AI-Powered Text Summarization Agent 📝🤖

🚀 Overview

This project provides a user-friendly AI-powered text summarization bot that helps users quickly condense long texts while preserving key information. The system is designed as an API-based solution with a Streamlit frontend for interaction.

The backend is powered by BCG AgentKit, a robust framework for building AI agents, and integrates Dynatrace OpenTelemetry to export traces for monitoring AI performance and system health.

It leverages Ollama and OpenAI's language models for accurate and context-aware text summarization. The entire workflow is orchestrated via Langchain, ensuring smooth AI interactions.

---

🛠️ Tech Stack

Component	Technology Used
Frontend	Streamlit (Python-based web framework)
Backend	Python + BCG AgentKit
AI/ML	Ollama + OpenAI API
Monitoring	Dynatrace + OpenTelemetry
Orchestration	Langchain
Containerization	Docker
Cloud Platform	Microsoft Azure (Azure Container Instances)
Version Control	Git

---

📌 Features

✅ AI-based summarization: Uses OpenAI's models for concise, meaningful text summaries.
✅ User-friendly interface: Simple web app built with Streamlit.
✅ Backend powered by BCG AgentKit: Utilizes BCG AgentKit for efficient AI agent interactions.
✅ Scalable & cloud-deployable: Runs on Azure Container Instances for seamless scaling.
✅ End-to-end monitoring: Integrated with Dynatrace via OpenTelemetry for exporting traces and performance insights.
✅ Efficient orchestration: Uses Langchain to manage AI tasks efficiently.
✅ Containerized deployment: Packaged with Docker for portability.

🖥️ Installation & Setup

1️⃣ Clone the Repository

git clone https://github.com/Chaimaaorg/SummarizeBot.git
cd SummarizeBot

2️⃣ Configure Environment Variables

• Rename .env.example to .env:

  mv .env.example .env

• Open .env and fill in your personal API keys (e.g., OpenAI API key, Dynatrace credentials, etc.).

3️⃣ Build and Run the Docker Containers

Use docker-compose to set up and launch the services:

docker-compose up --build

4️⃣ Add Ollama Models

Once the Ollama image is built, you need to add the required models:

docker exec -it ollama_summarizer bash
ollama pull llama3.2:1b  # Example model
exit

(Replace llama3.2:1b with the model of your choice if needed.)

5️⃣ Restart the Containers

After pulling the models, restart the service:

docker-compose up --build

6️⃣ Access the Bot

Once running, navigate to:
➡️ http://localhost:8501/ to interact with the summarization bot! 🎯🚀

---

📊 Monitoring with Dynatrace & OpenTelemetry

This project exports traces to Dynatrace using OpenTelemetry, ensuring real-time observability and debugging capabilities.

How Monitoring Works

✔ Traces AI interactions and API requests.
✔ Monitors response times, latency, and errors.
✔ Provides performance insights into AI-generated summaries.

📊 Dynatrace dashboards allow real-time visualization of AI agent activity and system health.

🛠️ Architecture Overview

+-----------------------+
|      User Input       |
+-----------------------+
         ↓
+-----------------------+
|      Streamlit UI     |
+-----------------------+
         ↓
+-----------------------+
|   FastAPI Backend     |
|  (BCG AgentKit-based) |
+-----------------------+
         ↓
+-----------------------+
|   Langchain Agent     |
+-----------------------+
         ↓
+-----------------------+
|   Ollama + OpenAI API |
+-----------------------+
         ↓
+-----------------------+
|   Dynatrace Tracing   |
|  (via OpenTelemetry)  |
+-----------------------+
         ↓
+-----------------------+
|     Summarized Text   |
+-----------------------+

---

👨‍💻 Author & Contributions

Developed by Chaimaâ Ourgani 🚀
Feel free to contribute via Pull Requests!