janssen/mailstore/mailstore_folder_v1.0.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
================================================================================
 Nazev:   mailstore_folder_v1.0.py
 Verze:   1.0
 Datum:   2026-06-11
 Autor:   Vladimir Buzalka (asistovano Claude)
 Popis:   Vypise obsah jedne MailStore slozky jako seznam zprav
          (datum | od | predmet) pres davkovy IMAP FETCH hlavicek.
          Predstupen ingestu - overuje davkove cteni hlavicek.

          Argument = plna cesta slozky (fullName z mapy), napr.:
            "vladimir.buzalka@buzalka.cz/Exchange vladimir.buzalka/Sent Items"

 Zdroj:   MailStore IMAP server, port 143, STARTTLS, auth Prosty text (LOGIN).
          IMAP FETCH BODY.PEEK[HEADER.FIELDS (...)] = hlavicky bez oznaceni
          jako precteno. Davkove jednim prikazem, ne po jedne zprave.

 Spusteni:
   python mailstore_folder_v1.0.py "...slozka..."            # poslednich 50
   python mailstore_folder_v1.0.py "...slozka..." --limit 200
   python mailstore_folder_v1.0.py "...slozka..." --all       # vse (pozor velke slozky)
   python mailstore_folder_v1.0.py "...slozka..." --oldest    # od nejstarsich
================================================================================
"""

from __future__ import annotations

import argparse
import email
import imaplib
import re
import ssl
import sys
from email.header import decode_header
from email.utils import parsedate_to_datetime

# --- konfigurace ------------------------------------------------------------
HOST = "192.168.1.53"
PORT = 143
USER = "admin"
PASS = "*$N(B)vMUym!%"

DEFAULT_LIMIT = 50


# --- helpery ----------------------------------------------------------------

def connect() -> imaplib.IMAP4:
    ctx = ssl.create_default_context()
    ctx.check_hostname = False
    ctx.verify_mode = ssl.CERT_NONE
    M = imaplib.IMAP4(HOST, PORT)
    M.starttls(ssl_context=ctx)
    M.login(USER, PASS)
    return M


def encode_mutf7(s: str) -> str:
    """Nazev IMAP slozky -> modified UTF-7 (RFC 3501) kvuli diakritice
    (MailStore neumi UTF8=ACCEPT). Vysledek je cisty ASCII."""
    import base64 as _b64
    res = []
    i, n = 0, len(s)
    while i < n:
        ch = s[i]; o = ord(ch)
        if 0x20 <= o <= 0x7e:
            res.append("&-" if ch == "&" else ch); i += 1
        else:
            j = i
            while j < n and not (0x20 <= ord(s[j]) <= 0x7e):
                j += 1
            enc = _b64.b64encode(s[i:j].encode("utf-16-be")).decode("ascii").rstrip("=").replace("/", ",")
            res.append("&" + enc + "-"); i = j
    return "".join(res)


def dec(s: str | None) -> str:
    """Dekoduje MIME-encoded hlavicku (=?utf-8?...?=) na citelny text."""
    if not s:
        return ""
    out = []
    for txt, enc in decode_header(s):
        if isinstance(txt, bytes):
            out.append(txt.decode(enc or "utf-8", errors="replace"))
        else:
            out.append(txt)
    return "".join(out).replace("\r", " ").replace("\n", " ").strip()


def fmt_date(raw: str | None) -> str:
    if not raw:
        return "?"
    try:
        dt = parsedate_to_datetime(raw)
        return dt.strftime("%Y-%m-%d %H:%M")
    except Exception:
        return (raw or "")[:16]


def short(s: str, n: int) -> str:
    s = s or ""
    return s if len(s) <= n else s[: n - 1] + "…"


# IMAP FETCH header bloky prijdou jako tuple (b'N (BODY[...] {len}', b'<headers>')
_NUM_RX = re.compile(rb"^(\d+)\s")


def main() -> int:
    ap = argparse.ArgumentParser(description="Vypis obsahu MailStore slozky")
    ap.add_argument("folder", help="Plna cesta slozky (fullName z mapy)")
    ap.add_argument("--limit", type=int, default=DEFAULT_LIMIT,
                    help=f"Pocet zprav (default {DEFAULT_LIMIT})")
    ap.add_argument("--all", action="store_true", help="Vsechny zpravy (ignoruje --limit)")
    ap.add_argument("--oldest", action="store_true",
                    help="Od nejstarsich (default: od nejnovejsich)")
    args = ap.parse_args()

    M = connect()
    typ, data = M.select(f'"{encode_mutf7(args.folder)}"', readonly=True)
    if typ != "OK":
        print(f"Slozku nelze otevrit: {data}", file=sys.stderr)
        return 1
    total = int(data[0]) if data and data[0] else 0
    print(f"Slozka: {args.folder}")
    print(f"Zprav celkem: {total:,}")
    if total == 0:
        M.logout()
        return 0

    # urci rozsah porad. cisel (1 = nejstarsi, total = nejnovejsi)
    if args.all:
        lo, hi = 1, total
    else:
        n = min(args.limit, total)
        lo, hi = (1, n) if args.oldest else (total - n + 1, total)
    rng = f"{lo}:{hi}"
    shown = hi - lo + 1
    order = "nejstarsi" if args.oldest else "nejnovejsi"
    print(f"Zobrazuji {shown} zprav ({order} prvni), rozsah #{rng}")
    print("=" * 100)

    # davkovy FETCH hlavicek
    typ, msgs = M.fetch(rng, "(BODY.PEEK[HEADER.FIELDS (DATE FROM SUBJECT)])")
    rows = []
    for item in msgs:
        if not isinstance(item, tuple):
            continue
        meta, hdr_bytes = item[0], item[1]
        m = _NUM_RX.match(meta or b"")
        seqno = int(m.group(1)) if m else 0
        hdr = email.message_from_bytes(hdr_bytes)
        rows.append((seqno, fmt_date(hdr.get("Date")),
                     dec(hdr.get("From")), dec(hdr.get("Subject"))))

    rows.sort(key=lambda r: r[0], reverse=not args.oldest)

    print(f"{'#':>6}  {'Datum':<16}  {'Od':<32}  Predmet")
    print("-" * 100)
    for seqno, d, frm, subj in rows:
        print(f"{seqno:>6}  {d:<16}  {short(frm, 32):<32}  {short(subj, 40)}")

    M.logout()
    print("=" * 100)
    print(f"Vypsano {len(rows)} zprav.")
    return 0


if __name__ == "__main__":
    try:
        sys.exit(main())
    except KeyboardInterrupt:
        print("\nPreruseno", file=sys.stderr)
        sys.exit(1)