rgevolve 0.2.0

rgevolve metapackage: install all rgevolve components at once

rgevolve - Renormalization Group Evolution Matrices for the SMEFT and the WET

rgevolve is the meta-package of a set of Python namespace packages for fast renormalization group (RG) evolution of Wilson coefficients in the Standard Model Effective Field Theory (SMEFT) and the Weak Effective Theory (WET) using the evolution matrix formalism. Installing it pulls in the core runtime and every currently available EFT/basis companion package in one go.

The packages are organized as follows:

• rgevolve-core: Core functionality for loading and processing evolution matrices.
• rgevolve.{eft}.{basis}: Precomputed evolution matrices for specific EFTs and bases, e.g., rgevolve.smeft.warsaw for the SMEFT in the Warsaw basis.
• rgevolve: A meta-package that installs the core and all available EFT/basis packages.

Installation

To install the core package and all available EFT/basis packages, use pip:

pip install rgevolve

To install only the core package, use:

pip install rgevolve-core

To install a specific EFT/basis package, use:

pip install rgevolve.{eft}.{basis}  # e.g., pip install rgevolve.smeft.warsaw

Available EFT/basis packages

The current lockstep release bundles the following companion distributions (scales, sector counts and Wilson-coefficient counts are read directly from each package's data.h5):

package	EFT	basis	scales (GeV)	sectors	# WCs
rgevolve.smeft.warsaw	SMEFT	Warsaw	91.1876 … 1000000 (10 pts)	11	2511
rgevolve.smeft.warsaw_up	SMEFT	Warsaw up	91.1876 … 1000000 (10 pts)	11	2511
rgevolve.wet.flavio	WET	flavio	2 … 91.1876 (13 pts)	125	3127
rgevolve.wet.jms	WET	JMS	2 … 91.1876 (13 pts)	219	4277
rgevolve.wet_4.flavio	WET-4	flavio	1.3 … 4.2 (4 pts)	81	1859
rgevolve.wet_4.jms	WET-4	JMS	1.3 … 4.2 (4 pts)	143	2545
rgevolve.wet_3.flavio	WET-3	flavio	0.77526 … 2 (5 pts)	33	720
rgevolve.wet_3.jms	WET-3	JMS	0.77526 … 2 (5 pts)	57	868

Usage

The main entry point is rgevolve.tools.run_and_match. It returns a 2-D NumPy array — a matrix that propagates a vector of input Wilson coefficients (at scale_in in (eft_in, basis_in)) to a vector of output Wilson coefficients (at scale_out in (eft_out, basis_out)), composing all relevant RG evolution, matching (when eft_in != eft_out), and basis translation (when basis_in != basis_out).

Signature

from rgevolve.tools import run_and_match

run_and_match(
    eft_in: str,
    eft_out: str,
    basis_in: str,
    basis_out: str,
    scale_in: float,
    scale_out: float,
    sector_out: str | None = None,
    wcs_in: Sequence[tuple[str, str]] | None = None,
    wcs_out: Sequence[tuple[str, str]] | None = None,
) -> numpy.ndarray

Arguments

• eft_in, eft_out — EFT identifiers, one of 'SMEFT', 'WET', 'WET-4', 'WET-3'. Matching is supported in the downward direction SMEFT → WET → WET-4 → WET-3.
• basis_in, basis_out — basis names within each EFT, e.g. 'Warsaw' / 'Warsaw up' (SMEFT) or 'JMS' / 'flavio' (WET / WET-3 / WET-4).
• scale_in, scale_out — initial and final renormalization scales, in GeV.
• sector_out — optional sector name (a block of Wilson coefficients). When given, the call runs in per-sector mode; the returned matrix is the full block for that output sector. When None, the call runs in full-EFT (cross-sector) mode and both wcs_in and wcs_out are required.
• wcs_in, wcs_out — optional lists of Wilson coefficients given as (name, kind) tuples, where kind is 'R' (real part, or a naturally real coefficient) or 'I' (imaginary part of a complex coefficient). In per-sector mode they subset and reorder the matrix's rows (wcs_out) and columns (wcs_in). In full-EFT mode both are mandatory and the assembled matrix has shape (len(wcs_out), len(wcs_in)); entries connecting Wilson coefficients in non-matched sectors are exactly zero.

Per-sector mode

Build the full evolution matrix for one output sector — here, the SMEFT-to-WET map for the sb sector (b→s transitions) from 1 TeV down to the b-quark scale, expressed in the Warsaw basis at the high scale and the flavio basis at the low scale:

from rgevolve.tools import run_and_match

M = run_and_match(
    eft_in='SMEFT', basis_in='Warsaw',
    eft_out='WET',  basis_out='flavio',
    scale_in=1000.0,            # GeV
    scale_out=4.2,              # GeV (b-quark mass scale)
    sector_out='sb',
)
# M.shape == (n_out, n_in)
# n_out: basis size of `sector_out` in (eft_out, basis_out)
# n_in:  basis size of the implied input sector in (eft_in, basis_in)

Restrict rows / columns to specific Wilson coefficients:

M_sub = run_and_match(
    eft_in='SMEFT', basis_in='Warsaw',
    eft_out='WET',  basis_out='flavio',
    scale_in=1000.0, scale_out=4.2,
    sector_out='sb',
    wcs_out=[('C9_bsmumu', 'R'), ('C10_bsmumu', 'R')],
    wcs_in =[('lq1_2223', 'R'), ('lq3_2223', 'R')],
)
# M_sub.shape == (len(wcs_out), len(wcs_in)) == (2, 2)

Full-EFT (cross-sector) mode

Specify both input and output Wilson coefficient lists; they may span multiple sectors. Cross-sector entries (where no matching/evolution path connects the two Wilson coefficients) are exactly zero:

M = run_and_match(
    eft_in='SMEFT', basis_in='Warsaw',
    eft_out='WET',  basis_out='flavio',
    scale_in=1000.0, scale_out=4.2,
    sector_out=None,
    wcs_out=[('C9_bsmumu', 'R'), ('C9_bdmumu', 'R')],
    wcs_in =[('lq1_2223', 'R'), ('lq1_2213', 'R')],
)
# M.shape == (2, 2); block-diagonal across sectors.

Discovering what's available at runtime

from rgevolve.tools.functions import bases_installed
from rgevolve.tools import get_wc_basis

bases_installed                       # {'SMEFT': ['Warsaw', ...], 'WET': [...], ...}
get_wc_basis('SMEFT', 'Warsaw',       # canonical (name, kind) list of one sector
             sector='dB=de=dmu=dtau=0')

Citation

If you use rgevolve in a scientific publication, please cite:

A. Smolkovič, P. Stangl

"Differentiable Multi-scale Effective Field Theory Likelihoods for Beyond the Standard Model Phenomenology"

arXiv:2603.15801

Bugs and feature requests

Please report bugs and request features via the GitHub issues pages of the relevant rgevolve repository.

Contributors

Authors:

• Aleks Smolkovič (@alekssmolkovic)
• Peter Stangl (@peterstangl)

License

rgevolve is licensed under the MIT License — see LICENSE.