Jose-dev commited on
Commit
a4e813a
·
1 Parent(s): 52c0bb0

datatables: tablas cargables (ley T4, clasificación, verificaciones) + configs y sección Uso en la card

Browse files
README.md CHANGED
@@ -22,6 +22,19 @@ size_categories:
22
  - n<1K
23
  thumbnail: assets/banner.png
24
  doi: 10.5281/zenodo.22747350
 
 
 
 
 
 
 
 
 
 
 
 
 
25
  ---
26
 
27
  <p align="center">
@@ -41,6 +54,30 @@ doi: 10.5281/zenodo.22747350
41
 
42
  ---
43
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
44
  ## Abstract
45
 
46
  We classify the operators under which the **color refinement** (1-WL)
 
22
  - n<1K
23
  thumbnail: assets/banner.png
24
  doi: 10.5281/zenodo.22747350
25
+ configs:
26
+ - config_name: ley_t4
27
+ data_files:
28
+ - split: train
29
+ path: data/ley-t4-exp102.csv
30
+ - config_name: clasificacion
31
+ data_files:
32
+ - split: train
33
+ path: data/clasificacion-exp105.csv
34
+ - config_name: verificaciones
35
+ data_files:
36
+ - split: train
37
+ path: data/verificaciones-v1-v19.csv
38
  ---
39
 
40
  <p align="center">
 
54
 
55
  ---
56
 
57
+ ## What's in this dataset?
58
+
59
+ This repository is the **mirror of [graphkind-universos-v2](https://github.com/cripto-bot/graphkind-universos-v2)** — paper, engine and frozen results of the *observation universes* arc (T4). The original frozen JSONs live untouched in [`resultados/`](resultados); the three tables below are derived from them and are **loadable with `datasets`**:
60
+
61
+ | config | rows | what it is | source |
62
+ |---|---|---|---|
63
+ | `ley_t4` | 34 | the sweep of 17 count-compressions × 2 involutions (EXP-102): `VIVE`/`MUERE`, violations, counterexample witness, and `compatible` — the second hypothesis, refuted by the table | `resultados/EXP-102_results_frozen.json` |
64
+ | `clasificacion` | 6 | the six involutions classified on the atlas (EXP-105): `trivial` / `dual` / `local` with the formal conditions (well-defined, injective, `g = id`) | `resultados/EXP-105_results_frozen.json` |
65
+ | `verificaciones` | 19 | the V1–V19 verification suite with the exact asserted numbers — every headline number of the paper is anchored here | this README, §7 / §7c |
66
+
67
+ ### Usage
68
+
69
+ ```python
70
+ from datasets import load_dataset
71
+
72
+ ley = load_dataset("Jose-dev/graphkind-universos", "ley_t4")["train"] # 34 universes
73
+ clas = load_dataset("Jose-dev/graphkind-universos", "clasificacion")["train"] # 6 involutions
74
+ verif = load_dataset("Jose-dev/graphkind-universos", "verificaciones")["train"] # 19 certificates
75
+ ```
76
+
77
+ The tables are regenerated deterministically (stdlib only) with `python3 tools/build_datatables.py`.
78
+
79
+ ---
80
+
81
  ## Abstract
82
 
