NLP/lab1/scripts/package_submission.py

#!/usr/bin/env python3
"""Собирает итоговый ZIP-архив лабораторной работы.

Запуск из каталога lab1:

    python scripts/package_submission.py

Скрипт копирует только нужные файлы во временный каталог, формирует
MANIFEST.txt и CHECKSUMS.sha256, затем упаковывает всё в
dist/nlp_lab1_submission_YYYYMMDD.zip. Рядом кладётся .sha256 самого архива.
"""

from __future__ import annotations

import argparse
import hashlib
import os
import shutil
import zipfile
from datetime import date
from pathlib import Path

ROOT = Path(__file__).resolve().parents[1]
DIST = ROOT / "dist"
DEFAULT_NAME = f"nlp_lab1_submission_{date.today():%Y%m%d}"

EXCLUDE_PREFIXES = (
    "data/processed/",
    "features/",
    "results/",
    "nltk_data/",
    ".venv/",
    ".git/",
    "presentation/preview/",
    "report/build/",
    ".pytest_cache/",
    "dist/",
    "exports/nlp_lab1_results_",
    "exports/nlp_lab1_final_results_20261026_120000.tar.gz.sha256",
)

EXCLUDE_NAMES = {
    "stardew_stylometry.bak",
    "START_HERE.md",
    "ARCHIVE_README.md",
}

LATEX_AUX_SUFFIXES = {".aux", ".log", ".nav", ".out", ".snm", ".toc"}

KEY_FILES = [
    "lab1/report/REPORT.pdf",
    "lab1/report/REPORT.docx",
    "lab1/report/REPORT.md",
    "lab1/presentation/stardew_stylometry.pdf",
    "lab1/presentation/stardew_stylometry.pptx",
    "lab1/presentation/latex/stardew_stylometry.pdf",
    "lab1/presentation/latex/stardew_stylometry.tex",
    "lab1/FINAL_RESULTS.csv",
    "lab1/data/dialogues.csv",
    "lab1/exports/nlp_lab1_final_results_20261026_120000.tar.gz",
]


def sha256(path: Path) -> str:
    h = hashlib.sha256()
    with path.open("rb") as fh:
        for chunk in iter(lambda: fh.read(1024 * 1024), b""):
            h.update(chunk)
    return h.hexdigest()


def should_copy(rel: Path) -> bool:
    posix = rel.as_posix()
    if rel.name in EXCLUDE_NAMES:
        return False
    if "__pycache__" in rel.parts or rel.suffix == ".pyc":
        return False
    if rel.suffix in LATEX_AUX_SUFFIXES:
        return False
    if any(posix == prefix.rstrip("/") or posix.startswith(prefix)
           for prefix in EXCLUDE_PREFIXES):
        return False
    return True


def copy_tree(stage: Path) -> int:
    target_root = stage / "lab1"
    count = 0
    for path in sorted(ROOT.rglob("*")):
        if not path.is_file():
            continue
        rel = path.relative_to(ROOT)
        if not should_copy(rel):
            continue
        dst = target_root / rel
        dst.parent.mkdir(parents=True, exist_ok=True)
        shutil.copy2(path, dst)
        count += 1
    return count


def write_manifest(stage: Path) -> int:
    files = sorted(p for p in stage.rglob("*") if p.is_file())
    lines = [
        "Архив: " + stage.name,
        f"Файлов в манифесте: {len(files)} (сам MANIFEST.txt не включён)",
        "",
        f"{'SHA-256':64}  {'Размер':>12}  Путь",
        "-" * 100,
    ]
    total = 0
    for path in files:
        size = path.stat().st_size
        total += size
        rel = path.relative_to(stage).as_posix()
        lines.append(f"{sha256(path)}  {size:>12}  {rel}")
    lines.append("-" * 100)
    lines.append(f"Итого байт: {total}")
    (stage / "MANIFEST.txt").write_text("\n".join(lines) + "\n", encoding="utf-8")
    return len(files)


def write_checksums(stage: Path) -> None:
    lines = []
    for rel_name in KEY_FILES:
        path = stage / rel_name
        if path.exists():
            lines.append(f"{sha256(path)}  {rel_name}")
        else:
            print(f"ВНИМАНИЕ: нет ключевого файла {rel_name}")
    (stage / "CHECKSUMS.sha256").write_text("\n".join(lines) + "\n", encoding="utf-8")


def zip_stage(stage: Path, archive: Path) -> None:
    archive.parent.mkdir(parents=True, exist_ok=True)
    with zipfile.ZipFile(archive, "w", compression=zipfile.ZIP_DEFLATED,
                         compresslevel=6) as zf:
        for path in sorted(stage.rglob("*")):
            if not path.is_file():
                continue
            rel = path.relative_to(stage).as_posix()
            arcname = f"{stage.name}/{rel}"
            info = zipfile.ZipInfo.from_file(path, arcname)
            info.compress_type = zipfile.ZIP_DEFLATED
            with path.open("rb") as src, zf.open(info, "w") as dst:
                shutil.copyfileobj(src, dst)


def human(n: int) -> str:
    for unit in ("Б", "КБ", "МБ", "ГБ"):
        if n < 1024 or unit == "ГБ":
            return f"{n:.1f} {unit}" if unit != "Б" else f"{n} {unit}"
        n /= 1024
    return f"{n} Б"


def main() -> int:
    parser = argparse.ArgumentParser(description=__doc__)
    parser.add_argument("--name", default=DEFAULT_NAME, help="имя каталога и ZIP")
    parser.add_argument("--keep-stage", action="store_true",
                        help="не удалять промежуточный каталог dist/stage")
    args = parser.parse_args()

    name = args.name
    stage = DIST / "stage" / name
    if stage.exists():
        shutil.rmtree(stage)
    stage.mkdir(parents=True, exist_ok=True)

    print(f"1/4. Копирование файлов в {stage} ...")
    copied = copy_tree(stage)
    shutil.copy2(ROOT / "START_HERE.md", stage / "START_HERE.md")
    shutil.copy2(ROOT / "ARCHIVE_README.md", stage / "ARCHIVE_README.md")

    print("2/4. Формирование CHECKSUMS.sha256 ...")
    write_checksums(stage)

    print("3/4. Формирование MANIFEST.txt ...")
    total = write_manifest(stage)

    archive = DIST / f"{name}.zip"
    print(f"4/4. Упаковка {archive} ...")
    zip_stage(stage, archive)

    digest = sha256(archive)
    (DIST / f"{name}.zip.sha256").write_text(
        f"{digest}  {archive.name}\n", encoding="utf-8"
    )

    print()
    print(f"Файлов в lab1: {copied}")
    print(f"Всего файлов в архиве: {total + 1}")
    print(f"Размер ZIP: {human(archive.stat().st_size)}")
    print(f"SHA-256 ZIP: {digest}")
    print(f"Архив: {archive}")
    print(f"Контрольная сумма: {archive}.sha256")

    if not args.keep_stage:
        shutil.rmtree(stage, ignore_errors=True)
        try:
            (DIST / "stage").rmdir()
        except OSError:
            pass
    return 0


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