \
    nutrition_assistant

Test the API:

URL=http://localhost:5000
QUESTION="Is the Vegetarian Chili recipe suitable for a vegan diet?"
DATA='{
    "question": "'${QUESTION}'"
}'

curl -X POST \
    -H "Content-Type: application/json" \
    -d "${DATA}" \
    ${URL}/question

Performance Metrics

1. Retrieval Evaluation:

• Hit Rate: 97%
• MRR (Mean Reciprocal Rank): 91%

Improved Version (with tuned boosting):

• Hit Rate: 97%
• MRR: 95%

RAG Flow Evaluation:

• Using the LLM-as-a-Judge metric, we evaluated the quality of our RAG flow:
• 95% RELEVANT
• 5% PARTLY_RELEVANT

Future Directions

We are continuously working on enhancing the monitoring aspects of the application. This will allow for better insights and adjustments based on user interaction and feedback.

Conclusion

The AI Nutrition Assistant is set to revolutionize meal planning by providing personalized nutrition guidance, making healthy eating habits more accessible and engaging for everyone. We invite you to explore the application and take the first step toward a healthier lifestyle!

link to the github code:

https://github.com/buriihenry/AI-Nutrition-Assistance