Skip to main content

graphkind-harness

Benchmark de invariantes estructurales sobre universos (motor GraphKind). Mide, sobre corpus no isomorfos, lo que el paper mide:

  • Colisiones y frontera por observador y por n (las curvas C_n).
  • Anclas: la tabla de separación de los pares canónicos.
  • T4 por universo: estándar (1-WL y k-FWL), multicapa (total vs por capa) e hipergrafo k-uniforme.
  • Controles: invarianza bajo relabeling + control negativo (un observador débil debe colisionar).

Todo sale de un JSON congelable (versión del motor, corpus, semilla, wall time) y es determinista.

Instalar

pip install graphkind-harness
# para el corpus exhaustivo (no isomorfos):
sudo apt install nauty        # provee `geng`

Uso (CLI)

# frontera completa hasta n=7 sobre todos los grafos no isomorfos
graphkind-harness run --corpus geng --n-max 7 --out freeze.json

# solo las anclas (no necesita geng)
graphkind-harness anclas

# reporte de texto de un freeze
graphkind-harness reporte freeze.json

Salida de anclas:

par         n     grado   wl      kw2     kw3     ir1     ir2
C6 vs 2·C3  6     separa  colisiona separa  separa  separa  separa
Petersen vs prisma 10 ...
Rook vs Shrikhande 16 ...

Uso (librería)

from graphkind_harness import bench

res = bench.run(corpus="geng", n_max=7, observadores="wl,kw2,kw3")
print(bench.reporte(res))
print(res["frontera"]["kw3"])      # colisiones por n

Corpus

nombre qué es
geng todos los grafos no isomorfos hasta --n-max (nauty)
pairs los pares canónicos (C₆/2·C₃, Petersen/prisma, Rook/Shrikhande)
random m grafos G(n,p) con semilla
file un archivo: graph6 o n u v u v … por línea
all todos los grafos etiquetados (solo pruebas chicas)

Observadores

grado (control negativo), wl, kw2, kw3, ir1, ir2. El motor no incluye el observador SNF del Laplaciano (línea del laboratorio, EXP-161); se puede sumar registrando una función en observers.OBSERVADORES.

Qué NO es

  • No es el motor: la matemática vive en graphkind (o graphkind-wl).
  • No mide tiempo de cómputo por observador (solo colisiones y leyes).
  • No promete completitud universal: mide en la clase del corpus.

Cita y licencia

Argaña Silguero, J. (2026). GraphKind — Universes v2. Zenodo. https://doi.org/10.5281/zenodo.22747350 · CC-BY-4.0.

Metadata

Release files for graphkind-harness 0.1.1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for graphkind-harness 0.1.1
File Size Uploaded
graphkind_harness-0.1.1.tar.gz 12.9 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for graphkind-harness 0.1.1
File Interpreter ABI Platform
graphkind_harness-0.1.1-py3-none-any.whl Python 3 none any Details

Total release size: 24.8 kB

Release files / graphkind_harness-0.1.1.tar.gz

Download URL graphkind_harness-0.1.1.tar.gz
Size 12.9 kB
Tags Source
SHA-256 checksum
How to use checksums
0d11ed99afdb81224c8b9b471835639b96867b974c7296a0444215e9bc70ea9f
BLAKE2b-256 checksum
How to use checksums
40b49288dee224f22768e95be876b5a9f934af6be714a6fa6a710f806ebaa674
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.12.3

Release files / graphkind_harness-0.1.1-py3-none-any.whl

Download URL graphkind_harness-0.1.1-py3-none-any.whl
Size 11.9 kB
Tags Python 3
SHA-256 checksum
How to use checksums
3ab29c1033e307fd1b67a3d06b5b69fa76ff3f72667f602a786ed0d9ce2be735
BLAKE2b-256 checksum
How to use checksums
a9b102c47137647d078bf08612e2dbdba402f9172ca00c797065ce91958b103b
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.12.3

Release history Release notifications | RSS feed

This release

0.1.1 This release

2 release files

0.1.0

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page