Files
fotkyBuzalkovi/00 PictureCollector/collect_pictures_windows.py
T
administrator db84c3e501 notebookVb
2026-05-25 06:39:49 +02:00

307 lines
9.9 KiB
Python

"""
collect_pictures_windows.py
---------------------------
Windows verze zálohovacího skriptu.
Prochází všechny lokální disky (ne síťové, ne CD), hledá JPG/JPEG,
kopíruje (deduplikace BLAKE3) na Tower1 přes UNC cestu a zapisuje do PostgreSQL.
Cesta zálohy: \\Tower1\ZalohaVsechObrazku\{hostname}\{písmeno_disku}\{cesta}
Příklad: \\Tower1\ZalohaVsechObrazku\JMENO-PC\D\Foto\2023\img.jpg
Bezpečné pro opakované spuštění (pokračuje tam kde skončilo).
Bezpečné pro souběžný běh na více strojích (ON CONFLICT v SQL).
"""
import os
import sys
import shutil
import socket
import string
import ctypes
import logging
from pathlib import Path
from datetime import datetime
import blake3
import psycopg2
from psycopg2.extras import execute_values
# ── Konfigurace ───────────────────────────────────────────────────────────────
DB_CONFIG = {
"host": "192.168.1.76",
"port": 5432,
"user": "vladimir.buzalka",
"password": "Vlado7309208104++",
"database": "fotky_buzalkovi",
}
ZALOHA_DIR = Path(r"\\Tower1\ZalohaVsechObrazku")
JPEG_EXTENSIONS = {".jpg", ".jpeg"}
EXCLUDED_DIR_NAMES_LOWER = {
"zalohavsechobrazku",
"zálohavsechobrázku",
"zálohavsechobrazku",
"$recycle.bin",
"system volume information",
"recovery",
}
EXCLUDED_FULL_PATHS = {
Path(os.environ.get("WINDIR", r"C:\Windows")),
}
BATCH_SIZE = 500
LOG_FILE = Path(__file__).parent / "collect_pictures_windows.log"
# ── Logging ───────────────────────────────────────────────────────────────────
logging.basicConfig(
level=logging.INFO,
format="%(asctime)s %(levelname)-7s %(message)s",
datefmt="%Y-%m-%d %H:%M:%S",
handlers=[
logging.StreamHandler(sys.stdout),
logging.FileHandler(LOG_FILE, encoding="utf-8"),
],
)
log = logging.getLogger(__name__)
# ── SQL ───────────────────────────────────────────────────────────────────────
SQL_INSERT_ZALOHA = """
INSERT INTO zaloha_obrazku (blake3_hash, cesta_zalohy, nazev_souboru, velikost)
VALUES (%s, %s, %s, %s)
ON CONFLICT (blake3_hash) DO NOTHING
RETURNING id
"""
SQL_INSERT_ZDROJ = """
INSERT INTO zdrojove_soubory (hostname, cesta_zdroje, nazev_souboru, velikost, blake3_hash, zaloha_id)
VALUES (%s, %s, %s, %s, %s, %s)
ON CONFLICT (hostname, cesta_zdroje) DO NOTHING
"""
SQL_GET_ZALOHA_ID = "SELECT id FROM zaloha_obrazku WHERE blake3_hash = %s"
# ── Pomocné funkce ────────────────────────────────────────────────────────────
def get_local_drives() -> list[Path]:
DRIVE_FIXED = 3
drives = []
bitmask = ctypes.windll.kernel32.GetLogicalDrives()
for letter in string.ascii_uppercase:
if bitmask & 1:
drive = Path(f"{letter}:\\")
drive_type = ctypes.windll.kernel32.GetDriveTypeW(str(drive))
if drive_type == DRIVE_FIXED:
drives.append(drive)
bitmask >>= 1
return drives
def is_excluded_path(path: Path) -> bool:
for excl in EXCLUDED_FULL_PATHS:
try:
path.relative_to(excl)
return True
except ValueError:
pass
return False
def compute_blake3(path: Path, chunk: int = 1 << 20) -> str:
h = blake3.blake3()
with open(path, "rb") as f:
while data := f.read(chunk):
h.update(data)
return h.hexdigest()
def dest_path_for(source: Path, hostname: str) -> Path:
drive_letter = source.drive.rstrip(":")
relative = source.relative_to(source.drive + "\\")
return ZALOHA_DIR / hostname / drive_letter / relative
def copy_to_backup(source: Path, dest: Path) -> None:
dest.parent.mkdir(parents=True, exist_ok=True)
if dest.exists():
return
shutil.copy2(source, dest)
def iter_jpeg_files(drives: list[Path]):
for drive in drives:
log.info(f"Skenuji disk: {drive}")
for root, dirs, files in os.walk(drive, followlinks=False):
root_path = Path(root)
if is_excluded_path(root_path):
dirs.clear()
continue
dirs[:] = [
d for d in dirs
if d.lower() not in EXCLUDED_DIR_NAMES_LOWER
and not is_excluded_path(root_path / d)
]
for fname in files:
if Path(fname).suffix.lower() in JPEG_EXTENSIONS:
yield root_path / fname
def load_known_sources(conn, hostname: str) -> set[str]:
with conn.cursor("src_cursor") as cur:
cur.itersize = 10000
cur.execute("SELECT cesta_zdroje FROM zdrojove_soubory WHERE hostname = %s", (hostname,))
return {row[0] for row in cur}
def load_known_hashes(conn) -> dict[str, int]:
with conn.cursor("hash_cursor") as cur:
cur.itersize = 10000
cur.execute("SELECT blake3_hash, id FROM zaloha_obrazku")
return {row[0]: row[1] for row in cur}
# ── Hlavní logika ─────────────────────────────────────────────────────────────
def process(conn, hostname, drives):
log.info(f"Hostname zdroje: {hostname}")
log.info("Načítám známé zdroje z DB...")
