405 lines
16 KiB
Python
405 lines
16 KiB
Python
# =============================================================================
|
|
# Název: download_lab_reports_v1.2.py
|
|
# Verze: 1.2
|
|
# Datum: 2026-06-16
|
|
# Popis: Stahuje PDF Lab Reports ze xsp.labcorp.com pro studii 77242113UCO3001
|
|
# (interni cislo 36940), filtrovane na 10 ceskych center (CZ),
|
|
# a uklada je PRIMO do MongoDB (db covance, kolekce labreports) —
|
|
# metadata z tabulky + skutecne PDF (inline Binary). Na disk NEUKLADA.
|
|
#
|
|
# KLIC: record_id = reportId (z dat AG Gridu) — stabilni 32-hex ID
|
|
# dokumentu, NAPRIC vsemi radky UNIKATNI (overeno: 997 radku ->
|
|
# 997 unikatnich reportId/fileId). Resi pripad, kdy se na portalu
|
|
# vyskytnou ruzna PDF se SHODNYMI viditelnymi metadaty (stejny
|
|
# site|subject|accession|visit|posted i na minutu) — to skutecne
|
|
# nastava (korekce vysledku reissue se shodnym casem Posted).
|
|
# Verze v1.1 klicovala podle metadat a tyto by chybne slucovala.
|
|
#
|
|
# Princip stahovani: Playwright + perzistentni profil, login, klik na
|
|
# "English" ve sloupci Download -> expect_download; PDF bajty se ctou
|
|
# z Playwright temp souboru (download.path()), save_as se nevola.
|
|
#
|
|
# INKREMENTALNE (stop-at-known): list je Posted DESC (nejnovejsi
|
|
# nahore). Nejdriv se z grid API precte SEZNAM (rowIndex, reportId)
|
|
# BEZ stahovani; od shora se hleda prvni uz ulozeny reportId -> vse
|
|
# pod nim je starsi a uz v Mongo je. Stahuji se jen nove (nahore).
|
|
#
|
|
# Prepinace:
|
|
# --full projit vsechny radky (bez stop-at-known); upsert
|
|
# chybejicich (rekonciliace).
|
|
# --dry-run nestahuje ani nepise do DB; jen vypise NOVE reporty.
|
|
# --limit N zpracovat max N novych radku (test).
|
|
# =============================================================================
|
|
from playwright.sync_api import sync_playwright
|
|
from datetime import datetime
|
|
from pymongo import MongoClient, ASCENDING
|
|
from bson.binary import Binary
|
|
import argparse
|
|
import hashlib
|
|
import json
|
|
import os
|
|
import re
|
|
import traceback
|
|
import urllib.parse
|
|
|
|
parser = argparse.ArgumentParser(description="Stahovani Lab Reports PDF (XSP) do MongoDB.")
|
|
parser.add_argument("--full", action="store_true", help="projit vse (bez stop-at-known)")
|
|
parser.add_argument("--dry-run", action="store_true", help="nestahovat ani nepsat do DB; jen vypsat nove")
|
|
parser.add_argument("--limit", type=int, default=0, help="max N novych radku (0 = vse)")
|
|
ARGS = parser.parse_args()
|
|
|
|
|
|
def log(msg):
|
|
print(f"[{datetime.now().strftime('%H:%M:%S')}] {msg}", flush=True)
|
|
|
|
|
|
# --- konfigurace ------------------------------------------------------------
|
|
EMAIL = "vbuzalka@its.jnj.com"
|
|
PASSWORD = "%zT3Wqfc9)cWua5"
|
|
LOGIN_URL = "https://xsp.labcorp.com/"
|
|
STUDY = "36940"
|
|
STUDY_CODE = "77242113UCO3001"
|
|
|
|
MONGO_URI = "mongodb://192.168.1.76:27017"
|
|
DB_NAME = "covance"
|
|
COLLECTION = "labreports"
|
|
|
|
SITES = [
|
|
"930539", "930547", "930555", "930556", "930553",
|
|
"930549", "930525", "930536", "930557", "930531",
|
|
]
|
|
|
|
_BASE = os.path.dirname(os.path.abspath(__file__))
|
|
PROFILE_DIR = os.path.join(_BASE, "browser_profile")
|
|
|
|
|
|
def lab_reports_url():
|
|
site_param = json.dumps(SITES, separators=(",", ":"))
|
|
return (f"https://xsp.labcorp.com/sponsor/study/{STUDY}/lab-reports"
|
|
f"?site={urllib.parse.quote(site_param)}")
|
|
|
|
|
|
# --- formatovani -------------------------------------------------------------
|
|
_MONTHS = {"Jan": "01", "Feb": "02", "Mar": "03", "Apr": "04", "May": "05",
|
|
"Jun": "06", "Jul": "07", "Aug": "08", "Sep": "09", "Oct": "10",
|
|
"Nov": "11", "Dec": "12"}
|
|
|
|
|
|
def safe(s: str) -> str:
|
|
return re.sub(r'[\\/:*?"<>|]', "", s or "").strip()
|
|
|
|
|
|
def fmt_date(s: str) -> str:
|
|
"""'Jun 10, 2026' i 'Jun 15, 2026 7:49 PM' -> '2026-06-10'."""
