VoxelForge/Standalone/Research/WorldAssemblerM3/Linux87/RunL87Timeline.py

#!/usr/bin/env python3
"""Interleaved L87 passage A/B readiness and worker timeline matrix."""
from __future__ import annotations

import json
import argparse
import re
import statistics
import subprocess
import time
from pathlib import Path

ROOT = Path(__file__).resolve().parents[4]
OUT = ROOT / "Saved/Linux/l87/timeline/switch-fixed"
DATA = ROOT / "Standalone/data/owner_world.json"
BINARY = ROOT / "Saved/Linux/l87/bin/VoxelForgeCoreHeadless-clang19"
SOURCE_STAMP = "ac3d8ea64e9ac4bd5526c61990070fb504528ff56dae52d33c1937327374032b"
STRATA = {
    "FI": {"id": 1, "camera": (25.8369140625, 3.19921875, -737.0),
           "near": 1216, "whole": 2249},
    "TN": {"id": 2, "camera": (19.166015625, -14.2001953125, -753.3740234375),
           "near": 1344, "whole": 2233},
}
WORKERS = (8, 12)
MODES = ("direct", "passage_off")
PIPELINE = re.compile(r"^\[StreamPipeline\] (.*)$", re.M)
FIELDS = re.compile(r"([a-zA-Z0-9_]+)=([^ ]+)")
NEAR = re.compile(
    r"^\[WorldAssemblerNearLODReady\] lods=0-2 tiles=(\d+)/(\d+) cold_ready_s=([0-9.]+)$",
    re.M,
)
PUBLISHED = re.compile(
    r"^\[WorldAssemblerNearLODPublished\] lods=0-2 tiles=(\d+)/(\d+) cold_ready_s=([0-9.]+)$",
    re.M,
)
PIPELINE_READY = re.compile(
    r"^\[WorldAssemblerNearLODPipelineReady\] lods=0-2 tiles=(\d+)/(\d+) cold_ready_s=([0-9.]+)$",
    re.M,
)
SHAPE = re.compile(
    r"^\[FarShapeStage\] parent_shapes=([0-9.]+) ms spatial_bins=([0-9.]+) ms total=([0-9.]+) ms",
    re.M,
)
CONFIG = re.compile(r"^\[HeadlessConfig\] (.*)$", re.M)
STAMP = re.compile(r"^\[VoxelForgeSourceStamp\] sha256=(\S+)$", re.M)
VOLUME = re.compile(r"^\[PipelineVolume\] tiles=(\d+) lod=", re.M)
TIMELINE_EVENTS = re.compile(r"^\[CpuTileTimeline\] event=(\S+)", re.M)


def command(stratum: str, workers: int, mode: str, timeline: bool) -> list[str]:
    camera = ",".join(f"{v:.12f}" for v in STRATA[stratum]["camera"]) + ",0,0"
    child = [
        str(BINARY), str(DATA), "--world-assembler-prototype", "--stream-benchmark",
        "--stream-pipeline", f"--workers={workers}", "--clip-radius=3",
        "--max-clip-level=8", "--max-level=8", f"--camera={camera}",
        "--profile-lod-work", "--stream-scheduler-policy=3",
    ]
    if timeline:
        child.append("--cold-worker-timeline")
    env = ["VOXELFORGE_L81_PASSAGE_BOUND_PADDING_VOXELS=0"]
    if mode == "passage_off":
        env.append("VOXELFORGE_L86_DISABLE_PASSAGE_COMPOSITION=1")
    return [
        "taskset", "-c", "0-11", "prlimit", "--as=6442450944", "--", "env",
        *env, "timeout", "1740s", *child,
    ]