83
  We classify the operators under which the **color refinement** (1-WL)
data/clasificacion-exp105.csv ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ involution,class,well_defined,injective,g_identity
2
+ comp,dual,True,True,False
3
+ comp_parcial,local,False,False,False
4
+ flip_v0,local,False,False,False
5
+ swap01,trivial,True,True,True
6
+ reverse,trivial,True,True,True
7
+ comp_rev,dual,True,True,False
data/ley-t4-exp102.csv ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ f,involution,t4,violations,compatible,n,witness_mask
2
+ id,T-M,VIVE,0,True,6,
3
+ id,T+M mod2,VIVE,0,True,6,
4
+ mod2,T-M,VIVE,0,False,6,
5
+ mod2,T+M mod2,VIVE,0,True,6,
6
+ mod3,T-M,VIVE,0,False,6,
7
+ mod3,T+M mod2,VIVE,0,False,6,
8
+ mod4,T-M,VIVE,0,False,6,
9
+ mod4,T+M mod2,VIVE,0,True,6,
10
+ mod5,T-M,VIVE,0,True,6,
11
+ mod5,T+M mod2,VIVE,0,True,6,
12
+ umbral1,T-M,MUERE,1080,False,6,657
13
+ umbral1,T+M mod2,MUERE,1080,False,6,657
14
+ umbral2,T-M,VIVE,0,False,6,
15
+ umbral2,T+M mod2,VIVE,0,False,6,
16
+ umbral3,T-M,MUERE,720,False,6,692
17
+ umbral3,T+M mod2,MUERE,720,False,6,692
18
+ umbral4,T-M,MUERE,1080,False,6,657
19
+ umbral4,T+M mod2,MUERE,1080,False,6,657
20
+ trunc1,T-M,MUERE,1080,False,6,657
21
+ trunc1,T+M mod2,MUERE,1080,False,6,657
22
+ trunc2,T-M,VIVE,0,False,6,
23
+ trunc2,T+M mod2,VIVE,0,False,6,
24
+ trunc3,T-M,VIVE,0,False,6,
25
+ trunc3,T+M mod2,VIVE,0,False,6,
26
+ log2,T-M,VIVE,0,False,6,
27
+ log2,T+M mod2,VIVE,0,False,6,
28
+ cuadrado,T-M,VIVE,0,True,6,
29
+ cuadrado,T+M mod2,VIVE,0,True,6,
30
+ cubo,T-M,VIVE,0,True,6,
31
+ cubo,T+M mod2,VIVE,0,True,6,
32
+ raiz,T-M,MUERE,1080,False,6,657
33
+ raiz,T+M mod2,MUERE,1080,False,6,657
34
+ const1,T-M,MUERE,1080,False,6,657
35
+ const1,T+M mod2,MUERE,1080,False,6,657
data/verificaciones-v1-v19.csv ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ id,verification,assert
2
+ V1,"the law f(1)≠f(2) (17 f × 2 involutions, n≤6)",34/34
3
+ V2,characterization n=4 exhaustive,96 and 48
4
+ V3,T4_k complement-invariance (k-FWL),51/51
5
+ V4,observer cascade n≤7,26 → 0
6
+ V5,Rook vs Shrikhande (k-FWL; 3-WL std does NOT),k*=3
7
+ V6,deterministic Z₃,does not hold
8
+ V7,random R1,204/1096
9
+ V8,R3 measure dependence,96/528
10
+ V9,autonomous search,cycle 99.63%
11
+ V10,hash certificate (12-hex == 64-hex),no collisions
12
+ V11,convergence certificate (stable == n+2),OK
13
+ V12,classification constant across graphs (n=4),OK
14
+ V13,"multilayer: total holds, per-layer dies in the canonical variant, channels hold",4160/4160 · failures > 0 · 8320/8320
15
+ V14,3-uniform hypergraphs: T4 holds,6042/6042 (n≤5 + n=6 sample)
16
+ V15,per-arity dual: partition + saving,1024/1024 · 5120→3392 edges
17
+ V16,mixed arities: per-arity holds (repair),1024/1024
18
+ V17,complement quotient (T4's price),"52 graphs · 24 pairs (SG-01 n≤8: 6,168)"
19
+ V18,"IR: IR_1 does not separate Rook/Shri, IR_2 does",i\*=2
20
+ V19,"SG-04 certificate: n=10 exhaustive, no failures","12,005,168 · 0 collisions"
tools/build_datatables.py ADDED
@@ -0,0 +1,84 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python3
2
+ """Genera las tablas cargables del dataset a partir de los freezes congelados.
3
+
4
+ Escribe (determinista, solo stdlib):
5
+ data/ley-t4-exp102.csv <- resultados/EXP-102_results_frozen.json (34 universos)
6
+ data/clasificacion-exp105.csv <- resultados/EXP-105_results_frozen.json (6 involuciones)
7
+ data/verificaciones-v1-v19.csv <- tabla V1..V19 de README.md
8
+
9
+ Uso: python3 tools/build_datatables.py
10
+ """
11
+ import csv
12
+ import json
13
+ import re
14
+ import sys
15
+ from pathlib import Path
16
+
17
+ ROOT = Path(__file__).resolve().parents[1]
18
+ DATA = ROOT / "data"
19
+
20
+
21
+ def write_csv(path, header, rows):
22
+ with path.open("w", encoding="utf-8", newline="") as fh:
23
+ writer = csv.writer(fh, lineterminator="\n")
24
+ writer.writerow(header)
25
+ writer.writerows(rows)
26
+ print(f"{path.relative_to(ROOT)}: {len(rows)} filas")
27
+
28
+
29
+ def ley_t4():
30
+ src = ROOT / "resultados/EXP-102_results_frozen.json"
31
+ tabla = json.loads(src.read_text(encoding="utf-8"))["tabla"]
32
+ rows = []
33
+ for clave, v in tabla.items():
34
+ f_nombre, iota = clave.split(" x ", 1)
35
+ testigo = v.get("testigo") or {}
36
+ rows.append([f_nombre, iota, v["T4"], v["violaciones"],
37
+ v["compatible"], v["n"], testigo.get("mask", "")])
38
+ assert len(rows) == 34, f"EXP-102: esperaba 34 universos, hay {len(rows)}"
39
+ write_csv(DATA / "ley-t4-exp102.csv",
40
+ ["f", "involution", "t4", "violations", "compatible", "n",
41
+ "witness_mask"], rows)
42
+
43
+
44
+ def clasificacion():
45
+ src = ROOT / "resultados/EXP-105_results_frozen.json"
46
+ d = json.loads(src.read_text(encoding="utf-8"))
47
+ rows = [[k, v["clase"], v["bien_definida"], v["inyectiva"],
48
+ v["g_identidad"]] for k, v in d["clasificacion"].items()]
49
+ assert len(rows) == 6, f"EXP-105: esperaba 6 involuciones, hay {len(rows)}"
50
+ write_csv(DATA / "clasificacion-exp105.csv",
51
+ ["involution", "class", "well_defined", "injective",
52
+ "g_identity"], rows)
53
+
54
+
55
+ def _limpio(celda):
56
+ return celda.replace("**", "").replace("`", "").strip()
57
+
58
+
59
+ def verificaciones():
60
+ readme = (ROOT / "README.md").read_text(encoding="utf-8")
61
+ rows = []
62
+ for line in readme.splitlines():
63
+ m = re.match(r"^\|\s*V(\d+)\s*\|(.+?)\|(.+?)\|\s*$", line)
64
+ if m:
65
+ rows.append((int(m.group(1)), _limpio(m.group(2)),
66
+ _limpio(m.group(3))))
67
+ rows.sort()
68
+ ids = [r[0] for r in rows]
69
+ assert ids == list(range(1, 20)), f"README: esperaba V1..V19, hay {ids}"
70
+ write_csv(DATA / "verificaciones-v1-v19.csv",
71
+ ["id", "verification", "assert"],
72
+ [[f"V{i}", desc, esperado] for i, desc, esperado in rows])
73
+
74
+
75
+ def main():
76
+ DATA.mkdir(exist_ok=True)
77
+ ley_t4()
78
+ clasificacion()
79
+ verificaciones()
80
+ return 0
81
+
82
+
83
+ if __name__ == "__main__":
84
+ sys.exit(main())