notebook

2026-06-08 07:20:37 +02:00
parent 0d3407e664
commit 70899149e4
14 changed files with 1162 additions and 14 deletions
@@ -0,0 +1,28 @@
+import sys, glob, os
+sys.stdout.reconfigure(encoding="utf-8")
+import extract_msg
+
+files = glob.glob(r"\\tower\JNJEMAILS\*.msg")
+f = files[0]
+fname = os.path.basename(f)
+print(f"Soubor: {fname}")
+print("(filename = posledních 20 znaků entry_id)")
+print()
+
+m = extract_msg.Message(f)
+print(f"messageId (Internet Message-ID): {m.messageId!r}")
+print()
+
+print("--- MAPI properties v souboru ---")
+try:
+    for pid in sorted(m.props.keys()):
+        prop = m.props[pid]
+        name = getattr(prop, "name", "")
+        print(f"  {pid}  {name}")
+except Exception as e:
+    print(f"  (props nedostupné: {e})")
+
+print()
+for attr in ("entryId", "entryID", "entry_id"):
+    print(f"  m.{attr} = {getattr(m, attr, '<není>')!r}")
+m.close()
@@ -0,0 +1,19 @@
+import sys, glob, os, sqlite3
+sys.stdout.reconfigure(encoding="utf-8")
+
+files = sorted(glob.glob(r"\\tower\JNJEMAILS\db\jnjemails_*.db"))
+db = files[-1]
+print(f"DB: {os.path.basename(db)}\n")
+
+conn = sqlite3.connect(db)
+for (tbl,) in conn.execute("SELECT name FROM sqlite_master WHERE type='table' ORDER BY name"):
+    print(f"=== {tbl} ===")
+    for cid, name, ctype, notnull, dflt, pk in conn.execute(f"PRAGMA table_info({tbl})"):
+        flags = []
+        if pk: flags.append("PK")
+        if notnull: flags.append("NOT NULL")
+        if dflt is not None: flags.append(f"default={dflt}")
+        print(f"  {name:14} {ctype:10} {' '.join(flags)}")
+    cnt = conn.execute(f"SELECT COUNT(*) FROM {tbl}").fetchone()[0]
+    print(f"  → {cnt} řádků\n")
+conn.close()
@@ -0,0 +1,120 @@
+"""
+backfill_entry_id.py | v1.0 | 2026-06-08
+Dohledá entry_id pro záznamy v jnjemails.db které ho nemají (69k starých emailů
+přenesených skriptem v1.1). Prochází celý Outlook MAPI strom a páruje emaily
+dle Internet Message-ID.
+
+Spouštět na JNJ PC s běžícím Outlookem.
+Bezpečné opakovat — přeskočí záznamy které už entry_id mají.
+"""
+
+import sqlite3
+import win32com.client
+from datetime import datetime
+
+DB_PATH = r"C:\Users\vbuzalka\SQLITE\jnjemails.db"
+PR_INTERNET_MESSAGE_ID = "http://schemas.microsoft.com/mapi/proptag/0x1035001E"
+
+
+def load_missing(conn) -> dict:
+    """Vrátí dict {message_id: db_id} pro záznamy bez entry_id."""
