janssen/Outlook/save_first_email_per_folder_v1.0.py

"""
=======================================================================
 Název:   save_first_email_per_folder_v1.0.py
 Verze:   1.0
 Datum:   2026-06-04
 Popis:   Testovací skript: projde všechny složky OWA, z každé stáhne
          první email jako EML a uloží ho do downloads/.

          Používá persistent profil z outlook_login_v1.0.py.
=======================================================================
"""

import tempfile
from pathlib import Path

from playwright.sync_api import sync_playwright

BASE_DIR    = Path(__file__).resolve().parent
PROFILE_DIR = BASE_DIR / "outlook_profile"
OUT_DIR     = BASE_DIR / "downloads"
START_URL   = "https://outlook.cloud.microsoft/mail/"

# Složky: (zobrazovaný název, způsob navigace, hodnota)
# způsob: "url" = přímá URL, "click" = klik na text v levém panelu
FOLDERS = [
    ("Inbox",              "url",   "https://outlook.cloud.microsoft/mail/"),
    ("Sent Items",         "url",   "https://outlook.cloud.microsoft/mail/sentitems"),
    ("Deleted Items",      "url",   "https://outlook.cloud.microsoft/mail/deleteditems"),
    ("Drafts",             "url",   "https://outlook.cloud.microsoft/mail/drafts"),
    ("Junk Email",         "url",   "https://outlook.cloud.microsoft/mail/junkemail"),
    ("Archive",            "url",   "https://outlook.cloud.microsoft/mail/archive"),
    ("77242113UCO3002",    "click", "77242113UCO3002"),
    ("Hrabak",             "click", "Hrabak"),
    ("Request assessment", "click", "Request assessment"),
    ("SAM tickets",        "click", "SAM tickets"),
    ("TMP",                "click", "TMP"),
]

SEARCH_READY = (
    '[placeholder*="Search"], [aria-label*="Search"], '
    '[placeholder*="Hledat"], [aria-label*="Hledat"]'
)


def wait_ready(page):
    page.wait_for_load_state("domcontentloaded")
    page.wait_for_selector(SEARCH_READY, timeout=30_000)


def navigate_to_folder(page, nav_type, value):
    if nav_type == "url":
        page.goto(value)
        wait_ready(page)
    else:
        # klik na text v levém panelu — čekej na viditelný element
        loc = page.locator(f'div[role="treeitem"]:has-text("{value}")').last
        loc.wait_for(state="visible", timeout=10_000)
        loc.click()
    page.wait_for_timeout(1_500)


def download_first_eml(page, tmp_dir):
    """Klikne na první zprávu, pravý klik → Download → EML.
    Vrátí (path, suggested_filename) nebo None pokud složka prázdná."""

    msgs = page.locator('div[role="option"]')
    if msgs.count() == 0:
        # zkus počkat chvíli
        try:
            page.wait_for_selector('div[role="option"]', timeout=5_000)
        except Exception:
            return None

    first = msgs.first
    if not first.is_visible():
        return None

    first.click()
    page.wait_for_timeout(800)
    first.click(button="right")
    page.wait_for_timeout(700)

    # Najdi "Download" v menu
    download_parent = None
    for name in ("Download", "Stáhnout"):
        loc = page.get_by_role("menuitem", name=name).first
        if loc.count() and loc.is_visible():
            download_parent = loc
            break

    if download_parent is None:
        print("     ! 'Download' položka v menu nenalezena — přeskakuji")
        page.keyboard.press("Escape")
        return None

    download_parent.hover()
    page.wait_for_timeout(600)

    eml_item = None
    for name in ("Download as EML", "Stáhnout jako EML", "Stáhnout jako .eml"):
        loc = page.get_by_role("menuitem", name=name).first
        if loc.count() and loc.is_visible():
            eml_item = loc
            break

    try:
        target = eml_item if eml_item else download_parent
        with page.expect_download(timeout=20_000) as dl_info:
            target.click()
        dl = dl_info.value
        out_path = Path(tmp_dir) / (dl.suggested_filename or "email.eml")
        dl.save_as(str(out_path))
        return out_path
    except Exception as e:
        print(f"     ! Stažení selhalo: {e}")
        page.keyboard.press("Escape")
        return None


def save_eml(folder_name, eml_path):
    OUT_DIR.mkdir(exist_ok=True)
    # Název souboru: složka__původní_název.eml
    safe = folder_name.replace(" ", "_").replace("/", "-")
    dest = OUT_DIR / f"{safe}__{eml_path.name}"
    if dest.exists():
        dest.unlink()
    dest.write_bytes(eml_path.read_bytes())
    print(f"     + uloženo: {dest.name} ({dest.stat().st_size:,} B)")
    return dest


def main():
    if not PROFILE_DIR.exists():
        print(f"Profil nenalezen: {PROFILE_DIR}")
        print("Nejprve spusť outlook_login_v1.0.py.")
        return

    with tempfile.TemporaryDirectory() as tmp_dir:
        with sync_playwright() as p:
            context = p.chromium.launch_persistent_context(
                user_data_dir=str(PROFILE_DIR),
                headless=False,
                no_viewport=True,
                accept_downloads=True,
                args=[
                    "--disable-blink-features=AutomationControlled",
                    "--start-maximized",
                ],
            )
            page = context.pages[0] if context.pages else context.new_page()

            print("Otevírám Outlook...")
            page.goto(START_URL)
            wait_ready(page)

            results = []
            for folder_name, nav_type, value in FOLDERS:
                print(f"\n[{folder_name}]")
                try:
                    navigate_to_folder(page, nav_type, value)
                    eml_path = download_first_eml(page, tmp_dir)
                    if eml_path is None:
                        print("     - složka prázdná nebo stažení selhalo")
                        results.append((folder_name, "skip"))
                        continue
                    dest = save_eml(folder_name, eml_path)
                    results.append((folder_name, f"ok → {dest.name}"))
                except Exception as e:
                    print(f"     ! Chyba: {e}")
                    results.append((folder_name, f"error: {e}"))

            context.close()

    print("\n=== Výsledky ===")
    for name, status in results:
        print(f"  {name:<25} {status}")


if __name__ == "__main__":
    main()