ORCA: ORchestrating Causal Agent

This repository contains the official implementation of the paper
“ORCA: Orchestrating Causal Agent”, which proposes an interactive multi-agent framework for end-to-end causal analysis on relational databases.

---

Setup

1. Create conda environment

conda create -n ORCA python=3.11 -y
conda activate ORCA

2. Install dependencies and initialize

zsh setup_server.sh

3. Environment variables

Create a .env file in the project root with PostgreSQL, Redis, and API key settings.

cp env.example .env

---

Running ORCA

Run once (non-interactive)

python main.py \
  --query "Estimate the causal effect of coupon usage on total order amount" \
  --db-id reef_db \

Optional flags:

• --interactive : enable human-in-the-loop checkpoints

---

Interactive session

python main.py --db-id reef_db --interactive

License

MIT License