|
|
m = re.match(r"\s*([A-Za-z]{3})\w*\s+(\d{1,2}),\s+(\d{4})", s or "")
|
|
if m and m.group(1)[:3] in _MONTHS:
|
|
return f"{m.group(3)}-{_MONTHS[m.group(1)[:3]]}-{int(m.group(2)):02d}"
|
|
return safe(s)
|
|
|
|
|
|
def build_basename(meta: dict) -> str:
|
|
"""Lidsky citelny nazev (pro budouci materializacni skript), ulozen jako fileName."""
|
|
return safe(
|
|
f"{STUDY_CODE} {fmt_date(meta['collected'])} {meta['site']} "
|
|
f"{meta['subject']} {meta['visit']} {meta['accession']} "
|
|
f"posted {fmt_date(meta['postedDisplay'])}"
|
|
) + ".pdf"
|
|
|
|
|
|
# --- JS helpery (AG Grid) ---------------------------------------------------
|
|
# Seznam vsech radku (rowIndex + reportId + data, ktera nepotrebuji vykresleni).
|
|
JS_ALL_ROWS = r"""() => {
|
|
let holder = null;
|
|
for (const el of document.querySelectorAll('.ag-root-wrapper, .ag-root, ag-grid-angular')) {
|
|
if (el.__agComponent) { holder = el.__agComponent; break; }
|
|
}
|
|
if (!holder) return null;
|
|
const api = (holder.gridApi && holder.gridApi.forEachNode) ? holder.gridApi : holder;
|
|
if (!api || !api.getDisplayedRowCount) return null;
|
|
const cnt = api.getDisplayedRowCount();
|
|
const out = [];
|
|
for (let i = 0; i < cnt; i++) {
|
|
const n = api.getDisplayedRowAtIndex(i);
|
|
if (!n || !n.data) continue;
|
|
const d = n.data;
|
|
const fl = (d.fileLinks || []).find(f => f.language === 'English') || (d.fileLinks || [])[0] || {};
|
|
out.push({
|
|
rowIndex: i,
|
|
reportId: d.reportId,
|
|
fileId: fl.fileId,
|
|
serverFileName: fl.fileName,
|
|
postedIso: d.postedDateTime,
|
|
site: d.siteNum,
|
|
subject: d.subjectNumber,
|
|
});
|
|
}
|
|
return out;
|
|
}"""
