⚙️ Workflow Engine — Python + FastAPI

A minimal graph-based workflow engine built for the AI Engineering Assignment. Supports node execution, looping, shared state, and an example Code Review workflow.

Features

#Workflow / Graph Engine

• Nodes as Python functions
• Shared state passed between nodes
• Edges defining execution flow
• Looping via cyclic edges
• Early stopping using state["done"] = True
• Execution logs for debugging

FastAPI Endpoints

• GET / — Health check
• POST /run-test-graph — Verify engine
• POST /graph/create — Create custom workflow
• POST /graph/run — Run stored workflow
• GET /graph/state/{run_id} — Retrieve run output
• POST /workflow/code-review/run — Example workflow with looping

---

Project Structure

app/ ├── main.py # API endpoints ├── graph_engine.py # Core workflow engine ├── tools.py # Optional tool registry └── workflows/ └── codereview.py # Example Code Review workflow

▶️ Run Locally

1. Create a virtual environment

python3 -m venv venv
source venv/bin/activate
(Windows)

Copy code
venv\Scripts\activate
2. Install dependencies
bash
Copy code
pip install -r requirements.txt
3. Start the API server
bash
Copy code
uvicorn app.main:app --reload
4. Open Swagger docs
http://127.0.0.1:8000/docs

Improvements (If more time)
Add conditional branching rules

Add WebSocket streaming logs

Add persistent storage (SQLite/Postgres)

Add async execution for heavy tasks

Add a small UI to design workflows visually

Summary
This project implements a clean, extensible workflow engine suitable for AI/ML pipelines, backend orchestration, and agent-like graph execution.