Physics SQL Lab

A small scientific computing project for learning SQL, PostgreSQL, Python, testing, GitHub workflow, and CI/CD through computational physics examples.

The first implemented model is a basic 2D Ising model. Simulation results are stored in a PostgreSQL database and can be queried later for analysis.

Project goals

• Learn SQL from scratch
• Connect Python with PostgreSQL
• Store computational physics experiments in a database
• Use Git and GitHub properly
• Add automated tests with pytest
• Run tests automatically with GitHub Actions CI
• Build toward a reproducible research-style workflow

Current features

• Basic 2D Ising lattice
• Periodic boundary conditions
• Magnetization calculation
• Energy calculation
• PostgreSQL connection with SQLAlchemy
• Experiment logging into database
• Basic pytest tests
• GitHub Actions CI pipeline

Project structure

physics-sql-lab/
├── .github/workflows/ci.yml
├── notebooks/
├── src/
│   ├── analysis/
│   ├── db/
│   ├── models/
│   │   └── ising.py
│   └── run_experiment.py
├── tests/
│   └── test_ising.py
├── .gitignore
├── pytest.ini
├── README.md
└── requirements.txt

## Setup

Create and activate conda environment:

```bash
conda create -n physics_sql python=3.12
conda activate physics_sql

Install dependencies:

pip install -r requirements.txt

Create a .env file in the project root:

DATABASE_URL=postgresql+psycopg2://postgres:YOUR_PASSWORD@localhost:5432/physics_lab

The .env file is ignored by Git and should not be committed.

Run Ising model

python src/models/ising.py

Run experiment and save to PostgreSQL

python src/run_experiment.py

View saved experiments

python src/db/select_experiments.py

Run tests

pytest

Database table: experiments

Current columns:

• id
• model_name
• lattice_size
• temperature
• coupling_j
• steps
• created_at
• magnetization
• energy

Roadmap

• Add temperature sweep for the Ising model

• Store many simulation runs automatically

• Add SQL queries for average energy and magnetization

• Add plots from database results

• Add more tests

• Add PostgreSQL service to GitHub Actions

• Add Docker Compose

• Add Streamlit dashboard

• Later: add Bose-Hubbard or Fermi-Hubbard toy models

• Monte Carlo temperature sweep

This project includes a Metropolis Monte Carlo implementation of the 2D Ising model. The simulation can be run for multiple lattice sizes and temperatures. For each run, the measured physical observables are stored in a PostgreSQL database.

Stored observables include:

• mean energy per spin
• energy standard deviation
• mean magnetization per spin
• mean absolute magnetization per spin
• magnetization standard deviation
• heat capacity
• magnetic susceptibility

The main idea is to avoid recomputing expensive simulations. Results are saved permanently in SQL tables and can later be reused for analysis and plotting.

Run temperature sweep

python src/run_temperature_sweep.py

Load saved results

python src/analysis/load_ising_results.py

Plot magnetization

python src/analysis/plot_magnetization.py

Current database table

The main Monte Carlo results are stored in:

ising_mc_results

This table contains simulation parameters, measured observables, and timestamps.

Scientific goal

The project aims to reproduce basic finite-size behavior of the 2D Ising model near the critical temperature:

Tc ≈ 2.269

Later extensions will include energy plots, heat capacity peaks, magnetic susceptibility, larger lattice sizes, cached runs, and more advanced statistical analysis.