03. ノートブックでの定量データチャート描画
Drawlib には 10 種類の定量チャート(BarChart、LineChart、AreaChart、PieChart、RadarChart、BoxChart、ScatterChart、GanttChart、HeatmapChart、HeatmapOverlay)が用意されており、カードや注釈と同一キャンバス上に配置できます。
View Drawlib Python Code
from drawlib.canvas import save, setup
from drawlib.charts.bar import BarChart
from drawlib.charts.line import LineChart
from drawlib.shapes import rectangle
from drawlib.styles import Colors, Styles
setup(width=144, height=58)
card_bg = Styles.White.patch(shape_r=2.0, shape_line_color=Colors.Gray3)
rectangle((38.0, 29.0), width=64.0, height=48.0, style=card_bg)
rectangle((106.0, 29.0), width=64.0, height=48.0, style=card_bg)
bar = BarChart(
axis_line_style=Styles.MutedDashed,
categories=["Q1", "Q2", "Q3", "Q4"],
width=56.0,
height=34.0,
bar_mode="group",
bar_width_ratio=0.62,
bar_r=0.8,
axis_text_style=Styles.Muted.patch(text_size=10.0),
grid_style=Styles.MutedThin,
value_text_style=Styles.DarkBold.patch(text_size=10.0),
value_format="{:.1f}",
title="Throughput (k req/s)",
title_style=Styles.DarkBold.patch(text_size=11.5),
)
bar.add_series("Baseline", [14.0, 18.0, 21.0, 25.0], style=Styles.SecondaryNeutral)
bar.add_series("Cached", [24.0, 31.5, 39.0, 48.5], style=Styles.PrimaryFlat)
bar.configure_y_axis(min_value=0.0, max_value=60.0, tick_step=15.0)
bar.draw(xy=(10.0, 12.0))
line_chart = LineChart(
axis_line_style=Styles.MutedDashed,
categories=["Q1", "Q2", "Q3", "Q4"],
width=56.0,
height=34.0,
show_points=True,
point_size=0.75,
axis_text_style=Styles.Muted.patch(text_size=10.0),
grid_style=Styles.MutedThin,
title="P99 Tail Latency (ms)",
title_style=Styles.DarkBold.patch(text_size=11.5),
)
line_chart.add_series("P99 (ms)", [42.0, 34.0, 24.5, 16.0], style=Styles.PrimaryFlat, line_width=2.2)
line_chart.configure_y_axis(min_value=0.0, max_value=50.0, tick_step=10.0, format="{:.0f}", unit="ms")
line_chart.draw(xy=(78.0, 12.0))
save()
1. Python セルでメトリクスを集計する
ノートブックのセル間で変数が共有されるため、まず Python コードで生データを集計してから Drawlib のチャートに渡すことができます。
In [setup]:
!pip install -q drawlib
In [1]:
raw_latencies_ms = {
"v0.3.8": [48.2, 46.0, 49.5, 45.1],
"v0.3.9": [34.0, 32.5, 35.1, 31.8],
"v0.3.10": [22.4, 21.0, 23.1, 20.5],
"v0.3.11": [15.8, 14.9, 16.2, 14.5],
}
releases = list(raw_latencies_ms.keys())
mean_latencies = [round(sum(vals) / len(vals), 1) for vals in raw_latencies_ms.values()]
for rel, avg in zip(releases, mean_latencies, strict=True):
print(f"{rel}: mean={avg:.1f} ms")
mean_latencies
Out [1]:
v0.3.8: mean=47.2 ms v0.3.9: mean=33.4 ms v0.3.10: mean=21.8 ms v0.3.11: mean=15.3 ms
[47.2, 33.4, 21.8, 15.3]
2. 集計結果とサマリカードの組み合わせ描画
すべての Charts コンポーネントは 左下座標 (x0, y0) を基準に配置します。
In [2]:
from drawlib.canvas import show, setup
from drawlib.charts.bar import BarChart
from drawlib.shapes import rectangle
from drawlib.styles import Styles
from drawlib.text import text
setup(width=120, height=44, cwd=True)
chart = BarChart(
axis_line_style=Styles.MutedDashed,
categories=releases,
width=66.0,
height=30.0,
bar_width_ratio=0.58,
bar_r=0.8,
axis_text_style=Styles.Muted.patch(text_size=10.0),
grid_style=Styles.MutedThin,
value_text_style=Styles.DarkBold.patch(text_size=10.0),
value_format="{:.1f}",
title="Mean Rendering Latency by Release (ms)",
title_style=Styles.DarkBold.patch(text_size=11.5),
)
chart.add_series("Mean Latency", mean_latencies, style=Styles.PrimaryFlat)
chart.configure_y_axis(min_value=0.0, max_value=60.0, tick_step=15.0)
chart.draw(xy=(8.0, 8.0))
# Summary Callout Card on the Right (Center Anchor)
rectangle((97.0, 23.0), width=32.0, height=24.0, style=Styles.PrimaryFlat.patch(shape_r=2.0))
text((97.0, 29.5), "68% Faster", style=Styles.WhiteBold.patch(text_size=13))
text((97.0, 23.0), "v0.3.8 -> v0.3.11", style=Styles.White.patch(text_size=10.5))
text((97.0, 16.5), "15.4 ms avg", style=Styles.WhiteBold.patch(text_size=11))
show()
Out [2]:
![Output [2]](../images/03_data_charts/03_live_release_benchmark.png)