microsoft · arrayka · May 7, 2026 · May 8, 2026 · May 8, 2026 · May 8, 2026
diff --git a/.github/reports/build-stats-report.html b/.github/reports/build-stats-report.html
@@ -0,0 +1,196 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+  <meta charset="utf-8">
+  <title>DiskANN Release Build Stats</title>
+  <script src="https://cdn.jsdelivr.net/npm/vue@3/dist/vue.global.prod.js"></script>
+  <script src="https://cdn.jsdelivr.net/npm/chart.js@4"></script>
+  <script src="build-stats-report.js"></script>
+  <style>
+    body {
+      font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif;
+      max-width: 1400px; margin: 0 auto; padding: 20px; background: #fafafa;
+    }
+    h1 { color: #1b1e24; border-bottom: 2px solid #4e79a7; padding-bottom: 8px; }
+    h2 { color: #333; margin-top: 40px; }
+    .chart-container {
+      background: white; border-radius: 8px; padding: 20px;
+      box-shadow: 0 1px 3px rgba(0,0,0,0.12); margin-bottom: 30px;
+    }
+    canvas { max-height: 400px; }
+    pre {
+      background: #1b1e24; color: #d5dde5; padding: 16px; border-radius: 8px;
+      overflow-x: auto; font-size: 13px; line-height: 1.4;
+    }
+    .meta { color: #666; font-size: 14px; }
+    table { border-collapse: collapse; width: 100%; margin: 10px 0; }
+    th, td { border: 1px solid #ddd; padding: 8px 12px; text-align: left; }
+    th { background: #1b1e24; color: #d5dde5; }
+    tr:nth-child(even) { background: #f2f2f2; }
+  </style>
+</head>
+<body>
+<div id="app">
+  <h1>DiskANN Release Build Statistics</h1>
+  <p class="meta">Generated: {{ data.generated }} — Last 30 days ({{ data.dates.length }} runs)</p>
+
+  <h2>Total Build Time Trend</h2>
+  <div class="chart-container"><canvas ref="totalBuildTime"></canvas></div>
+
+  <h2>Build Time by Crate (Top 15)</h2>
+  <div class="chart-container"><canvas ref="crateBuildTime"></canvas></div>
+
+  <h2>Total Binary Size Trend</h2>
+  <div class="chart-container"><canvas ref="totalBinarySize"></canvas></div>
+
+  <h2>Binary Size per Binary</h2>
+  <div class="chart-container"><canvas ref="perBinarySize"></canvas></div>
+
+  <h2>Latest Cargo Bloat (diskann-benchmark)</h2>
+  <pre>{{ data.latest_cargo_bloat || 'No cargo bloat data available.' }}</pre>
+
+  <h2>Latest LLVM Lines (diskann-benchmark)</h2>
+  <pre>{{ data.latest_cargo_llvm_lines || 'No cargo llvm-lines data available.' }}</pre>
+
+  <template v-if="data.latest_run">
+    <h2>Latest Build Details (Top 20 Crates)</h2>
+    <p class="meta">
+      Run: {{ data.latest_run.created_at }} —
+      Commit: <code>{{ data.latest_run.head_sha }}</code> —
+      Total wall time: {{ data.latest_run.total_time_display }}
+    </p>
+
+    <table v-if="data.latest_run.units.length">
+      <thead><tr><th>#</th><th>Crate</th><th>Version</th><th>Duration</th></tr></thead>
+      <tbody>
+        <tr v-for="(u, i) in data.latest_run.units.slice(0, 20)" :key="i">
+          <td>{{ i + 1 }}</td>
+          <td>{{ u.name }}</td>
+          <td>{{ u.version }}</td>
+          <td>{{ u.duration.toFixed(1) }}s</td>
+        </tr>
+      </tbody>
+    </table>
+
+    <template v-if="data.latest_run.binary_sizes.length">
+      <h3>Binary Sizes</h3>
+      <table>
+        <thead><tr><th>Binary</th><th>Size (bytes)</th><th>Size</th></tr></thead>
+        <tbody>
+          <tr v-for="b in data.latest_run.binary_sizes" :key="b.name">
+            <td>{{ b.name }}</td>
+            <td>{{ b.bytes.toLocaleString() }}</td>
+            <td>{{ formatSize(b.bytes) }}</td>
+          </tr>
+        </tbody>
+      </table>
+    </template>
+  </template>
+</div>
+
+<script>
+const COLORS = [
+  '#4e79a7', '#f28e2b', '#e15759', '#76b7b2', '#59a14f',
+  '#edc948', '#b07aa1', '#ff9da7', '#9c755f', '#bab0ac',
+  '#86bcb6', '#8cd17d', '#b6992d', '#499894', '#d37295',
