devtools/bench_fast.py

"""Fast engine on a small fixed document (default: 4 AI-written subject paragraphs, ~460 words), all rounds allowed,
scored by the live detector ensemble. For comparing engine settings; speed on CPU is not meaningful.
    python devtools/bench_fast.py [--paras 4] [--tag name]       (FAST_* variables apply as on the Space)
Writes devtools/logs/bench_fast_<tag>.docx and prints the score after each round."""
import argparse
import os
import shutil
import sys
import time
from pathlib import Path

ROOT = Path(__file__).resolve().parent.parent
sys.path.insert(0, str(ROOT))
sys.path.insert(0, str(Path(__file__).resolve().parent))
import env  # noqa: E402
env.load()
os.environ.setdefault("LT_MODE", "off")
os.environ.setdefault("SKIP_MODEL_LOAD", "1")
sys.stdout.reconfigure(encoding="utf-8")

from docx import Document  # noqa: E402

PICK = ["civil engineering", "medicine", "management", "law", "computer science", "economics"]


def main():
    ap = argparse.ArgumentParser()
    ap.add_argument("--paras", type=int, default=4)
    ap.add_argument("--tag", default="run")
    a = ap.parse_args()
    from samples_domains import DOMAINS
    import fast as F
    d = Document()
    d.add_heading("Benchmark", 1)
    for k in PICK[:a.paras]:
        d.add_heading(k.title(), 2)
        d.add_paragraph(DOMAINS[k])
    src = ROOT / "devtools" / "logs" / "bench_fast_src.docx"
    d.save(str(src))
    t = time.time()
    r = F.humanize_docx(str(src), budget=100000)
    print(f"RESULT {a.tag}: {r['before']['pct']}% -> {r['after']['pct']}% | " + " | ".join(r["log"])
          + f" | {time.time() - t:.0f}s", flush=True)
    shutil.copy(r["path"], ROOT / "devtools" / "logs" / f"bench_fast_{a.tag}.docx")


if __name__ == "__main__":
    main()