# =============================================================================
# Název:   download_test_results_v1.2.py
# Verze:   1.2
# Datum:   2026-05-29
# Popis:   Stahuje Test Results ze xsp.labcorp.com pro 2 studie (36940, 35472),
#          oba typy reportu (Standard + Microbiology), pres vsechna centra.
#          Ceka na nacteni AG Grid radku (.ag-row); prazdne centrum ('No Data')
#          preskoci. Vystup: timestampovane CSV do adresare Source/.
# Zmeny v1.2: + paralelni beh pres sharding (--shard N --of M). Kazdy shard
#             vezme svuj podil reportu (REPORTS[shard-1::of]) a pouziva vlastni
#             profil browser_profile_{shard}, takze 4 procesy mohou bezet
#             soucasne (kazdy se prihlasi do sveho profilu sam). Bez argumentu
#             = puvodni serialni beh nad profilem browser_profile.
# Zmeny v1.1: + studie 35472, + report typ microbiology (driv jen 36940/standard).
# =============================================================================
from playwright.sync_api import sync_playwright
from datetime import datetime
import argparse
import os

# --- argumenty: sharding pro paralelni beh ----------------------------------
parser = argparse.ArgumentParser(description="Stahovani Test Results (XSP) s podporou shardingu.")
parser.add_argument("--shard", type=int, default=1, help="poradi tohoto shardu (1..of)")
parser.add_argument("--of",    type=int, default=1, help="celkovy pocet shardu")
ARGS = parser.parse_args()
SHARD, OF = ARGS.shard, ARGS.of
TAG = f"[S{SHARD}/{OF}]" if OF > 1 else ""


def log(msg):
    print(f"[{datetime.now().strftime('%H:%M:%S')}] {TAG} {msg}", flush=True)

EMAIL       = "vbuzalka@its.jnj.com"
PASSWORD    = "%zT3Wqfc9)cWua5"
LOGIN_URL   = "https://xsp.covance.com/"
OUT_DIR     = r"U:\PythonProject\Janssen\Covance_UCO3001\Source"
# Pri paralelnim behu MUSI mit kazdy shard vlastni profil (Chrome zamyka adresar
# profilu -> dve bezici instance nemohou sdilet jeden). Serialni beh (of=1)
# pouziva puvodni browser_profile.
_BASE = os.path.dirname(os.path.abspath(__file__))
PROFILE_DIR = os.path.join(_BASE, "browser_profile" if OF == 1 else f"browser_profile_{SHARD}")

# Studie + jejich interni cisla center.
#  36940 = 77242113UCO3001 (UC)  — zdroj center: download_equeries_report SITES
#  35472 = druha studie (MDD)
STUDIES = [
    {
        "study": "36940",
        "sites": [
            "930551", "930556", "930525", "930549", "930543", "930547",
            "930555", "930557", "930539", "930536", "930553", "930531",
        ],
    },
    {
        "study": "35472",
        "sites": [
            "898745", "898739", "898733", "898744", "898727",
        ],
    },
]

# Typy reportu: zalozka v URL + suffix v nazvu souboru.
REPORT_TYPES = [
    {"slug": "standard-test-results", "suffix": "standard"},
    {"slug": "microbiology",          "suffix": "microbiology"},
]

ALL_REPORTS = [
    {
        "site":     sid,
        "study":    st["study"],
        "type":     rt["suffix"],
        "url":      f"https://xsp.labcorp.com/sponsor/study/{st['study']}/test-results/{sid}/{rt['slug']}",
        "filename": f"sponsor-study-{st['study']}-test-results-{sid}-{rt['suffix']}.csv",
    }
    for st in STUDIES
    for sid in st["sites"]
    for rt in REPORT_TYPES
]

# Tento shard vezme kazdy of-ty report od indexu (shard-1). Rovnomerne rozdeleni
# a zadny report neudela dva shardy zaroven.
REPORTS = ALL_REPORTS[SHARD - 1::OF]


def login(page):
    log("LOGIN: otviram login stranku...")
    page.goto(LOGIN_URL)
    page.wait_for_load_state("networkidle")
    if not page.get_by_label("Email").is_visible():
        log(f"LOGIN: session uz aktivni, prihlaseni preskoceno ({page.url})")
        return
    log("LOGIN: zadavam email...")
    page.get_by_label("Email").fill(EMAIL)
    page.get_by_role("button", name="Next").click()
    page.wait_for_load_state("networkidle")
    log("LOGIN: zadavam heslo...")
    page.get_by_label("Password").fill(PASSWORD)
    page.get_by_role("button", name="Verify").click()
    log("LOGIN: cekam na presmerovani po prihlaseni...")
    page.wait_for_url(lambda url: "code=" not in url, timeout=60000)
    page.wait_for_load_state("networkidle", timeout=60000)
    page.wait_for_timeout(2000)
    log(f"LOGIN: prihlaseni OK ({page.url})")


def download_report(page, report):
    log(f"=== Centrum {report['site']} / {report['type']} (studie {report['study']}) ===")

    log(f"KROK 1/5: navigace na report URL...")
    page.goto(report["url"])
    log(f"KROK 1/5: stranka nactena ({page.url})")

