benchmark/generate-chart.py at main · codemod/benchmark · GitHub

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
185
186
187
188
189
190
191
192
193
#!/usr/bin/env python3
"""Generate a beautiful benchmark comparison chart."""
import os
import matplotlib.pyplot as plt
import matplotlib.ticker as ticker
import numpy as np

# ── Data ──────────────────────────────────────────────────────────
tools = ['ast-grep', 'GritQL', 'recast', 'jscodeshift', 'semgrep']

tasks = {
    'Search console.log': [43, 80, 200, 754, 7535],
    'Transform var → const': [41, 816, 200, 808, 13111],
    'Search async + try/catch': [44, 1486, 199, 764, 1433],
}

memory = {
    'Search console.log': [11, 58, 93, 153, 250],
    'Transform var → const': [10, 57, 95, 157, 327],
    'Search async + try/catch': [11, 58, 95, 154, 231],
}

# ── Colors ────────────────────────────────────────────────────────
colors = ['#3b82f6', '#22c55e', '#a855f7', '#f59e0b', '#ef4444']
bg_color = '#0f172a'
card_color = '#1e293b'
text_color = '#e2e8f0'
grid_color = '#334155'
accent_color = '#94a3b8'

# ── Figure ────────────────────────────────────────────────────────
fig = plt.figure(figsize=(18, 11), facecolor=bg_color)

# Title — plenty of room at top
fig.text(0.5, 0.98, 'AST Tool Benchmark', ha='center', va='top',
         fontsize=30, fontweight='bold', color=text_color,
         fontfamily='sans-serif')
fig.text(0.5, 0.945, '500 files  ·  75,000 lines  ·  median of 3 runs  ·  Apple Silicon',
         ha='center', va='top', fontsize=14, color=accent_color,
         fontfamily='sans-serif')

task_names = list(tasks.keys())
n_tools = len(tools)
bar_height = 0.52

# Use gridspec for precise control
gs = fig.add_gridspec(2, 3, left=0.08, right=0.96, top=0.87, bottom=0.06,
                      hspace=0.45, wspace=0.35)

# ── Top row: 3 task charts ───────────────────────────────────────
for col, task in enumerate(task_names):
    ax = fig.add_subplot(gs[0, col])
    ax.set_facecolor(card_color)

    times = tasks[task]
    y_pos = np.arange(n_tools)

    bars = ax.barh(y_pos, times, height=bar_height, color=colors,
                   edgecolor='none', zorder=3)

    for bar, c in zip(bars, colors):
        bar.set_linewidth(0)

    # Value labels — always outside the bar
    for i, t in enumerate(times):
        ax.text(t * 1.15, i, f'{t:,}ms',
                ha='left', va='center', fontsize=11,
                fontweight='bold', color=colors[i], fontfamily='sans-serif')

    ax.set_yticks(y_pos)
    ax.set_yticklabels(tools, fontsize=12, color=text_color, fontfamily='sans-serif')
    ax.invert_yaxis()

    ax.set_title(task, fontsize=15, fontweight='bold', color=text_color,
                 fontfamily='sans-serif', pad=14)

    ax.set_xscale('log')
    ax.set_xlim(20, max(times) * 6)
    ax.xaxis.set_major_formatter(ticker.FuncFormatter(lambda x, _: f'{int(x):,}'))
    ax.tick_params(axis='x', colors=accent_color, labelsize=9)
    ax.tick_params(axis='y', length=0)

    ax.set_xlabel('Time (ms, log scale)', fontsize=10, color=accent_color,
                  fontfamily='sans-serif')

    ax.spines['top'].set_visible(False)
    ax.spines['right'].set_visible(False)
    ax.spines['bottom'].set_color(grid_color)
    ax.spines['left'].set_visible(False)
    ax.grid(axis='x', color=grid_color, linewidth=0.5, zorder=0)

# ── Bottom left: Memory chart ────────────────────────────────────
ax_mem = fig.add_subplot(gs[1, 0])
ax_mem.set_facecolor(card_color)

avg_mem = [np.mean([memory[t][i] for t in task_names]) for i in range(n_tools)]
y_pos = np.arange(n_tools)

bars = ax_mem.barh(y_pos, avg_mem, height=bar_height, color=colors,
                   edgecolor='none', zorder=3)

for i, m in enumerate(avg_mem):
    ax_mem.text(m + max(avg_mem) * 0.04, i, f'{m:.0f} MB',
                ha='left', va='center', fontsize=11,
                fontweight='bold', color=colors[i], fontfamily='sans-serif')

