1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
|
#!/usr/bin/env python3
from __future__ import annotations
import argparse
import csv
import math
from collections import defaultdict
from pathlib import Path
import matplotlib.pyplot as plt
DEFAULT_BACKBONES = [
"gin",
"gine",
"gcn",
"graphsage",
"gatv2",
"graphconv",
"transformer",
"pna",
"gen",
"film",
"resgated",
"tag",
"sgc",
"cheb",
"arma",
"mf",
"appnp",
]
DEFAULT_DATASETS = [
"zinc-cycle56",
"ogbg-molhiv",
"ogbg-molbbbp",
"ogbg-molbace",
"ogbg-moltox21",
"ogbg-molclintox",
"ogbg-molsider",
"ogbg-molesol",
"ogbg-molfreesolv",
"ogbg-mollipo",
]
def mean(xs: list[float]) -> float:
return sum(xs) / len(xs) if xs else math.nan
def parse_list(value: str) -> list[str]:
return [x.strip() for x in value.replace(",", " ").split() if x.strip()]
def parse_label_overrides(value: str) -> dict[str, str]:
overrides = {}
for item in value.split(","):
item = item.strip()
if not item:
continue
if "=" not in item:
raise ValueError(f"label override must be DATASET=LABEL, got {item!r}")
dataset, label = item.split("=", 1)
overrides[dataset.strip().lower()] = label.strip()
return overrides
def load_cells(
path: Path,
compute_label: str,
datasets: list[str],
label_overrides: dict[str, str],
) -> dict[tuple[str, str], float]:
requested = {x.lower() for x in datasets}
values: dict[tuple[str, str], list[float]] = defaultdict(list)
with path.open() as f:
for row in csv.DictReader(f):
dataset = row["dataset"].lower()
if requested and dataset not in requested:
continue
expected_label = label_overrides.get(dataset, compute_label)
if row["compute_label"] != expected_label:
continue
values[(dataset, row["view"])].append(float(row["test_delta"]))
return {key: mean(xs) for key, xs in values.items()}
def render_table(
cells: dict[tuple[str, str], float],
datasets: list[str],
backbones: list[str],
title: str,
out_path: Path,
digits: int,
) -> None:
n_rows = max(1, len(datasets))
n_cols = max(1, len(backbones))
fig_w = max(12.0, 1.05 * n_cols + 2.8)
fig_h = max(2.4, 0.72 * n_rows + 1.8)
fig, ax = plt.subplots(figsize=(fig_w, fig_h), dpi=220)
ax.axis("off")
data = []
for dataset in datasets:
row = []
for backbone in backbones:
value = cells.get((dataset.lower(), backbone))
if value is None or math.isnan(value) or value < 0:
row.append("-")
else:
row.append(f"{value:+.{digits}f}")
data.append(row)
table = ax.table(
cellText=data,
rowLabels=datasets,
colLabels=backbones,
cellLoc="center",
rowLoc="center",
loc="center",
)
table.auto_set_font_size(False)
table.set_fontsize(8.5)
table.scale(1.0, 1.65)
header_bg = "#f1f5f9"
edge = "#cbd5e1"
positive = "#15803d"
missing = "#64748b"
for (row_idx, col_idx), cell in table.get_celld().items():
cell.set_edgecolor(edge)
cell.set_linewidth(0.7)
text = cell.get_text()
if row_idx == 0 or col_idx == -1:
cell.set_facecolor(header_bg)
text.set_fontweight("bold")
text.set_color("#0f172a")
continue
raw = text.get_text()
if raw == "-":
text.set_text("-")
text.set_color(missing)
continue
value = float(raw)
if value > 0:
text.set_color(positive)
text.set_fontweight("bold")
else:
text.set_color("#334155")
ax.set_title(title, fontsize=13, fontweight="bold", pad=14)
fig.tight_layout(pad=0.8)
out_path.parent.mkdir(parents=True, exist_ok=True)
fig.savefig(out_path, bbox_inches="tight", facecolor="white")
plt.close(fig)
def main() -> None:
ap = argparse.ArgumentParser()
ap.add_argument("--deltas", default="analysis/paired_deltas.csv")
ap.add_argument("--compute-label", default="fixed-rrog-T1-ns3")
ap.add_argument(
"--label-overrides",
default="zinc-cycle56=fixed-rrog-T1-ns3+trace",
help="comma-separated DATASET=COMPUTE_LABEL overrides",
)
ap.add_argument("--datasets", default=" ".join(DEFAULT_DATASETS))
ap.add_argument("--backbones", default=" ".join(DEFAULT_BACKBONES))
ap.add_argument("--out", default="analysis/fixed_rrog_delta_matrix.png")
ap.add_argument("--title", default="")
ap.add_argument("--digits", type=int, default=3)
args = ap.parse_args()
datasets = parse_list(args.datasets)
backbones = parse_list(args.backbones)
label_overrides = parse_label_overrides(args.label_overrides)
cells = load_cells(Path(args.deltas), args.compute_label, datasets, label_overrides)
title = args.title or "Test Delta vs Classic: fixed RRoG"
render_table(cells, datasets, backbones, title, Path(args.out), args.digits)
print(f"wrote {args.out}")
if __name__ == "__main__":
main()
|
