From ec611f4deafca43e07424e569111340e1e4a4c71 Mon Sep 17 00:00:00 2001
From: "vladimir.buzalka" <vladimir.buzalka@buzalka.cz>
Date: Tue, 30 Dec 2025 10:55:03 +0100
Subject: [PATCH] z230

---
 TestduplicitSyncThing.py | 153 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 153 insertions(+)
 create mode 100644 TestduplicitSyncThing.py

diff --git a/TestduplicitSyncThing.py b/TestduplicitSyncThing.py
new file mode 100644
index 0000000..b4dc75a
--- /dev/null
+++ b/TestduplicitSyncThing.py
@@ -0,0 +1,153 @@
+import os
+import pymysql
+import sys
+
+# ==============================
+# ⚙️ NASTAVENÍ
+# ==============================
+LOCAL_ROOT = r"U:\#Syncthing"
+
+# DB Připojení
+DB_HOST = "192.168.1.76"
+DB_PORT = 3307  # Port pro tvou DB
+DB_USER = "root"
+DB_PASS = "Vlado9674+"  # Tvé heslo
+DB_NAME = "torrents"
+DB_TABLE = "file_md5_index"
+REMOTE_SHARE_KEY = "#ColdData"
+
+
+# ==============================
+
+def get_db_connection():
+    try:
+        return pymysql.connect(
+            host=DB_HOST,
+            port=DB_PORT,
+            user=DB_USER,
+            password=DB_PASS,
+            database=DB_NAME,
+            cursorclass=pymysql.cursors.DictCursor
+        )
+    except Exception as e:
+        print(f"❌ CHYBA: Nelze se připojit k MySQL na {DB_HOST}:{DB_PORT}.")
+        print(f"   Detail: {e}")
+        sys.exit(1)
+
+
+def normalize_path(path):
+    """
+    1. Převede Windows '\\' na Linux '/'
+    2. Převede vše na malá písmena (aby 'Foto.jpg' == 'foto.JPG')
+    """
+    return path.replace("\\", "/").lower()
+
+
+def get_local_files(root_path):
+    print(f"📂 Skenuji lokální disk: {root_path}")
+    local_files = set()
+
+    if not os.path.exists(root_path):
+        print(f"❌ Chyba: Cesta {root_path} neexistuje.")
+        return local_files
+
+    count = 0
+    for root, dirs, files in os.walk(root_path):
+        # Ignorovat systémové složky Syncthingu
+        if ".stfolder" in dirs: dirs.remove(".stfolder")
+        if ".stversions" in dirs: dirs.remove(".stversions")
+
+        for file in files:
+            full_path = os.path.join(root, file)
+
+            try:
+                rel_path = os.path.relpath(full_path, root_path)
+            except ValueError:
+                continue
+
+            # NORMALIZACE (Lomítka + Malá písmena)
+            clean_path = normalize_path(rel_path)
+
+            local_files.add(clean_path)
+            count += 1
+
+            if count % 1000 == 0:
+                print(f"   ... načteno {count} lokálních souborů ...", end="\r")
+
+    print(f"✅ Lokálně nalezeno: {len(local_files):,} souborů.")
+    return local_files
+
+
+def get_remote_files_from_db():
+    print(f"📡 Připojuji se k DB '{DB_NAME}' a stahuji seznam...")
+    conn = get_db_connection()
+    remote_files = set()
+
+    try:
+        with conn.cursor() as cursor:
+            sql = f"SELECT full_path FROM {DB_TABLE} WHERE full_path LIKE %s"
+            cursor.execute(sql, (f"%/{REMOTE_SHARE_KEY}/%",))
+
+            rows = cursor.fetchall()
+            print(f"   ... DB vrátila {len(rows):,} řádků. Zpracovávám...")
+
+            for row in rows:
+                full_path_db = row['full_path']
+
+                if f"/{REMOTE_SHARE_KEY}/" in full_path_db:
+                    parts = full_path_db.split(f"/{REMOTE_SHARE_KEY}/")
+                    if len(parts) > 1:
+                        rel_path = parts[1]
+
+                        # NORMALIZACE (Lomítka + Malá písmena)
+                        clean_path = normalize_path(rel_path)
+                        remote_files.add(clean_path)
+
+    finally:
+        conn.close()
+
+    print(f"✅ V DB nalezeno: {len(remote_files):,} unikátních cest.")
+    return remote_files
+
+
+def main():
+    print("🚀 HLEDÁNÍ DUPLIKÁTŮ (Case Insensitive)")
+    print("=======================================")
+
+    local_set = get_local_files(LOCAL_ROOT)
+    remote_set = get_remote_files_from_db()
+
+    print("\n🔍 Porovnávám (všechna písmena ignorují velikost)...")
+
+    # Průnik
+    duplicates = local_set.intersection(remote_set)
+    only_local = local_set - remote_set
+
+    print("\n📊 VÝSLEDKY")
+    print("=======================================")
+    print(f"🏠 Celkem lokálně:      {len(local_set):,}")
+    print(f"☁️  Celkem v DB:         {len(remote_set):,}")
+    print("---------------------------------------")
+    print(f"👯 SHODA (DUPLIKÁTY):   {len(duplicates):,}")
+    print(f"🆕 UNIKÁTNÍ (JEN DOMA): {len(only_local):,}")
+    print("=======================================\n")
+
+    if duplicates:
+        file_dup = "report_DUPLIKATY.txt"
+        with open(file_dup, "w", encoding="utf-8") as f:
+            for item in sorted(duplicates):
+                f.write(f"{item}\n")
+        print(f"💾 Seznam duplikátů (malá písmena) uložen do: {file_dup}")
+
+    if only_local:
+        file_new = "report_NOVE_SOUBORY.txt"
+        with open(file_new, "w", encoding="utf-8") as f:
+            for item in sorted(only_local):
+                f.write(f"{item}\n")
+        print(f"💾 Seznam nových souborů uložen do: {file_new}")
+
+    input("\nStiskni ENTER pro ukončení...")
+
+
+if __name__ == "__main__":
+    main()
\ No newline at end of file