ax_mem.set_yticks(y_pos)
ax_mem.set_yticklabels(tools, fontsize=12, color=text_color, fontfamily='sans-serif')
ax_mem.invert_yaxis()
ax_mem.set_title('Peak Memory (avg across tasks)', fontsize=15, fontweight='bold',
                 color=text_color, fontfamily='sans-serif', pad=14)
ax_mem.set_xlim(0, max(avg_mem) * 1.5)
ax_mem.tick_params(axis='x', colors=accent_color, labelsize=9)
ax_mem.tick_params(axis='y', length=0)
ax_mem.set_xlabel('Memory (MB)', fontsize=10, color=accent_color, fontfamily='sans-serif')
ax_mem.spines['top'].set_visible(False)
ax_mem.spines['right'].set_visible(False)
ax_mem.spines['bottom'].set_color(grid_color)
ax_mem.spines['left'].set_visible(False)
ax_mem.grid(axis='x', color=grid_color, linewidth=0.5, zorder=0)

# ── Bottom center: Speedup multiplier chart ──────────────────────
ax_sp = fig.add_subplot(gs[1, 1])
ax_sp.set_facecolor(card_color)

geo_means = []
for i in range(n_tools):
    ratios = [tasks[t][i] / tasks[t][0] for t in task_names]
    geo = np.exp(np.mean(np.log(ratios)))
    geo_means.append(geo)

y_pos = np.arange(n_tools)
bars = ax_sp.barh(y_pos, geo_means, height=bar_height, color=colors,
                  edgecolor='none', zorder=3)

for i, g in enumerate(geo_means):
    label = f'{g:.0f}x' if g >= 2 else f'{g:.1f}x'
    ax_sp.text(g + max(geo_means) * 0.03, i, label,
               ha='left', va='center', fontsize=12,
               fontweight='bold', color=colors[i], fontfamily='sans-serif')

ax_sp.set_yticks(y_pos)
ax_sp.set_yticklabels(tools, fontsize=12, color=text_color, fontfamily='sans-serif')
ax_sp.invert_yaxis()
ax_sp.set_title('Slowdown vs ast-grep (geometric mean)', fontsize=15, fontweight='bold',
                color=text_color, fontfamily='sans-serif', pad=14)
ax_sp.set_xlim(0, max(geo_means) * 1.35)
ax_sp.tick_params(axis='x', colors=accent_color, labelsize=9)
ax_sp.tick_params(axis='y', length=0)
ax_sp.set_xlabel('Times slower', fontsize=10, color=accent_color, fontfamily='sans-serif')
ax_sp.spines['top'].set_visible(False)
ax_sp.spines['right'].set_visible(False)
ax_sp.spines['bottom'].set_color(grid_color)
ax_sp.spines['left'].set_visible(False)
ax_sp.grid(axis='x', color=grid_color, linewidth=0.5, zorder=0)

# ── Bottom right: Key takeaways ──────────────────────────────────
ax_info = fig.add_subplot(gs[1, 2])
ax_info.set_facecolor(card_color)
ax_info.set_xlim(0, 10)
ax_info.set_ylim(0, 10)
ax_info.axis('off')

# Rounded border effect
for spine in ax_info.spines.values():
    spine.set_visible(False)

takeaways = [
    ('●', colors[0], 'ast-grep: ~42ms avg, 11 MB peak'),
    ('●', colors[0], '5x faster than next fastest tool'),
    ('●', colors[0], 'No slowdown on complex patterns'),
    ('', '', ''),
    ('●', colors[1], 'GritQL: fast search, slower transforms'),
    ('●', colors[2], 'recast: consistent ~200ms baseline'),
    ('●', colors[3], 'jscodeshift: worker overhead (~770ms)'),
    ('●', colors[4], 'semgrep: built for CI, not speed'),
]

ax_info.text(0.3, 9.5, 'Key Takeaways', fontsize=16, fontweight='bold',
             color=text_color, fontfamily='sans-serif', va='top')

for i, (bullet, color, text) in enumerate(takeaways):
    y = 8.2 - i * 1.05
    if bullet:
        ax_info.text(0.3, y, bullet, fontsize=14, color=color,
                     fontfamily='sans-serif', va='center')
        ax_info.text(1.1, y, text, fontsize=12, color=text_color,
                     fontfamily='sans-serif', va='center')

# ── Save ──────────────────────────────────────────────────────────
fig.savefig(os.path.join(os.path.dirname(__file__), 'benchmark.png'), dpi=200,
            facecolor=bg_color, bbox_inches='tight', pad_inches=0.4)
print('Saved benchmark.png')