+    rows = conn.execute(
+        "SELECT id, message_id FROM messages WHERE entry_id IS NULL"
+    ).fetchall()
+    return {r[1]: r[0] for r in rows}
+
+
+def update_entry_id(conn, db_id: int, entry_id: str):
+    conn.execute(
+        "UPDATE messages SET entry_id = ? WHERE id = ?",
+        (entry_id, db_id)
+    )
+
+
+def scan_folder(conn, folder, lookup: dict, stats: dict, path: str = ""):
+    current = f"{path}/{folder.Name}"
+    try:
+        items = folder.Items
+        for item in items:
+            try:
+                if not item.MessageClass.upper().startswith("IPM.NOTE"):
+                    continue
+
+                stats["checked"] += 1
+
+                try:
+                    mid = item.PropertyAccessor.GetProperty(PR_INTERNET_MESSAGE_ID)
+                except Exception:
+                    mid = None
+                if not mid:
+                    mid = f"entryid:{item.EntryID}"
+
+                if mid in lookup:
+                    db_id = lookup.pop(mid)
+                    update_entry_id(conn, db_id, item.EntryID)
+                    stats["updated"] += 1
+                    if stats["updated"] % 100 == 0:
+                        conn.commit()
+                        print(f"  [{datetime.now().strftime('%H:%M:%S')}] "
+                              f"aktualizováno {stats['updated']} | "
+                              f"zbývá {len(lookup)} | složka: {current}")
+
+            except Exception as e:
+                stats["errors"] += 1
+
+    except Exception as e:
+        print(f"  CHYBA složka {current}: {e}")
+        stats["errors"] += 1
+        return  # nelze ani procházet podsložky
+
+    try:
+        subfolders = list(folder.Folders)
+    except Exception as e:
+        print(f"  CHYBA podsložky {current}: {e}")
+        return
+
+    for subfolder in subfolders:
+        if not lookup:
+            return
+        scan_folder(conn, subfolder, lookup, stats, current)
+
+
+def main():
+    print(f"=== backfill_entry_id v1.0 ===")
+    print(f"Start: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}")
+
+    conn = sqlite3.connect(DB_PATH)
+    lookup = load_missing(conn)
+    total_missing = len(lookup)
+    print(f"Záznamy bez entry_id: {total_missing}")
+
+    if not lookup:
+        print("Nic k doplnění.")
+        conn.close()
+        return
+
+    outlook = win32com.client.Dispatch("Outlook.Application")
+    ns = outlook.GetNamespace("MAPI")
+
+    stats = {"checked": 0, "updated": 0, "errors": 0}
+
+    for i in range(1, ns.Folders.Count + 1):
+        if not lookup:
+            break
+        root = ns.Folders.Item(i)
+        print(f"\nSložka: {root.Name}")
+        scan_folder(conn, root, lookup, stats, "")
+
+    conn.commit()
+    conn.close()
+
+    print(f"\n=== Hotovo ===")
+    print(f"Zkontrolováno emailů: {stats['checked']}")
+    print(f"Doplněno entry_id:    {stats['updated']} / {total_missing}")
+    print(f"Nenalezeno:           {len(lookup)}")
+    print(f"Chyby:                {stats['errors']}")
+    print(f"Konec: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}")
+
+
+if __name__ == "__main__":
+    main()
@@ -0,0 +1,90 @@
+"""
+check_msg_files.py
+Zkontroluje, zda má každý záznam v jnjemails SQLite odpovídající .msg soubor
+fyzicky uložený na \\\\tower\\JNJEMAILS\\.
+
+DB:    \\\\tower\\JNJEMAILS\\db\\jnjemails_*.db  (nejnovější)
+Soubory: \\\\tower\\JNJEMAILS\\*.msg
+
+Název souboru = entry_id[-20:] + ".msg"
+Záznamy bez entry_id mají fallback message_id "entryid:..." — ty se přeskočí
+zvlášť (server je nemohl uložit standardním názvem).
