horizon/tools/test_world_spawn_parity.py

#!/usr/bin/env python3
"""Compare a new world's spawn with vanilla's, seed by seed.

A new overworld's spawn comes from two steps (MinecraftServer.setInitialSpawn):

1. the origin: NoiseBasedChunkGenerator.getOrigin runs NoiseSpawnFinder over
   the noise settings' `spawn_target` points (a coarse then a fine radial
   search of the climate fields) and takes the chunk of the best column;
2. the spawn: an 11 by 11 chunk spiral around that chunk, the first chunk
   with a column PlayerSpawnFinder.getSpawnPosInChunk accepts (a full-topped
   block under open sky, no liquid on top) gives the world spawn.

Origins: VanillaWorldSpawnOracle calls NoiseSpawnFinder.findSpawnPosition and
getOrigin for every seed; the spawn_origin example asks Horizon's generator.
World spawns: the pinned vanilla dedicated server creates each world and
stops; Horizon's (`horizon-client --dedicated --nogui`) creates its own from
the same server.properties and stops; the `spawn` position of both
level.dat files is compared.

Vanilla is not deterministic everywhere: trees of neighbouring chunks
overlap on chunk corners, and which one is placed last depends on the order
the server's worker threads decorate the chunks. A spawn column on such a
corner can then stand on different leaves from one vanilla run to the next.
When the spawns differ, vanilla creates the world up to RETRIES more times;
Horizon's spawn passes if one of vanilla's runs gives it too, and the seed is
reported as varying in vanilla.

Seeds: the player report's seed (1891623539983467248: Horizon spawned at
-48 -48, vanilla at 152 145), the chunk generation harness's seeds and
--random seeds from a fixed generator.

usage: test_world_spawn_parity.py [--random N] [--jobs N] [--origins-only]
"""
import argparse
import json
import os
import random
import shutil
import subprocess
import sys
import time
from concurrent.futures import ThreadPoolExecutor
from pathlib import Path

try:
    from . import anvil_dump, lab, oracle_harness as oh, test_rust_parity as base
    from .test_dedicated_server_parity import Server, free_port
except ImportError:
    import anvil_dump
    import lab
    import oracle_harness as oh
    import test_rust_parity as base
    from test_dedicated_server_parity import Server, free_port

NAME = 'world-spawn-parity'
REPORT = 1891623539983467248
RETRIES = 3
SEEDS = [REPORT, 12345, -7046029254386353131, 987654321987, 4242, -1, 0x5DEECE66D, 20260607, 777, 0]
RESOURCES = lab.RESOURCES


def seeds(count):
    rng = random.Random(20261014)
    return SEEDS + [rng.randrange(-2**63, 2**63) for _ in range(count)]


def target_dir():
    metadata = json.loads(subprocess.check_output(['cargo', 'metadata', '--no-deps', '--format-version', '1'],
                                                  cwd=lab.ROOT, text=True))
    return Path(metadata['target_directory']) / 'release'


def origins(build, classes, dependencies, all_seeds):
    vanilla_out = build / 'origins-vanilla.txt'
    oh.run_logged(['java', '-cp', str(classes) + os.pathsep + dependencies, 'dev.horizonlab.parity.VanillaWorldSpawnOracle',
                   vanilla_out, *all_seeds], build, 'origins-vanilla')
    rust = subprocess.run([target_dir() / 'examples/spawn_origin', RESOURCES, *map(str, all_seeds)],
                          capture_output=True, text=True, check=True, cwd=lab.ROOT).stdout
    vanilla = vanilla_out.read_text()
    parse = lambda text: {int(l.split()[0]): tuple(map(int, l.split()[1:])) for l in text.splitlines() if l}
    return parse(vanilla), parse(rust)


def world_spawn(command, directory, seed, log_name):
    """Creates a world with `seed` in `directory`, stops, returns its level.dat spawn position."""
    if directory.exists():
        shutil.rmtree(directory)
    directory.mkdir(parents=True)
    props = {'online-mode': 'false', 'level-seed': str(seed), 'server-port': str(free_port()),
             'view-distance': '4', 'simulation-distance': '4', 'pause-when-empty-seconds': '-1'}
    (directory / 'server.properties').write_text(''.join(f'{k}={v}\n' for k, v in props.items()), encoding='iso-8859-1')
    (directory / 'eula.txt').write_text('eula=true\n')
    server = Server(command, directory, log_name)
    try:
        server.wait_for(r'Done \(', timeout=900)
    finally:
        server.stop()
    data = anvil_dump.read_dat(directory / 'world/level.dat')['Data']
    return tuple(data['spawn']['pos'][1])


