git

Seedbox/80 DeleteWhatWeHave.py

@@ -0,0 +1,142 @@
+#!/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
+from pathlib import Path
+
+# ============================================================
+# CONFIG
+# ============================================================
+
+SCAN_DIR = "//tower/torrents/ultracc"
+
+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) -> bool:
+    cursor.execute(
+        "SELECT 1 FROM file_md5_index WHERE blake3 = %s LIMIT 1",
+        (digest,)
+    )
+    return cursor.fetchone() is not None
+
+
+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 main():
+    dry_run = DRY_RUN
+
+    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
+
+        if hash_in_db(cursor, digest):
+            print(f"  [SMAZAT]  {file_path}  ({size:,} B)")
+            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()