blog-agentic-ai

Overview

This repository provides a multi-agent, LangGraph observability application that integrates with Memgraph for data ingestion, semantic search, and signal observability. A Panel-based UI (app/app.py) is included for interacting with the agents.

Prerequisites

1. Python >= 3.10
2. Memgraph

Environment Setup

1. Clone this repository.
2. Install dependencies:

   pip install -r requirements.txt

3. Set environment variables or a .env file with:

   export MEMGRAPH_URI="bolt://localhost:7687"
   export MEMGRAPH_USER="memgraphUser"
   export MEMGRAPH_PASSWORD="MemgraphPassword1233"
   export OPENAI_API_KEY=""
   export MODEL_PROVIDER="local || openai"

4. Run Llama 3.1 8B locally (Optional):
   • Setup Llama 3.1 using ollama

   brew install ollama
   ollama pull llama3.1:8b-instruct-fp16
   ollama pull nomic-embed-text

Data Ingestion

1. Run the graph_ingest script to load them into Memgraph, including vector indexing and embeddings:

   python -m scripts.graph_ingest

2. Verify the data is loaded. You can open Memgraph in the browser or run queries to check the nodes and edges.

Starting the App

1. Launch the UI by running:

   python -m copilot.app

   This starts a Panel server in app.py.
2. Interact with the agents:
   • Provide your user prompt
   • The system automatically references org_id, role_id, user_id, conversation_id from the application or session
   • The multi-agent system chooses which agent/tool to call
   • The application displays the final response

Architecture

• LangGraph orchestrates multi-agent workflows with ReAct-style agents.
• Each agent forcibly reads guard rails (org_id, etc.) from a config["configurable"].
• Memgraph provides adjacency-based queries and local vector indexes for semantic search.
• The UI is powered by Panel (pn.template.*) and Holoviews.

---