+];
+
+const { createApp, ref, onMounted } = Vue;
+
+createApp({
+  setup() {
+    const data = BUILD_DATA;
+    const totalBuildTime = ref(null);
+    const crateBuildTime = ref(null);
+    const totalBinarySize = ref(null);
+    const perBinarySize = ref(null);
+
+    function formatSize(bytes) {
+      return bytes > 1048576
+        ? (bytes / 1048576).toFixed(1) + ' MiB'
+        : (bytes / 1024).toFixed(1) + ' KiB';
+    }
+
+    onMounted(() => {
+      new Chart(totalBuildTime.value, {
+        type: 'line',
+        data: {
+          labels: data.dates,
+          datasets: [{
+            label: 'Total Wall Time (s)',
+            data: data.total_build_times,
+            borderColor: '#4e79a7',
+            backgroundColor: '#4e79a733',
+            fill: true, tension: 0.3,
+          }]
+        },
+        options: {
+          responsive: true,
+          plugins: { title: { display: true, text: 'Total Release Build Time' } },
+          scales: { y: { beginAtZero: true, title: { display: true, text: 'Seconds' } } }
+        }
+      });
+
+      new Chart(crateBuildTime.value, {
+        type: 'line',
+        data: {
+          labels: data.dates,
+          datasets: data.crate_datasets.map((ds, i) => ({
+            ...ds,
+            borderColor: COLORS[i % COLORS.length],
+            backgroundColor: COLORS[i % COLORS.length] + '33',
+            tension: 0.3, spanGaps: true,
+          }))
+        },
+        options: {
+          responsive: true,
+          plugins: { title: { display: true, text: 'Build Time by Crate (Top 15)' } },
+          scales: { y: { beginAtZero: true, title: { display: true, text: 'Seconds' } } }
+        }
+      });
+
+      new Chart(totalBinarySize.value, {
+        type: 'line',
+        data: {
+          labels: data.dates,
+          datasets: [{
+            label: 'Total Binary Size (MiB)',
+            data: data.total_binary_sizes_mib,
+            borderColor: '#e15759',
+            backgroundColor: '#e1575933',
+            fill: true, tension: 0.3,
+          }]
+        },
+        options: {
+          responsive: true,
+          plugins: { title: { display: true, text: 'Total Binary Size' } },
+          scales: { y: { beginAtZero: true, title: { display: true, text: 'MiB' } } }
+        }
+      });
+
+      new Chart(perBinarySize.value, {
+        type: 'line',
+        data: {
+          labels: data.dates,
+          datasets: data.binary_datasets.map((ds, i) => ({
+            ...ds,
+            borderColor: COLORS[i % COLORS.length],
+            backgroundColor: COLORS[i % COLORS.length] + '33',
+            tension: 0.3, spanGaps: true,
+          }))
+        },
+        options: {
+          responsive: true,
+          plugins: { title: { display: true, text: 'Binary Size per Binary' } },
+          scales: { y: { beginAtZero: true, title: { display: true, text: 'MiB' } } }
+        }
+      });
+    });
+
+    return { data, formatSize, totalBuildTime, crateBuildTime, totalBinarySize, perBinarySize };
+  }
+}).mount('#app');
+</script>
+</body>
+</html>
+
diff --git a/.github/scripts/build-stats-report-data.py b/.github/scripts/build-stats-report-data.py
@@ -0,0 +1,165 @@
+"""Aggregate build-stats artifacts into a JS data file for the HTML report.
+
+Reads from:
+  collected/runs.tsv        — tab-separated: run_id, created_at, head_sha
+  collected/<run_id>/       — contains cargo-timing.html, build-stats-size.json, cargo-bloat.txt
+
+Usage: python build-stats-report-data.py <collected_dir> <output_dir>
+"""
+import json
+import re
+import sys
+from datetime import datetime, timezone
+from pathlib import Path
+
+
+def parse_cargo_timing(html_path: Path) -> dict:
+    """Parse build times from a cargo-timing.html file."""
