# ABJM large N and the Airy approximation

Read the calculation, assumptions, exercises and interpretation at:
https://susy.qft.org/labs/abjm-large-n/

## Start in the extracted folder

Extract the ZIP, then open a terminal in the `abjm-large-n` folder containing this
README and `requirements.txt`. Use Python 3.13; the reference calculations
were run with Python 3.13.7. Installation of the pinned packages requires
network access unless they are already cached. The calculations themselves
use local files and do not need the website or a repository checkout.

### Windows PowerShell

```powershell
python -m venv .venv
.venv\Scripts\python.exe -m pip install -r requirements.txt
.venv\Scripts\python.exe -I large_n.py --json result.json --svg regenerated.svg --check-expected expected-output.json
```

### macOS or Linux

```sh
python3 -m venv .venv
.venv/bin/python -m pip install -r requirements.txt
.venv/bin/python -I large_n.py --json result.json --svg regenerated.svg --check-expected expected-output.json
```

The explicit interpreter paths avoid shell activation settings. If a package
is already installed elsewhere, the new environment still needs its own
installation. Keep the pinned versions when reproducing the reference run.
Use `--help` after the script name to inspect supported options.

## Read the result

The program performs the reference comparison itself and prints its result.

The command checks the default physical columns against the declared absolute tolerance of 5e-11, writes result.json, and regenerates the plot without replacing the supplied SVG. Float64 diagnostics can vary by platform. The resolution indicator is empirical, and the Airy expression omits nonperturbative corrections.

Default parameters and expected-output files belong together. A changed
input need not match the default reference. A passing computational check
supports its stated scope, not an unrestricted physical or numerical claim.

## Included files

- `expected-output.json`
- `large_n.py`
- `large_n.svg`
- `requirements.txt`
- `README.md`: these instructions.
- `manifest.json`: file sizes and SHA-256 hashes for every other bundle file.

The original individual downloads remain available from the lab page.
The archive preserves their bytes. `manifest.json` does not list itself;
its `files` list covers the scientific payload and this README.