def main():
    parser = argparse.ArgumentParser(description=__doc__)
    parser.add_argument('--random', type=int, default=50, help='random seeds besides the fixed ones')
    parser.add_argument('--jobs', type=int, default=max(1, min(4, int(os.environ.get('FLEET_LOAD_JOBS', '4')) // 4)),
                        help='worlds created at once (each side)')
    parser.add_argument('--origins-only', action='store_true', help='compare the spawn search origins only')
    parser.add_argument('--seed', type=int, action='append', help='only these seeds (repeatable)')
    args = parser.parse_args()
    lock, jars, libraries = base.verify_artifacts()
    build = lab.LOCAL / NAME
    build.mkdir(parents=True, exist_ok=True)
    dependencies = os.pathsep.join(map(str, [jars['server'], *libraries]))
    classes = build / 'classes'
    classes.mkdir(exist_ok=True)
    oh.run_logged(['javac', '--release', lock['minecraft']['java_major'], '-proc:none', '-nowarn', '-cp', dependencies,
                   '-d', classes, lab.ROOT / 'mods/horizon/src/test/java/dev/horizonlab/parity/VanillaWorldSpawnOracle.java'],
                  build, 'compile-java')
    oh.run_logged(['cargo', 'build', '--release', '-p', 'horizon-world', '--example', 'spawn_origin'], build, 'compile-probe')
    all_seeds = args.seed or seeds(args.random)
    started = time.time()

    vanilla, rust = origins(build, classes, dependencies, all_seeds)
    origin_diffs = [s for s in all_seeds if vanilla.get(s) != rust.get(s)]
    for s in origin_diffs:
        print(f'origin {s}: vanilla column {vanilla.get(s)} / horizon {rust.get(s)}')
    print(f'Origins: {len(all_seeds) - len(origin_diffs)}/{len(all_seeds)} seeds match '
          f'(column x z, chunk x z) in {time.time() - started:.0f}s', flush=True)
    if args.origins_only:
        sys.exit(1 if origin_diffs else 0)

    oh.run_logged(['cargo', 'build', '--release', '-p', 'horizon-client', '--bin', 'horizon-client'], build,
                  'compile-rust', timeout=3600)
    binary = target_dir() / ('horizon-client.exe' if os.name == 'nt' else 'horizon-client')
    vanilla_server = ['java', '-Xmx2g', '-cp', dependencies, 'net.minecraft.server.Main', 'nogui']
    horizon_server = [binary, '--resources', RESOURCES, '--dedicated', '--nogui']
    os.environ['HORIZON_NO_AUDIO'] = '1'

    def one(seed):
        t0 = time.time()
        a = world_spawn(vanilla_server, build / f'vanilla-{seed}', seed, f'vanilla-{seed}')
        b = world_spawn(horizon_server, build / f'horizon-{seed}', seed, f'horizon-{seed}')
        runs = [a]
        while b not in runs and len(runs) <= RETRIES:
            runs.append(world_spawn(vanilla_server, build / f'vanilla-{seed}-{len(runs)}', seed, f'vanilla-{seed}-{len(runs)}'))
        return seed, runs, b, time.time() - t0

    spawn_diffs, varying = [], []
    with ThreadPoolExecutor(args.jobs) as pool:
        for seed, runs, b, took in pool.map(one, all_seeds):
            ok = b in runs
            if not ok:
                spawn_diffs.append(seed)
            elif len(set(runs)) > 1:
                varying.append(seed)
            seen = ', '.join(' '.join(map(str, r)) for r in dict.fromkeys(runs))
            print(f'{"ok  " if ok else "DIFF"} {seed}: vanilla {seen} / horizon {b[0]} {b[1]} {b[2]} '
                  f'(origin chunk {vanilla[seed][2]} {vanilla[seed][3]}; {took:.0f}s)', flush=True)
            if ok:
                for path in build.glob(f'*-{seed}*'):
                    shutil.rmtree(path, ignore_errors=True)
    print(f'World spawns: {len(all_seeds) - len(spawn_diffs)}/{len(all_seeds)} seeds match in {time.time() - started:.0f}s'
          + (f'; vanilla itself gave several spawns for {len(varying)}: {varying}' if varying else ''))
    if origin_diffs or spawn_diffs:
        sys.exit(1)


if __name__ == '__main__':
    main()