Installation

Before installing, ensure that you have the following:

• Python 3.7+ installed. You can check your version with:

python --version

• pip (Python package manager) installed. You can check with:

pip --version

The framework relies on several open-source libraries, including:

• SentenceTransformers (for embeddings)
• FAISS (for fast similarity search)
• Pillow (for image processing)

To install all dependencies, run:

pip install sentence-transformers faiss-cpu pillow

If you plan to use GPU acceleration for FAISS, install the GPU-supported version:

pip install faiss-gpu

To manage dependencies efficiently, it's recommended to set up a virtual environment:

# Create a virtual environment (Linux/Mac)
python -m venv venv
source venv/bin/activate

# (Windows)
python -m venv venv
venv\Scripts\activate

Then, install the required dependencies inside the virtual environment:

pip install -r requirements.txt

If you don't have a requirements.txt file, create one and add the following:

sentence-transformers
faiss-cpu
pillow

To ensure everything is installed correctly, run the following Python snippet:

import torch
from sentence_transformers import SentenceTransformer
import faiss
from PIL import Image

# Check SentenceTransformers
model = SentenceTransformer('all-MiniLM-L6-v2')
print("SentenceTransformer loaded successfully.")

# Check FAISS
index = faiss.IndexFlatL2(384)  # 384 is the embedding dimension
print("FAISS index initialized.")

# Check Pillow
image = Image.new('RGB', (100, 100), color='red')
print("Pillow image created.")

If no errors occur, your environment is set up correctly.

If this framework is hosted on GitHub, you can clone it using:

git clone https://github.com/your-repo-name/ai-agent-framework.git
cd ai-agent-framework

After installation, test the framework by creating an agent and processing a query:

from agent import create_agent, run_agent

# Create an agent from the default config
agent = create_agent("agent_config.json", llm_provider="openai")

# Test query
response = run_agent(agent, "Hello, what can you do?")
print("Agent Response:", response)