+"""
+
+import sqlite3
+import sys
+from pathlib import Path
+
+sys.stdout.reconfigure(encoding="utf-8")
+
+DB_DIR   = Path(r"\\tower\JNJEMAILS\db")
+MSGS_DIR = Path(r"\\tower\JNJEMAILS")
+
+
+def get_latest_db() -> Path:
+    files = sorted(DB_DIR.glob("jnjemails_*.db"), key=lambda f: f.name)
+    if not files:
+        raise FileNotFoundError(f"Žádný jnjemails_*.db v {DB_DIR}")
+    return files[-1]
+
+
+def main():
+    db_path = get_latest_db()
+    print(f"DB: {db_path.name}")
+
+    conn = sqlite3.connect(db_path)
+    conn.row_factory = sqlite3.Row
+    rows = conn.execute(
+        "SELECT id, message_id, subject, sender, received_at, entry_id, source FROM messages"
+    ).fetchall()
+    conn.close()
+
+    print(f"Celkem záznamů: {len(rows)}\n")
+
+    missing = []
+    no_entry_id = []
+
+    for row in rows:
+        entry_id = row["entry_id"]
+
+        if not entry_id:
+            no_entry_id.append(dict(row))
+            continue
+
+        expected_file = MSGS_DIR / (entry_id[-20:] + ".msg")
+        if not expected_file.exists():
+            missing.append({**dict(row), "expected_file": expected_file.name})
+
+    msg_files = sum(1 for _ in MSGS_DIR.glob("*.msg"))
+    print(f"Záznamy bez entry_id (nelze zkontrolovat): {len(no_entry_id)}")
+    print(f"Záznamy s entry_id:                        {len(rows) - len(no_entry_id)}")
+    print(f"Chybějící .msg soubory:                    {len(missing)}")
+    print(f"\n--- POROVNÁNÍ POČTŮ ---")
+    print(f"Záznamy v DB celkem:     {len(rows)}")
+    print(f"Soubory .msg na serveru: {msg_files}")
+    diff = msg_files - len(rows)
+    if diff >= 0:
+        print(f"Rozdíl: +{diff} souborů navíc (OK — všechny záznamy mají soubor)")
+    else:
+        print(f"Rozdíl: {diff} — CHYBÍ {abs(diff)} souborů!")
+
+    if missing:
+        print("\n--- CHYBĚJÍCÍ SOUBORY ---")
+        for r in missing:
+            print(f"  id={r['id']} | {r['received_at']} | {r['subject'][:60]!r}")
+            print(f"    sender={r['sender']} | source={r['source']}")
+            print(f"    entry_id={r['entry_id']}")
+            print(f"    očekávaný soubor: {r['expected_file']}")
+
+    if no_entry_id:
+        print(f"\n--- ZÁZNAMY BEZ ENTRY_ID ({len(no_entry_id)}) ---")
+        for r in no_entry_id[:20]:
+            print(f"  id={r['id']} | {r['received_at']} | {r['subject'][:60]!r} | source={r['source']}")
+        if len(no_entry_id) > 20:
+            print(f"  ... a dalších {len(no_entry_id) - 20}")
+
+    print("\nHotovo.")
+
+
+if __name__ == "__main__":
+    main()
@@ -0,0 +1,129 @@
+"""
+wipe_jnj_mailbox.py | 2026-06-08
+Vyčistí složku Inbox/JNJ ve schránce vladimir.buzalka@buzalka.cz PŘED testem mirroru.
+
+- Zachová samotnou složku Inbox/JNJ
+- Trvale smaže (permanentDelete — obchází Deleted Items) všechny zprávy v JNJ
+  i ve všech podsložkách
+- Smaže všechny podsložky JNJ (Inbox, Sent Items, Deleted Items, ...)
+
+Výsledek: Inbox/JNJ existuje a je prázdná. Mirror si podsložky vytvoří znovu.
