DFiP_Budget_planing/perf/perfkit/aggregate.py

"""Агрегатор: python -m perfkit.aggregate <run_dir> -> perf/report-data/<run_id>/summary.{json,md}."""
from __future__ import annotations

import json
import math
import sys
from pathlib import Path

from . import REPORT_DATA_DIR
from .results import read_all

GROUP_KEYS = ("scenario", "target", "role", "scale", "cpu_rate")
# основная метрика сравнения (решение человека 2026-10-25): суммарное время main thread на действие - первыми строками
PRIMARY_METRICS = ("main_thread_ms_per_step", "main_thread_ms")


def _metric_order(m: str):
    return (PRIMARY_METRICS.index(m), "") if m in PRIMARY_METRICS else (len(PRIMARY_METRICS), m)


def _p95(sorted_vals):
    """Nearest-rank."""
    k = math.ceil(0.95 * len(sorted_vals))
    return sorted_vals[max(k, 1) - 1]


def _median(s):
    n = len(s)
    return s[n // 2] if n % 2 else (s[n // 2 - 1] + s[n // 2]) / 2


def _num(v):
    return isinstance(v, (int, float)) and not isinstance(v, bool)


def summarize(run_dir) -> dict:
    run_dir = Path(run_dir)
    groups: dict[tuple, dict] = {}
    for r in read_all(run_dir):
        key = tuple(r[k] for k in GROUP_KEYS)
        g = groups.setdefault(key, {"vals": {}, "n": 0, "incomplete_n": 0, "traced_n": 0})
        if r["metrics"].get("traced") == 1:
            # повтор под трассировкой (O-9): искажён накладными расходами trace, в статистику не входит
            g["traced_n"] += 1
            continue
        if r.get("incomplete"):
            g["incomplete_n"] += 1
            continue
        g["n"] += 1
        for m, v in r["metrics"].items():
            if m != "traced" and _num(v):
                g["vals"].setdefault(m, []).append(v)
    out = []
    for key in sorted(groups, key=lambda k: tuple(str(x) for x in k)):
        g = groups[key]
        metrics = {}
        for m, vals in sorted(g["vals"].items(), key=lambda kv: _metric_order(kv[0])):
            s = sorted(vals)
            metrics[m] = {"median": _median(s), "p95": _p95(s), "min": s[0], "max": s[-1], "n": len(s)}
        out.append({**dict(zip(GROUP_KEYS, key)), "n": g["n"], "incomplete_n": g["incomplete_n"], "traced_n": g["traced_n"], "metrics": metrics})
    return {"run_id": run_dir.name, "groups": out}


def _fmt(x):
    return f"{x:.1f}" if isinstance(x, float) and not x.is_integer() else str(int(x)) if isinstance(x, float) else str(x)


def render_md(summary: dict) -> str:
    lines = [f"# Summary: {summary['run_id']}", ""]
    for g in summary["groups"]:
        lines.append(f"## {g['scenario']} | target={g['target']} | role={g['role']} | scale={g['scale']} | cpu_rate={g['cpu_rate']}")
        lines.append("")
        lines.append(f"Замеров: {g['n']}, незавершённых: {g['incomplete_n']}" + (f", исключено трассировочных: {g['traced_n']}" if g.get("traced_n") else ""))
        lines.append("")
        lines.append("| metric | median | p95 | min | max | n |")
        lines.append("|---|---|---|---|---|---|")
        for m, s in g["metrics"].items():
            lines.append(f"| {m} | {_fmt(s['median'])} | {_fmt(s['p95'])} | {_fmt(s['min'])} | {_fmt(s['max'])} | {s['n']} |")
        lines.append("")
    return "\n".join(lines)


def main(argv=None) -> int:
    argv = sys.argv[1:] if argv is None else argv
    if len(argv) != 1:
        print("usage: python -m perfkit.aggregate <run_dir>", file=sys.stderr)
        return 2
    run_dir = Path(argv[0])
    s = summarize(run_dir)
    out = REPORT_DATA_DIR / s["run_id"]
    out.mkdir(parents=True, exist_ok=True)
    (out / "summary.json").write_text(json.dumps(s, ensure_ascii=False, indent=2), encoding="utf-8")
    (out / "summary.md").write_text(render_md(s), encoding="utf-8")
    print(out)
    return 0


if __name__ == "__main__":
    raise SystemExit(main())