def run(stratum: str, workers: int, mode: str, phase: str,
        replicate: int, timeline: bool) -> dict:
    OUT.mkdir(parents=True, exist_ok=True)
    rep = f"-r{replicate}" if phase == "measured" else ""
    label = f"{stratum.lower()}-w{workers}-{mode}-{phase}{rep}"
    log = OUT / f"{label}.log"
    started = time.monotonic()
    with log.open("w", encoding="utf-8") as output:
        result = subprocess.run(command(stratum, workers, mode, timeline), cwd=ROOT,
                                stdout=output, stderr=subprocess.STDOUT,
                                check=False, timeout=1740)
    content = log.read_text(encoding="utf-8", errors="replace")
    near, published, pipeline_ready = (
        NEAR.search(content), PUBLISHED.search(content), PIPELINE_READY.search(content)
    )
    pipeline_match, far = PIPELINE.search(content), SHAPE.search(content)
    stamp_match = STAMP.search(content)
    config_match, volume_match = CONFIG.search(content), VOLUME.search(content)
    if result.returncode or not all((near, published, pipeline_ready, pipeline_match, far,
                                    config_match, volume_match, stamp_match)):
        raise RuntimeError(f"run failed or required timeline is missing: {log}; exit={result.returncode}")
    if stamp_match.group(1) != SOURCE_STAMP:
        raise RuntimeError(f"source stamp mismatch in {log}: {stamp_match.group(1)}")
    expected = STRATA[stratum]
    config_fields = dict(FIELDS.findall(config_match.group(1)))
    if int(config_fields["workers"]) != workers:
        raise RuntimeError(f"worker mismatch in {log}: {config_fields['workers']}")
    if int(volume_match.group(1)) != expected["whole"]:
        raise RuntimeError(f"wrong camera/strate window in {log}: {volume_match.group(1)} tiles")
    if ("[L86PassageCeiling]" in content) != (mode == "passage_off"):
        raise RuntimeError(f"passage switch state mismatch in {log}")
    for match in (near, published, pipeline_ready):
        if (int(match.group(1)), int(match.group(2))) != (expected["near"], expected["near"]):
            raise RuntimeError(f"incomplete near set in {log}: {match.group(1)}/{match.group(2)}")
    fields = dict(FIELDS.findall(pipeline_match.group(1)))
    whole_done, whole_total = (int(v) for v in fields["whole_tiles"].split("/"))
    if (whole_done, whole_total) != (expected["whole"], expected["whole"]):
        raise RuntimeError(f"incomplete whole view in {log}: {whole_done}/{whole_total}")
    cpu_events = TIMELINE_EVENTS.findall(content)
    if timeline and not cpu_events:
        raise RuntimeError(f"worker timeline missing in {log}")
    row = {
        "stratum": stratum, "stratum_id": expected["id"], "workers": workers,
        "mode": mode, "phase": phase, "replicate": replicate,
        "timeline": timeline, "headless_recipe_label": config_fields.get("recipe"),
        "cpu_tile_timeline_events": len(cpu_events),
        "exit_code": result.returncode, "wall_s": round(time.monotonic() - started, 3),
        "source_stamp": stamp_match.group(1),
        "lod02_ready_s": float(near.group(3)),
        "lod02_published_s": float(published.group(3)),
        "lod02_pipeline_ready_s": float(pipeline_ready.group(3)),
        "whole_view_ready_s": float(fields["whole_view_ready"].removesuffix("s")),
        "shape_s": float(fields["shape"].removesuffix("s")),
        "far_shape_s": float(far.group(3)) / 1000.0,
        "far_parent_ms": float(far.group(1)), "far_bins_ms": float(far.group(2)),
        "near_tiles": expected["near"], "whole_tiles": expected["whole"],
        "log": str(log.relative_to(ROOT)),
    }
    print("[L87TimelineRun] " + json.dumps(row, sort_keys=True), flush=True)
    return row


def main() -> int:
    global OUT
    parser = argparse.ArgumentParser()
    parser.add_argument("--out", type=Path, default=Path("Saved/Linux/l87/timeline/final-phase"))
    args = parser.parse_args()
    OUT = args.out if args.out.is_absolute() else ROOT / args.out
    if not BINARY.is_file() or not DATA.is_file():
        raise SystemExit(f"missing input: {BINARY} or {DATA}")
    # L80 warmup: once per camera window and worker count. Rotate both mode and
    # worker order across three paired measured replicates.
    for stratum in STRATA:
        for workers in WORKERS:
            for mode in MODES:
                run(stratum, workers, mode, "warmup", 0, False)

    measured = []
    orders = (
        ((8, "direct"), (8, "passage_off"), (12, "passage_off"), (12, "direct")),
        ((12, "direct"), (12, "passage_off"), (8, "passage_off"), (8, "direct")),
        ((8, "passage_off"), (8, "direct"), (12, "direct"), (12, "passage_off")),
    )
    for stratum in STRATA:
        timeline_pending = {(workers, mode) for workers in WORKERS for mode in MODES}
        for replicate, order in enumerate(orders, 1):
            for workers, mode in order:
                timeline = (workers, mode) in timeline_pending
                if timeline:
                    timeline_pending.remove((workers, mode))
                measured.append(run(stratum, workers, mode, "measured", replicate, timeline))
                (OUT / "runs-partial.json").write_text(
                    json.dumps(measured, indent=2) + "\n", encoding="utf-8")

    summary = []
    for stratum in STRATA:
        for workers in WORKERS:
            for mode in MODES:
                rows = [row for row in measured if row["stratum"] == stratum
                        and row["workers"] == workers and row["mode"] == mode]
                if len(rows) != 3:
                    raise RuntimeError(f"expected 3 measured replicates for {stratum} W{workers} {mode}")
                item = {"stratum": stratum, "workers": workers, "mode": mode,
                        "measured_runs": 3}
                for metric in ("lod02_ready_s", "lod02_published_s",
                               "lod02_pipeline_ready_s", "whole_view_ready_s",
                               "shape_s", "far_shape_s"):
                    values = [row[metric] for row in rows]
                    item[metric] = values
                    item[f"{metric}_median"] = statistics.median(values)
                summary.append(item)
    (OUT / "summary.json").write_text(
        json.dumps({"runs": measured, "summary": summary}, indent=2) + "\n",
        encoding="utf-8")
    return 0


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