+"""
+import sys
+import msal
+import requests
+
+sys.stdout.reconfigure(encoding="utf-8")
+
+GRAPH_TENANT_ID     = "7d269944-37a4-43a1-8140-c7517dc426e9"
+GRAPH_CLIENT_ID     = "4b222bfd-78c9-4239-a53f-43006b3ed07f"
+GRAPH_CLIENT_SECRET = "Txg8Q~MjhocuopxsJyJBhPmDfMxZ2r5WpTFj1dfk"
+GRAPH_MAILBOX       = "vladimir.buzalka@buzalka.cz"
+GRAPH_URL           = "https://graph.microsoft.com/v1.0"
+
+_token = None
+
+
+def token():
+    global _token
+    app = msal.ConfidentialClientApplication(
+        GRAPH_CLIENT_ID,
+        authority=f"https://login.microsoftonline.com/{GRAPH_TENANT_ID}",
+        client_credential=GRAPH_CLIENT_SECRET,
+    )
+    res = app.acquire_token_for_client(scopes=["https://graph.microsoft.com/.default"])
+    if "access_token" not in res:
+        raise RuntimeError(f"auth failed: {res}")
+    _token = res["access_token"]
+    return _token
+
+
+def H():
+    return {"Authorization": f"Bearer {_token or token()}"}
+
+
+def get_jnj_id():
+    r = requests.get(f"{GRAPH_URL}/users/{GRAPH_MAILBOX}/mailFolders/Inbox/childFolders?$top=100",
+                     headers=H(), timeout=20).json()
+    for f in r.get("value", []):
+        if f["displayName"] == "JNJ":
+            return f["id"]
+    return None
+
+
+def child_folders(fid):
+    out = []
+    url = f"{GRAPH_URL}/users/{GRAPH_MAILBOX}/mailFolders/{fid}/childFolders?$top=100"
+    while url:
+        r = requests.get(url, headers=H(), timeout=20).json()
+        out += r.get("value", [])
+        url = r.get("@odata.nextLink")
+    return out
+
+
+def all_descendants(root_id):
+    """Vrať [(id, displayName)] root + všech podsložek (BFS)."""
+    result = [(root_id, "JNJ")]
+    i = 0
+    while i < len(result):
+        fid = result[i][0]
+        i += 1
+        for f in child_folders(fid):
+            result.append((f["id"], f["displayName"]))
+    return result
+
+
+def wipe_messages(fid, name):
+    deleted = 0
+    while True:
+        r = requests.get(
+            f"{GRAPH_URL}/users/{GRAPH_MAILBOX}/mailFolders/{fid}/messages?$select=id&$top=100",
+            headers=H(), timeout=30).json()
+        msgs = r.get("value", [])
+        if not msgs:
+            break
+        for m in msgs:
+            pd = requests.post(
+                f"{GRAPH_URL}/users/{GRAPH_MAILBOX}/messages/{m['id']}/permanentDelete",
+                headers=H(), timeout=20)
+            if pd.status_code in (200, 204):
+                deleted += 1
+            else:
+                # fallback: běžné smazání
+                requests.delete(f"{GRAPH_URL}/users/{GRAPH_MAILBOX}/messages/{m['id']}",
+                                headers=H(), timeout=20)
+                deleted += 1
+    print(f"  {name}: smazáno {deleted} zpráv")
+    return deleted
+
+
+def main():
+    print("=== wipe_jnj_mailbox ===")
+    token()
+
+    jnj_id = get_jnj_id()
+    if not jnj_id:
+        print("Složka Inbox/JNJ neexistuje — není co mazat.")
+        return
+
+    folders = all_descendants(jnj_id)
+    print(f"Nalezeno složek pod JNJ (vč. JNJ): {len(folders)}\n")
+
+    print("Mažu zprávy (trvale)...")
+    total = 0
+    for fid, name in folders:
+        total += wipe_messages(fid, name)
+
+    # smaž podsložky JNJ (ne samotnou JNJ)
+    print("\nMažu podsložky JNJ...")
+    subs = child_folders(jnj_id)
+    for f in subs:
+        r = requests.delete(f"{GRAPH_URL}/users/{GRAPH_MAILBOX}/mailFolders/{f['id']}",
+                            headers=H(), timeout=20)
+        print(f"  podsložka {f['displayName']}: {'smazána' if r.status_code in (200,204) else 'CHYBA '+str(r.status_code)}")
+
+    print(f"\n=== Hotovo: smazáno {total} zpráv, Inbox/JNJ je prázdná ===")
+
+
+if __name__ == "__main__":
+    main()