+    if not html_path.exists():
+        return {}
+
+    html = html_path.read_text()
+
+    m = re.search(r"DURATION\s*=\s*(\d+(?:\.\d+)?)", html)
+    total_s = float(m.group(1)) if m else 0
+
+    m2 = re.search(r"Total time:</td><td>([^<]+)</td>", html)
+    total_display = m2.group(1).strip() if m2 else f"{total_s:.1f}s"
+
+    m = re.search(r"const UNIT_DATA\s*=\s*(\[.*?\]);", html, re.DOTALL)
+    if not m:
+        return {"total_wall_time_s": total_s, "total_time_display": total_display, "units": []}
+
+    units = json.loads(m.group(1))
+    units_sorted = sorted(units, key=lambda u: u.get("duration", 0), reverse=True)
+
+    return {
+        "total_wall_time_s": total_s,
+        "total_time_display": total_display,
+        "units": [{"name": u["name"], "version": u.get("version", ""), "duration": u.get("duration", 0)} for u in units_sorted],
+    }
+
+
+def main():
+    collected_dir = Path(sys.argv[1])
+    output_dir = Path(sys.argv[2])
+    output_dir.mkdir(parents=True, exist_ok=True)
+
+    runs_tsv = collected_dir / "runs.tsv"
+
+    # Parse runs.tsv and load per-run artifacts
+    runs = []
+    for line in runs_tsv.read_text().strip().splitlines():
+        parts = line.split("\t")
+        if len(parts) < 3:
+            raise ValueError(f"Malformed line in runs.tsv: {line!r}")
+        run_id, created_at, head_sha = parts[0], parts[1], parts[2]
+        run_dir = collected_dir / run_id
+
+        timing_path = run_dir / "target/cargo-timings/cargo-timing.html"
+        bs_path = run_dir / "build-stats-size.json"
+        cb_path = run_dir / "cargo-bloat.txt"
+        ll_path = run_dir / "cargo-llvm-lines.txt"
+
+        if not timing_path.exists():
+            raise FileNotFoundError(f"Missing cargo-timing.html for run {run_id} in {run_dir}")
+
+        runs.append({
+            "run_id": run_id,
+            "created_at": created_at,
+            "head_sha": head_sha,
+            "build_times": parse_cargo_timing(timing_path),
+            "binary_sizes": json.loads(bs_path.read_text()) if bs_path.exists() else [],
+            "cargo_bloat": cb_path.read_text() if cb_path.exists() else "",
+            "cargo_llvm_lines": ll_path.read_text() if ll_path.exists() else "",
+        })
+
+    runs.sort(key=lambda r: r["created_at"])
+
+    dates = []
+    total_build_times = []
+    crate_times: dict[str, list] = {}
+    total_binary_sizes = []
+    per_binary: dict[str, list] = {}
+
+    for run in runs:
+        dt_str = run.get("created_at", "")
+        dates.append(dt_str[:10] if dt_str else "?")
+
+        bt = run.get("build_times", {})
+        total_build_times.append(bt.get("total_wall_time_s", 0))
+
+        # Per-crate build times
+        units = bt.get("units", [])
+        seen = set()
+        for u in units:
+            name = u.get("name", "")
+            if name not in crate_times:
+                crate_times[name] = [None] * (len(dates) - 1)
+            crate_times[name].append(u.get("duration", 0))
+            seen.add(name)
+        for name in crate_times:
+            if name not in seen:
+                crate_times[name].append(None)
+
+        # Binary sizes
+        bs = run.get("binary_sizes", [])
+        total_binary_sizes.append(sum(b.get("bytes", 0) for b in bs))
+
+        seen_bins = set()
+        for b in bs:
+            bname = b.get("name", "")
+            if bname not in per_binary:
+                per_binary[bname] = [None] * (len(dates) - 1)
+            per_binary[bname].append(b.get("bytes", 0))
+            seen_bins.add(bname)
+        for bname in per_binary:
+            if bname not in seen_bins:
+                per_binary[bname].append(None)
+
+    # Top 15 crates by average duration
+    def avg(lst):
+        vals = [v for v in lst if v is not None]
+        return sum(vals) / len(vals) if vals else 0
+
+    top_crates = sorted(crate_times.keys(), key=lambda c: avg(crate_times[c]), reverse=True)[:15]
+
+    # Latest cargo bloat and llvm-lines
+    latest_bloat = next((r["cargo_bloat"] for r in reversed(runs) if r.get("cargo_bloat")), "")
+    latest_llvm_lines = next((r["cargo_llvm_lines"] for r in reversed(runs) if r.get("cargo_llvm_lines")), "")
+
+    # Latest run details
+    latest_run = None
+    if runs:
+        last = runs[-1]
+        bt = last.get("build_times", {})
+        latest_run = {
+            "created_at": last.get("created_at", "?"),
+            "head_sha": last.get("head_sha", "?")[:12],
+            "total_time_display": bt.get("total_time_display", "?"),
+            "units": bt.get("units", []),
+            "binary_sizes": last.get("binary_sizes", []),
+        }
+
+    build_data = {
+        "generated": datetime.now(timezone.utc).strftime("%Y-%m-%d %H:%M UTC"),
+        "dates": dates,
+        "total_build_times": total_build_times,
+        "total_binary_sizes_mib": [s / 1048576 for s in total_binary_sizes],
+        "crate_datasets": [{"label": name, "data": crate_times[name]} for name in top_crates],
+        "binary_datasets": [
+            {"label": name, "data": [b / 1048576 if b is not None else None for b in per_binary[name]]}
+            for name in sorted(per_binary.keys())
+        ],
+        "latest_cargo_bloat": latest_bloat,
+        "latest_cargo_llvm_lines": latest_llvm_lines,
+        "latest_run": latest_run,
+    }
+
+    js_path = output_dir / "build-stats-report.js"
+    js_path.write_text(f"const BUILD_DATA = {json.dumps(build_data, indent=2)};\n")
+    print(f"Generated {js_path} ({len(runs)} runs)")
+
+
+if __name__ == "__main__":
+    main()