    # Grid je AG Grid uvnitř <covance-ag-grid>. Data jsou nactena, jakmile
    # se v gridu objevi radky (.ag-row jde z 0 -> N). Pockej na prvni radek
    # a pak na stabilizaci poctu (proti castecnemu renderu).
    log("KROK 2/5: cekam na radky gridu (.ag-row) nebo prazdny grid ('No Data')...")
    # AG Grid radky jsou position-absolute (virtualni render), takze nejsou
    # "visible" dle Playwrightu -> cekej na pritomnost v DOM, ne na viditelnost.
    # Prazdne centrum: AG Grid vykresli no-rows overlay s textem "No Data" ve
    # wrapperu .ag-overlay-no-rows-wrapper. POZOR: trida NENI -no-rows-center;
    # navic jsou na strance 2 overlaye (jeden skryty) -> kontroluj viditelny
    # (offsetParent != null). Detekuj, aby to u centra bez dat necekalo 120 s.
    EMPTY_GRID_JS = """() => {
        if (document.querySelectorAll('div.ag-row').length > 0) return false;
        return [...document.querySelectorAll('.ag-overlay-no-rows-wrapper')]
            .some(e => e.offsetParent !== null);
    }"""
    page.wait_for_function(
        f"""() => document.querySelectorAll('div.ag-row').length > 0
              || ({EMPTY_GRID_JS})()""",
        timeout=120000,
    )
    if page.evaluate(EMPTY_GRID_JS):
        log("KROK 2/5: centrum bez dat ('No Data' overlay) — preskakuji export.")
        return
    log("KROK 2/5: radky se objevily, cekam na stabilizaci poctu...")
    prev = -1
    for i in range(20):  # max ~40 s stabilizace
        cnt = page.locator("div.ag-row").count()
        log(f"          ...kontrola #{i+1}: {cnt} radku")
        if cnt == prev and cnt > 0:
            break
        prev = cnt
        page.wait_for_timeout(2000)
    page.wait_for_timeout(2000)  # buffer
    log(f"KROK 2/5: data stabilni ({prev} radku v gridu).")

    # Tri tecky: na strance jsou 2x <ag-export> (jeden skryty), klikni na
    # VIDITELNY more_horiz button.
    log("KROK 3/5: klikam na viditelne tri tecky (more_horiz)...")
    page.locator("ag-export button:visible", has_text="more_horiz").first.click()
    log("KROK 3/5: menu otevreno.")

    timestamp = datetime.now().strftime("%Y-%m-%d_%H%M%S")
    dest = os.path.join(OUT_DIR, f"{timestamp} {report['filename']}")
    log("KROK 4/5: klikam na 'Export to CSV' a cekam na stahovani...")
    with page.expect_download(timeout=60000) as dl:
        # 2x "Export to CSV" v DOM (jeden skryty) -> klikni na VIDITELNY
        page.locator("mdl-menu-item:visible", has_text="Export to CSV").first.click()
    log("KROK 4/5: stahovani zachyceno, ukladam soubor...")
    dl.value.save_as(dest)
    log(f"KROK 5/5: HOTOVO -> {dest}")


if __name__ == "__main__":
    log(f"START: shard {SHARD}/{OF}, profil '{os.path.basename(PROFILE_DIR)}', "
        f"{len(REPORTS)}/{len(ALL_REPORTS)} reportu k zpracovani.")
    with sync_playwright() as p:
        context = p.chromium.launch_persistent_context(
            user_data_dir=PROFILE_DIR,
            headless=False,
            args=[
                "--disable-blink-features=AutomationControlled",
                "--start-maximized",
                "--disable-restore-session-state",
                "--disable-session-crashed-bubble",
            ],
            no_viewport=True,
            user_agent="Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/136.0.0.0 Safari/537.36",
            accept_downloads=True,
        )
        context.add_init_script("Object.defineProperty(navigator, 'webdriver', {get: () => undefined})")
        page = context.new_page()
        log("START: prohlizec spusten.")
        login(page)
        ok, failed = 0, []
        for idx, report in enumerate(REPORTS, 1):
            log(f">>> Report {idx}/{len(REPORTS)}")
            try:
                download_report(page, report)
                ok += 1
            except Exception as e:
                failed.append(f"{report['site']}/{report['type']}")
                log(f"CHYBA u centra {report['site']}/{report['type']}: {e!r} — pokracuji dalsim.")
        log(f"KONEC: hotovo {ok}/{len(REPORTS)} reportu (shard {SHARD}/{OF}).")
        if failed:
            log(f"KONEC: SELHALA centra: {', '.join(failed)}")
        context.close()