🌱E-Mealio: A Chat Agent for Sustainable and Healthy Recipes Suggestions

A project originally developed by Antonio Raffaele Iacovazzi as part of his Master’s Thesis in Computer Science, with the goal of developing a chat-based agent that helps users adopt sustainable food habits.

Subsequently, Lorenzo Blanco, as part of his Bachelor's Thesis in Computer Science, further developed the chat-based agent by enhancing its functionalities, improving usability and overall user experience, and expanding its domain beyond sustainability to also address the promotion of healthy food habits.

---

🔌 Current Status

✅ The Telegram bot is currently active and operational. You can interact with it searching @emealio_bot on Telegam and typing the /start command.

---

How to Install

Instructions below are intended for running the project locally.
Python 3 is required.

1. Clone the Repository

Download or clone this repository into a local folder.

---

2. Install the Dataset

a. Install MongoDB

• MongoDB Community Edition
  Installation Guide

• MongoDB Compass (a handy frontend GUI)
  Compass Download

b. Import the Dataset

1. Extract emealio_food_db.zip to a location of your choice.
2. Open MongoDB Compass.
3. Create a new database and name it:
   ➤ emealio_food_db
4. Create a collection inside it called:
   ➤ ingredients
5. Import data:
   • Click on the collection, then use the "Add Data" > "Import JSON" function to import emealio_food_db.ingredients.json.
6. Repeat the process for the remaining .json files:
   • For example, emealio_food_db.recipes.json should be imported into a collection named recipes.
   • Make sure each collection is named exactly like the corresponding file (without the .json extension).

---

3. Compute Embeddings

The dataset does not include embeddings due to their size (~3GB).
Follow these steps to generate them locally.

a. Required Libraries

Make sure the following Python libraries are installed:

pip install pandas numpy pymongo sentence_transformers

b. Generate Embeddings

Run the script:

python datasetUtilities/compute_embeddings.py

• This process takes around 1.5 hours.
• Progress is shown every 100 items with a message like:
  Done N

---

4. Install and Run the Agent

The core agent code is located in the projectRoot folder.

a. Install Dependencies

Use pip to install required libraries:

pip install -r requirements.txt

b. Set Up the Telegram Bot

1. Create a new bot using BotFather on Telegram.
   (Or contact me at ar.iacovazzi@gmail.com to gain access to the existing bot.)

2. Create a .env file in the projectRoot folder with the following contents:

OPENAI_API_KEY=
TELEGRAM_BOT_TOKEN=
ANTHROPIC_API_KEY=

• Add your corresponding API keys.
• You can provide just one (OpenAI or Anthropic), or none if you plan to configure a different LLM via LangChain.

c. Run the Bot

Launch the agent:

python TelegramBot.py

• Send /start to the bot on Telegram.
• If it replies, everything is working! 🎉

---

5. Run Unit Tests

You can verify everything is set up correctly by running:

python test.py

• Tests are validated with:
  • OpenAI GPT-4o
  • Anthropic Claude Sonnet 3.5
• If all tests pass, the bot is ready to go.
• If you use a different LLM, ensure it passes all tests—otherwise, the agent may not work properly.

---

🐳 Run with Docker (Recommended)

Alternatively, you can run E-Mealio fully containerized with Docker and Docker Compose — no manual setup required.

✅ Prerequisites

• Docker
• Docker Compose

📁 Project Structure (Relevant Files)

E-Mealio/
├── docker-compose.yml
├── Dockerfile
├── mongo_dump/                  # Precomputed MongoDB dump (BSON format), WARNING: the actual files are not in the folder, check the readme inside to understand how to download them. 
├── .env                         # Contains API keys
└── projectRoot/                 # Main bot code

---

🔑 1. Set Environment Variables

Create a .env file in the root (same directory as docker-compose.yml) with:

OPENAI_API_KEY=your_openai_key
TELEGRAM_BOT_TOKEN=your_telegram_bot_token
ANTHROPIC_API_KEY=your_anthropic_key

You can leave some variables blank if you're not using a specific provider.

---

📥 2. Download the MongoDB Dump

Download the precomputed MongoDB dump from the following link:

🔗 Download emealio_food_db.zip

Then:

1. Unzip the file.
2. Place the contents inside the /mongo_dump directory.

---

🚀 3. Build and Launch the Containers

From the root of the project, run:

docker-compose up --build

This will:

• Start a MongoDB container
• Automatically restore the precomputed database from /mongo_dump
• Launch the Telegram bot

---

🧪 4. Verify It’s Working

• Open your Telegram bot and send /start
• You should receive a response from the bot within a few seconds 🎉

---

💾 MongoDB Persistence

MongoDB data is stored in a Docker-managed volume named mongo_data.

To completely reset the environment (including clearing the database), run:

docker-compose down -v