Axon Pharmacy | Function Calling With LLM

This is an app that allows admin to generate and post announcements to Telegram channels and groups, add medicines to the database, update stock, delete medicines, and update order status. It also allows users to search for medicines, view medicine details, get prices, place orders, view order status, and cancel orders. All done with a prompt to Gemini AI.

Live App: https://axon-automation.streamlit.app

Here is how it works:

Setup

1. Clone the repository:

   git clone https://github.com/alexkalll/axon-pharma.git
   cd axon-pharma

2. Install dependencies:

   pip install -r requirements.txt

3. Environment Variables: Create a .env file in the root directory and add the following (replace with your actual values):

   GEMINI_API_KEY=your_gemini_api_key
   TELEGRAM_BOT_TOKEN=your_telegram_bot_token
   CHANNEL_USERNAME=@your_telegram_channel_username
   GROUP_USERNAME=@your_telegram_group_username
   # Firebase credentials (which handled by firebase/db_manager.py)
   # and make sure to save firebase_credentials.json in the root directory.

Tech Stack

• Python
• Jupyter Notebook
• Streamlit
• Firebase
• Gemini AI
• Telegram API
• Telegram Bot Integration
• API Integration
• Database Management

How to Run

Run the User Application

To run the main user application, use the following command:

streamlit run app.py

Run the Admin Application

To run the admin dashboard, use the following command:

streamlit run admin.py # to run the admin page in Streamlit

Features

Admin Application (admin.py)

• Login: Admin login with email and password.
• Generate Announcement: Create and post pharmacy announcements to Telegram channel and group using Gemini AI.
• Add Medicine: Add new medicine entries to the Firebase database.
• Update Stock: Modify the stock quantity of existing medicines and notify users if a medicine is out of stock via Telegram.
• Delete Medicine: Remove medicine entries from the database and send Telegram notifications.
• Update Order Status: Change the status of customer orders in Firebase.

User Application (app.py)

• Guest Mode (no account required): Explore and try the app without signing up. Guests can check availability and prices.
• Search Medicines Availability: Search for medicines by name or category.
• Get Medicine Price: Ask for the price of a medicine; if available, returns the current unit price and stock status.
• Place Order: Place a new order for a medicine.
• Cancel Order: Cancel a pending or processing order.
• View Order Status: Check the status of your order.
• Get Professional Advice: Get advice from the llm based on his symptoms and profile details and order status.

Quick start on the live app:

• Click "Continue as Guest (No account needed)" on the login page.
• Try: "Do you have paracetamol?", "What’s the price of doxycycline?", "Is insulin in stock?"

Key Insights and Ouptuts

• Streamlit: Streamlit is a Python library for building interactive web applications.
• Firebase: Firebase is a platform for building and deploying apps, databases, and other cloud services.
• Gemini AI: Gemini AI is a platform for building and deploying LLMs.
• Telegram: Telegram is a messaging and voice communication platform.

This is a demo for How LLMs are powerful in automating pharmacy services. Take a look at in the sample output below that the model has done for the Pharmacy Admin...

• Here was the prompt given to the model

write a telegram announcement regarding this that a new medicine 
using telegram_post function called paracetamol is arrived and 
record this medicine, and add 200 stock to the aspirin medicine, 
and delete the asprin medicine since it is prohibited by the 
government, and update the order status of order 
0b527cd538844e7ab93c4f656f314cd7 to delivered.

It has been be able to call multiple functions at the same time.

CI pipline & Dockerfile:

• GitHub Actions workflow at .github/workflows/ci.yml that sets up Python 3.11 and installs requirements.txt.
• Simple Dockerfile to run the Streamlit app on port 8501.

Usage

• CI: push or open a PR against main; it will install dependencies.
• Docker:
  • Build: docker build -t axon-pharma .
  • Run: docker run -p 8501:8501 axon-pharma

Note: For local Docker runs, you’ll need to pass secrets via environment or mount a .streamlit/secrets.toml/.env if required.

You are welcome to contribute!🤝