"""Render measured medians/ranges; never substitute estimates for failed samples.""" from collections import defaultdict import json from pathlib import Path import statistics import matplotlib matplotlib.use('Agg') import matplotlib.pyplot as plt root=Path(__file__).resolve().parents[1] rows=json.loads((root/'evidence/scaling/summary.json').read_text(encoding='utf-8'))['results'] groups=defaultdict(list) for row in rows: groups[(row['label'],row['concurrency'])].append(row) fig,axes=plt.subplots(2,2,figsize=(12,8),layout='constrained') colors=['#2563eb','#e87924','#16a085'] for col,mode in enumerate(('fixed','growing')): for n,color in zip((1,2,4),colors): med,low,high,lat=[],[],[],[] for c in (8,24,64): values=[r['successful_rps'] for r in groups[(f'{mode}-{n}',c)]] center=statistics.median(values) med.append(center); low.append(center-min(values)); high.append(max(values)-center) lat.append(statistics.median(r['p95_ms'] for r in groups[(f'{mode}-{n}',c)])) axes[0,col].errorbar([8,24,64],med,yerr=[low,high],marker='o',capsize=4,color=color,label=f'{n} replicas') axes[1,col].plot([8,24,64],lat,marker='o',color=color) axes[0,col].set_title('Fixed total: 1 CPU / 768 MiB' if mode=='fixed' else 'Growing: 0.5 CPU / 192 MiB each') for ax in axes[:,col]: ax.set_xlabel('Concurrent clients'); ax.set_xticks([8,24,64]); ax.grid(alpha=.2) axes[0,col].set_ylabel('Successful HTTP requests / second'); axes[0,col].legend() axes[1,col].set_ylabel('p95 of all attempts (ms)') fig.suptitle('Lab 3: observed scaling; bars show min/max across repeats',fontsize=15) for suffix in ('png','svg'): path=root/('evidence/scaling.'+suffix) fig.savefig(path,dpi=160) if suffix=='svg': path.write_text('\n'.join(line.rstrip() for line in path.read_text(encoding='utf-8').splitlines())+'\n',encoding='utf-8') lines=['| Configuration | Clients | RPS median [min, max] | p95 median ms | Errors / attempts |', '|---|---:|---:|---:|---:|'] for (label,c),items in groups.items(): rps=[i['successful_rps'] for i in items] lines.append(f'| {label} | {c} | {statistics.median(rps):.2f} [{min(rps):.2f}, {max(rps):.2f}] | {statistics.median(i["p95_ms"] for i in items):.1f} | {sum(i["errors"] for i in items)} / {sum(i["attempts"] for i in items)} |') (root/'evidence/scaling/table.md').write_text('\n'.join(lines)+'\n',encoding='utf-8') print('\n'.join(lines))