torrents/Seedbox/81 DeleteWhatWeHaveUltraCC.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Projde /mnt/user/torrents/ultracc, pro každý soubor spočítá blake3
a porovná s tabulkou file_md5_index. Pokud je hash nalezen → soubor smaže.
Po smazání souborů odstraní prázdné adresáře.
"""

import os
import sys
import blake3
import pymysql
import paramiko
from pathlib import Path

# ============================================================
# CONFIG
# ============================================================

SCAN_DIR = "//tower/torrents/ultracc"

SSH_CONFIG = {
    "hostname": "192.168.1.76",
    "port":     22,
    "username": "root",
    "password": "7309208104",
}

ULTRACC_DIRS = [
    "/mnt/user/Torrents/UltraCC",
    "/mnt/user/Torrents/UltraCC1",
    "/mnt/user/Torrents/UltraCC2",
]

DB_CONFIG = {
    "host":     "192.168.1.76",
    "port":     3306,
    "user":     "root",
    "password": "Vlado9674+",
    "database": "torrents",
    "charset":  "utf8mb4",
}

CHUNK_SIZE  = 8 * 1024 * 1024  # 8 MB
DRY_RUN     = True             # True = pouze vypíše, nesmaže

# ============================================================
# HELPERS
# ============================================================

def compute_blake3(path: Path) -> bytes:
    """Vrátí blake3 digest jako 32 raw bytes."""
    h = blake3.blake3()
    with open(path, "rb") as f:
        while True:
            chunk = f.read(CHUNK_SIZE)
            if not chunk:
                break
            h.update(chunk)
    return h.digest()


def hash_in_db(cursor, digest: bytes):
    """Vrátí (host_name, full_path) prvního záznamu s daným hashem, nebo None."""
    cursor.execute(
        "SELECT host_name, full_path FROM file_md5_index WHERE blake3 = %s AND host_name = 'tower1' AND full_path LIKE '/mnt/user/#ColdData/Porno/%%' LIMIT 1",
        (digest,)
    )
    return cursor.fetchone()  # None nebo (host_name, full_path)


def remove_empty_dirs(root: str) -> int:
    """Rekurzivně smaže prázdné adresáře pod root. Vrátí počet smazaných."""
    removed = 0
    for dirpath, dirnames, filenames in os.walk(root, topdown=False):
        if dirpath == root:
            continue
        try:
            os.rmdir(dirpath)
            print(f"  [rmdir] {dirpath}")
            removed += 1
        except OSError:
            pass
    return removed


# ============================================================
# MAIN
# ============================================================

def set_ultracc_permissions():
    """Přes SSH nastaví na Tower chown nobody:users + chmod 777 pro všechny UltraCC adresáře."""
    print("Nastavuji práva na Tower (UltraCC*)...")
    ssh = paramiko.SSHClient()
    ssh.set_missing_host_key_policy(paramiko.AutoAddPolicy())
    ssh.connect(**SSH_CONFIG)
    for d in ULTRACC_DIRS:
        _, out, err = ssh.exec_command(
            'chown -R nobody:users "%s" && chmod -R 777 "%s" && echo OK' % (d, d)
        )
        result = out.read().decode().strip()
        error  = err.read().decode().strip()
        if result == "OK":
            print(f"  [OK]    {d}")
        else:
            print(f"  [CHYBA] {d}: {error}")
    ssh.close()
    print()


def main():
    dry_run = DRY_RUN

    set_ultracc_permissions()

    if dry_run:
        print("=== DRY RUN — nic se nesmaže ===\n")

    conn = pymysql.connect(**DB_CONFIG)
    cursor = conn.cursor()

    scan_root = Path(SCAN_DIR)
    if not scan_root.exists():
        print(f"CHYBA: Adresář neexistuje: {SCAN_DIR}")
        sys.exit(1)

    files_checked = 0
    files_deleted = 0
    files_kept    = 0
    bytes_deleted = 0

    for file_path in scan_root.rglob("*"):
        if not file_path.is_file():
            continue

        files_checked += 1
        size = file_path.stat().st_size

        try:
            digest = compute_blake3(file_path)
        except OSError as e:
            print(f"  [CHYBA čtení] {file_path}: {e}")
            continue

        db_match = hash_in_db(cursor, digest)
        if db_match:
            db_host, db_path = db_match
            print(f"  [SMAZAT]  {file_path}  ({size:,} B)")
            print(f"            ↳ originál v DB: [{db_host}] {db_path}")
            if not dry_run:
                try:
                    file_path.unlink()
                    files_deleted += 1
                    bytes_deleted += size
                except OSError as e:
                    print(f"  [CHYBA smazání] {file_path}: {e}")
            else:
                files_deleted += 1
                bytes_deleted += size
        else:
            print(f"  [zachovat] {file_path}  ({size:,} B)")
            files_kept += 1

    cursor.close()
    conn.close()

    print()
    print(f"Zkontrolováno:  {files_checked} souborů")
    print(f"Ke smazání:     {files_deleted} souborů  ({bytes_deleted / 1024**3:.2f} GB)")
    print(f"Zachováno:      {files_kept} souborů")

    if not dry_run and files_deleted > 0:
        print("\nOdstraňuji prázdné adresáře...")
        removed = remove_empty_dirs(SCAN_DIR)
        print(f"Odstraněno prázdných adresářů: {removed}")

    if dry_run:
        print("\n(Dry run — žádné změny nebyly provedeny)")


if __name__ == "__main__":
    main()