Files
walkfiles/WalkFilesOnBackupHDD/51 testthoseneedchangewhetherok.py
2026-02-03 13:26:11 +01:00

147 lines
3.3 KiB
Python

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
import os
import hashlib
import posixpath
import unicodedata
import pymysql
import time
# =========================
# CONFIG
# =========================
HOST_TO_CHECK = "Tower"
WINDOWS_UNC_BASE = r"\\tower"
DB_CONFIG = {
"host": "192.168.1.50",
"port": 3306,
"user": "root",
"password": "Vlado9674+",
"database": "torrents",
"charset": "utf8mb4",
}
PRINT_FIRST_CHANGES = 20
LOG_EVERY = 5000
# =========================
# CANONICAL
# =========================
def canonical_path(path_str):
path_str = path_str.replace("\\", "/")
path_str = posixpath.normpath(path_str)
path_str = unicodedata.normalize("NFC", path_str)
return path_str
def md5_bytes(path_str):
return hashlib.md5(path_str.encode("utf-8")).digest()
# =========================
# PATH MAP
# =========================
def linux_to_windows_unc(linux_path):
rel = linux_path[len("/mnt/user/"):]
return os.path.join(WINDOWS_UNC_BASE, *rel.split("/"))
# =========================
# MAIN
# =========================
def main():
print("=" * 70)
print("🔍 Tower Canonical Path SMB Verification")
print(f"Host: {HOST_TO_CHECK}")
print(f"UNC Base: {WINDOWS_UNC_BASE}")
print("=" * 70)
db = pymysql.connect(**DB_CONFIG)
cur = db.cursor(pymysql.cursors.SSCursor)
cur.execute("""
SELECT id, full_path, path_hash
FROM file_md5_index
WHERE host_name = %s
""", (HOST_TO_CHECK,))
total = 0
needs_change = 0
exists_ok = 0
missing = 0
printed_changes = 0
start = time.time()
for rec_id, full_path, stored_hash in cur:
total += 1
new_path = canonical_path(full_path)
new_hash = md5_bytes(new_path)
# Already canonical
if new_path == full_path and new_hash == stored_hash:
continue
needs_change += 1
win_path = linux_to_windows_unc(new_path)
exists = os.path.exists(win_path)
if exists:
exists_ok += 1
else:
missing += 1
# ---- Print first examples ----
if printed_changes < PRINT_FIRST_CHANGES:
print("\n🔧 CHANGE DETECTED")
print(f"ID : {rec_id}")
print(f"DB PATH : {full_path}")
print(f"NEW PATH : {new_path}")
print(f"WIN PATH : {win_path}")
print(f"Exists : {exists}")
printed_changes += 1
# ---- Progress ----
if total % LOG_EVERY == 0:
elapsed = time.time() - start
rate = total / elapsed if elapsed else 0
print(
f"📊 Checked {total:,} rows | "
f"Needs change {needs_change:,} | "
f"Exists {exists_ok:,} | "
f"Missing {missing:,} | "
f"{rate:,.0f} rows/sec"
)
# =========================
# SUMMARY
# =========================
elapsed = time.time() - start
print("\n" + "=" * 70)
print("✅ FINAL SUMMARY")
print("=" * 70)
print(f"Total scanned : {total:,}")
print(f"Needs change : {needs_change:,}")
print(f"Exists on Tower : {exists_ok:,}")
print(f"Missing on Tower : {missing:,}")
print(f"Runtime : {elapsed:.1f}s")
print("=" * 70)
cur.close()
db.close()
if __name__ == "__main__":
main()