known_sources = load_known_sources(conn, hostname)
log.info(f"Známých zdrojů v DB: {len(known_sources)}")
log.info("Načítám známé hashe z DB...")
known_hashes = load_known_hashes(conn)
log.info(f"Známých hashů v DB: {len(known_hashes)}")
stats = {"nalezeno": 0, "kopirovano": 0, "duplicit": 0, "chyb": 0, "preskoceno": 0}
pending_zdroje = []
def flush_zdroje():
if not pending_zdroje:
return
cur = conn.cursor()
execute_values(
cur,
"""INSERT INTO zdrojove_soubory
(hostname, cesta_zdroje, nazev_souboru, velikost, blake3_hash, zaloha_id)
VALUES %s
ON CONFLICT (hostname, cesta_zdroje) DO NOTHING""",
pending_zdroje,
)
conn.commit()
cur.close()
pending_zdroje.clear()
for source in iter_jpeg_files(drives):
stats["nalezeno"] += 1
src_str = str(source)
if src_str in known_sources:
stats["preskoceno"] += 1
if stats["preskoceno"] % 5000 == 0:
log.info(f"Přeskočeno (již v DB): {stats['preskoceno']}")
continue
try:
velikost = source.stat().st_size
hash_val = compute_blake3(source)
except (OSError, PermissionError) as e:
log.warning(f"CHYBA čtení: {source}{e}")
stats["chyb"] += 1
continue
zaloha_id = known_hashes.get(hash_val)
if zaloha_id is not None:
stats["duplicit"] += 1
else:
dest = dest_path_for(source, hostname)
try:
copy_to_backup(source, dest)
except (OSError, shutil.Error) as e:
log.warning(f"CHYBA kopírování: {source}{e}")
stats["chyb"] += 1
continue
cur = conn.cursor()
cur.execute(SQL_INSERT_ZALOHA, (hash_val, str(dest), source.name, velikost))
row = cur.fetchone()
if row:
zaloha_id = row[0]
else:
cur.execute(SQL_GET_ZALOHA_ID, (hash_val,))
zaloha_id = cur.fetchone()[0]
cur.close()
conn.commit()
known_hashes[hash_val] = zaloha_id
stats["kopirovano"] += 1
log.info(f"ZKOPÍROVÁNO [{stats['kopirovano']:>6}] {source}")
pending_zdroje.append((hostname, src_str, source.name, velikost, hash_val, zaloha_id))
known_sources.add(src_str)
if len(pending_zdroje) >= BATCH_SIZE:
flush_zdroje()
if stats["nalezeno"] % 5000 == 0:
log.info(
f"Průběh: nalezeno={stats['nalezeno']} "
f"nových={stats['kopirovano']} duplikátů={stats['duplicit']} "
f"chyb={stats['chyb']} přeskočeno={stats['preskoceno']}"
)
flush_zdroje()
return stats
def main():
hostname = socket.gethostname()
drives = get_local_drives()
log.info("=" * 60)
log.info(f"Spuštění: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}")
log.info(f"Hostname: {hostname}")
log.info(f"Disky: {[str(d) for d in drives]}")
log.info(f"Záloha: {ZALOHA_DIR}\\{hostname}\\")
try:
conn = psycopg2.connect(**DB_CONFIG)
conn.autocommit = False
log.info("PostgreSQL: připojeno")
except Exception as e:
log.error(f"Nelze se připojit k DB: {e}")
sys.exit(1)
stats = {"nalezeno": 0, "kopirovano": 0, "duplicit": 0, "chyb": 0, "preskoceno": 0}
try:
stats = process(conn, hostname, drives)
except KeyboardInterrupt:
log.warning("Přerušeno uživatelem — dosavadní záznamy jsou uloženy.")
conn.rollback()
except Exception as e:
log.error(f"Neočekávaná chyba: {e}", exc_info=True)
conn.rollback()
finally:
conn.close()
log.info("-" * 60)
log.info(f"Nalezeno JPG/JPEG: {stats['nalezeno']}")
log.info(f"Zkopírováno nových: {stats['kopirovano']}")
log.info(f"Duplikátů (hash): {stats['duplicit']}")
log.info(f"Přeskočeno (v DB): {stats['preskoceno']}")
log.info(f"Chyb: {stats['chyb']}")
log.info("Hotovo.")
if __name__ == "__main__":
main()