🧠 NutriHelp-AI

This repository contains the AI service for the NutriHelp project. It exposes machine learning model predictions through a FastAPI server, allowing other services (e.g., Backend) to send data and receive results.

---

🚀 Getting Started

Follow the steps below to clone and run the AI model server locally.

1. 📥 Clone the Repository

git clone https://github.com/Gopher-Industries/NutriHelp-AI.git
cd NutriHelp-AI

---

2. 🐍 (Optional) Create a Virtual Environment

Windows

python -m venv venv
venv\Scripts\activate

If you see a PowerShell error, try this instead:

venv\Scripts\activate.bat

macOS / Linux

python3 -m venv venv
source venv/bin/activate

---

3. 📦 Install Dependencies

pip install -r requirements.txt

---

4. 🧪 Run the AI Server

python run.py

By default, the FastAPI server will run at:

http://127.0.0.1:8000

You can access the Swagger UI documentation at:

http://127.0.0.1:8000/docs

If you want to run on different port, go to run.py file and change the port to different one.

---

📬 API Endpoint Example

POST /ai-model/obesity/predict

Accepts JSON-formatted input and returns a prediction result.

🔄 Example Request Payload:

{
  "Gender": "Male",
  "family_history_with_overweight": "yes",
  "Age": 25,
  "Height": 1.75,
  "Weight": 85,
}

---

📡 How the Backend Team Can Use the NutriHelp AI API

Once the AI server is running (locally or on a deployed environment), the Backend team can send requests to it over HTTP.

---

🔗 Endpoint Information

Method	Endpoint	Description
POST	/ai-model/obesity/predict	Get obesity prediction
POST	/ai-model/chatbot/query	Get response from AI

---

📤 Example Request

🔧 URL (Local Testing)

http://localhost:8000/ai-model/obesity/predict

🧾 Headers

Content-Type: application/json

📦 JSON Body

{
  "Gender": "Male",
  "Age": 25,
  "Height": 1.75,
  "Weight": 85,
}

✅ Example Successful Response

{
  "obesity_level": "Overweight"
}

---

🧪 How to Test

✅ Example: Send Obesity Prediction Request

const data = {
  "Gender": "Male",
  "Age": 25,
  "Height": 1.75,
  "Weight": 65,
  "family_history_with_overweight": "yes",
  "FAVC": "yes",
  "FCVC": 2.5,
  "NCP": 3,
  "CAEC": "Sometimes",
  "SMOKE": "no",
  "CH2O": 2.5,
  "SCC": "no",
  "FAF": 0.5,
  "TUE": 1,
  "CALC": "Sometimes",
  "MTRANS": "Public_Transportation"
};

fetch("http://localhost:8000/predict", {
  method: "POST",
  headers: {
    "Content-Type": "application/json"
  },
  body: JSON.stringify(data)
})
.then(response => response.json())
.then(result => console.log("Prediction:", result))
.catch(error => console.error("Error:", error));

---

NutriBot AI Setup Guide

NutriBot is an LLM-powered AI assistant designed for nutritional advice, document ingestion, emotion-aware dialogue, and smart retrieval of local or online health information. This system is integrated with LangChain, OpenAI models, Redis memory, and optional external tools (SERP API and Nutrition API).

---

🚀 Features

• ✅ Emotion-aware chat using GPT-4o or GPT-3.5-Turbo via OpenAI API
• ✅ Tool calling for nutrition lookup, document search (RAG), and web search
• ✅ Long-term memory via Redis
• ✅ PDF and webpage document ingestion with vector storage using Qdrant
• ✅ WebSocket streaming and REST support

---

🧩 Requirements

• Python 3.10 or 3.11
• Git
• pip
• Redis

---

📦 Installation Steps

🧠 Set Up Redis (Memory Storage) Windows Download Redis .msi from this link: https://github.com/tporadowski/redis/releases

Install it.

Add the Redis install folder to your system PATH. Example:

C:\Program Files\Redis

Run Redis server in terminal:

redis-server

macOS

brew install redis
brew services start redis

Ubuntu/Debian

sudo apt update
sudo apt install redis-server
sudo service redis-server start

🔐 Configure Environment Variables Create a .env file in the root directory: Paste in the following, replacing with your actual keys:

OPENAI_API_KEY=your-openai-key
OPENAI_API_BASE=https://api.openai.com/v1
SERPAPI_API_KEY=your-serpapi-key
NINJA_API_KEY=your-ninja-api-key

❗️ Never commit this file. It is listed in .gitignore. Ask person in charge for the actual file.