Try it in your browser#

Run ramanujan-qft without installing anything. The console below runs the package with Pyodide, a build of CPython for WebAssembly. Everything runs locally in your browser; no code or result is sent to a server.

  • Pick an example or write your own code, then press Run (or Ctrl+Enter). The value of the last line is shown, as in a notebook, and variables persist between runs.

  • The first run downloads Python, NumPy, mpmath and SymPy (tens of MB). Your browser caches them, so later visits start faster.

  • Stop ends a long calculation. Python then restarts on the next run.

  • Three reference-checked calculations runs a one-loop parameter check, exports a two-loop RMT derivation, and compares a hybrid three-loop result with a Fourier reference. It loads SciPy on first use. See Choose a method and check your first result for the full walkthrough.

  • To see the Ramanujan method on a coupled integral, choose Two-loop sunset: graph to multivariate RMT. It displays the exponent matrix, Gamma families, result and error diagnostics. The separate Taylor example covers equal-mass vacuum and checks it against a Clausen reference.

The console needs JavaScript.

Good to know#

  • Speed. WebAssembly is slower than native Python, typically by a factor of a few. Most examples take well under a second once Python is loaded. sunset_subtracted and sunset_d2 take 10–20 s natively, so expect minutes here.

  • Version. The console installs the package built from the same commit as this documentation.

  • Limits. One calculation runs at a time. Files and plots are not supported. NumPy, mpmath and SymPy are loaded before the package starts. Other Pyodide packages are loaded when your code imports them explicitly. For the hybrid graph evaluator, include import scipy before calling integrate_graph; its optional SciPy dependency is not needed for the sunset series examples.

  • If it does not start: the console needs a recent browser with WebAssembly and access to cdn.jsdelivr.net, where Pyodide is hosted. Corporate networks sometimes block it. Use Colab or a local installation instead.

Run it in Google Colab#

Open in Colab

The quickstart notebook installs the package in a Colab session and walks through the main functions. Colab runs native Python, so it is faster than the browser console.

Install it#

For your own calculations, install the package locally (Installation):

pip install --pre ramanujan-qft