Telegram-Bot

LipoOut Official Telegram bot for Miniapp, uploading food photo and recording users' information

An intelligent Telegram bot that helps users track their nutrition and fitness goals using AI-powered image analysis and natural language processing. The bot can analyze food photos, provide nutritional information, and offer personalized health advice.

Bot url: t.me/Fit_ioBot

In-bot Usage

• Use /start command to start the bot.
• The bot will ask you to select your health goal.
• After selecting your health goal, the bot will ask you to upload a photo.
• The bot will analyze the photo and give you a report and a chart.
• (Beta) You can choose to store your photo and data in your record.
• (Beta) Use /pay command to pay for the service.
• Use /cancel command to cancel the current operation.
• You can also chat with the bot by sending messages and the bot will remember your conversation.

How to run

Installation

• Install python 3.11 or later.
• Install poetry.
• Use command make install to install all the dependancies in poetry.
• Use command make run to run the telegram bot.

Key Features

• User goal setting and profile management
• Real-time food photo analysis
• Nutritional content estimation
• Visual nutrition charts
• Meal history tracking
• Interactive conversations about health and nutrition
• Payment integration

Technical Architecture &Implementation

1. Main Bot Application (main.py)
   • Handles bot initialization and conversation flow
   • Manages user states and interactions
   • Implements command handlers and callbacks
2. Utilities Package (utils/)
   • gpt4.py: OpenAI GPT-4 integration for image and text analysis
   • chart.py: Nutrition visualization generation
   • save_food_to_db.py: Database operations for meal tracking
   • health_rating.py: Health rating calculations and formatting

State Flow

graph TD
    A[Start] --> B[Health State]
    B --> C[Photo State]
    C --> D[Reply State]
    D --> E[End State]

Key Classes and Functions

ConversationHandler States

class State(Enum):
    HEALTH_STATE = 1
    PHOTO = 2
    REPLY_PHOTO = 3
    REPLY_TEXT = 4
    END = 5

Main Handlers

1. Start Handler
   • Initializes user profile
   • Stores user information in database
   • Sets up conversation context
2. Health State Handler
   • Manages user health goal selection
   • Provides interactive goal buttons
   • Updates user profile with selected goal
3. Photo Handler
   • Processes uploaded food images
   • Triggers AI analysis
   • Generates nutrition charts
   • Offers meal saving options

API Integration

OpenAI GPT-4

• Used for image analysis and natural language understanding
• Custom prompts for consistent formatting
• Nutrition extraction and health scoring

Backend API

• RESTful endpoints for user management
• Food entry storage and retrieval
• User profile updates

Data Models

User Profile

user_data = {
    "name": str,
    "age": int,
    "height": float,
    "weight": float,
    "telegram_id": int,
    "goal": str
}

Food Entry

food_data = {
    "user_id": int,
    "food_analysis": str,
    "food_photo": bytes,
    "calories": float,
    "carb": float,
    "protein": float,
    "fat": float
}

Setup and Configuration

Environment Variables

BOT_TOKEN=your_telegram_bot_token
OPENAI_API_KEY=your_openai_api_key
API_BASE_URL=your_backend_api_url

Dependencies

python-telegram-bot
openai
Pillow
matplotlib
requests
python-dotenv

Future Enhancements

1. Multi-language support
2. Progress tracking and analytics
3. Social sharing features (ex. Instagram)
4. Integration with fitness tracking devices