|
|
|
|
# Bunky daneho radku (potrebuji vykresleni -> nejdriv scroll). Accession/Visit/
|
|
# Collected nejsou v top-level datech (jsou ve 'visits'), beru je z bunek.
|
|
JS_CELLS = r"""(idx) => {
|
|
const dedup = s => {
|
|
s = (s || '').replace(/\s+/g, ' ').trim();
|
|
const h = s.slice(0, Math.floor(s.length / 2));
|
|
if (s === h + h) return h;
|
|
const m = s.match(/^(.*?)\s+\1$/);
|
|
if (m) return m[1];
|
|
return s;
|
|
};
|
|
const row = document.querySelector('.ag-body-container .ag-row[row-index="' + idx + '"]');
|
|
if (!row) return null;
|
|
const get = id => { const c = row.querySelector('[col-id="' + id + '"]'); return c ? dedup(c.textContent) : ''; };
|
|
return {
|
|
type: get('type'),
|
|
accession: get('accessionNumber'),
|
|
visit: get('visit'),
|
|
collected: get('visitCollectionDate'),
|
|
postedDisplay: get('postedDateTime'),
|
|
};
|
|
}"""
|
|
|
|
JS_SCROLL_TO = r"""(args) => {
|
|
const [idx, rh] = args;
|
|
const vp = document.querySelector('.ag-body-viewport');
|
|
if (!vp) return;
|
|
vp.scrollTop = Math.max(0, idx * rh - vp.clientHeight / 2);
|
|
}"""
|
|
|
|
JS_ROW_HEIGHT = r"""() => {
|
|
const r = document.querySelector('.ag-body-container .ag-row');
|
|
return r ? r.getBoundingClientRect().height || 25 : 25;
|
|
}"""
|
|
|
|
|
|
# --- login ------------------------------------------------------------------
|
|
def login(page):
|
|
log("LOGIN: otviram login stranku...")
|
|
page.goto(LOGIN_URL)
|
|
try:
|
|
page.get_by_label("Email").wait_for(state="visible", timeout=12000)
|
|
except Exception:
|
|
log(f"LOGIN: Email pole se neobjevilo -> session aktivni, login preskocen ({page.url})")
|
|
return
|
|
log("LOGIN: zadavam email...")
|
|
page.get_by_label("Email").fill(EMAIL)
|
|
page.get_by_role("button", name="Next").click()
|
|
log("LOGIN: cekam na pole pro heslo...")
|
|
page.get_by_label("Password").wait_for(state="visible", timeout=30000)
|
|
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...")
|
|
try:
|
|
page.wait_for_url(lambda url: "code=" not in url or "xsp." in url, timeout=60000)
|
|
except Exception:
|
|
log("LOGIN: wait_for_url vyprsel, pokracuji.")
|
|
page.wait_for_timeout(3000)
|
|
log(f"LOGIN: prihlaseni hotovo ({page.url})")
|
|
|
|
|
|
def open_grid(page):
|
|
log(f"GRID: navigace na Lab Reports ({len(SITES)} center)...")
|
|
page.goto(lab_reports_url())
|
|
log("GRID: cekam na radky (.ag-row)...")
|
|
page.wait_for_selector(".ag-body-container .ag-row", timeout=120000)
|
|
# cekej, az grid API hlasi stabilni pocet radku
|
|
prev = -1
|
|
rows = None
|
|
for i in range(25):
|
|
rows = page.evaluate(JS_ALL_ROWS)
|
|
cnt = len(rows) if rows else 0
|
|
log(f" ...kontrola #{i+1}: rows={cnt}")
|
|
if rows and cnt == prev and cnt > 0:
|
|
break
|
|
prev = cnt
|
|
page.wait_for_timeout(2000)
|
|
row_height = page.evaluate(JS_ROW_HEIGHT)
|
|
log(f"GRID: nacteno {len(rows) if rows else 0} radku, rowHeight={row_height}px.")
|
|
return rows or [], row_height
|
|
|
|
|
|
def download_pdf_bytes(page, idx):
|
|
link = page.locator(
|
|
f'.ag-body-container .ag-row[row-index="{idx}"] a.dl-link',
|
|
has_text="English",
|
|
).first
|
|
with page.expect_download(timeout=60000) as dl:
|
|
link.click()
|
|
with open(dl.value.path(), "rb") as f:
|
|
return f.read()
|
|
|
|
|
|
def upsert(col, rec, cells, data, now):
|
|
fields = {
|
|
"Type": cells["type"],
|
|
"Subject": rec["subject"],
|
|
"Accession": cells["accession"],
|
|
"Visit": cells["visit"],
|
|
"Collected Date": cells["collected"],
|
|
"Site Number": rec["site"],
|
|
"Posted": cells["postedDisplay"],
|
|
}
|
|
sha = hashlib.sha256(data).hexdigest()
|
|
meta = {"site": rec["site"], "subject": rec["subject"],
|
|
"accession": cells["accession"], "visit": cells["visit"],
|
|
"collected": cells["collected"], "postedDisplay": cells["postedDisplay"]}
|
|
derived = {
|
|
"study": STUDY,
|
|
"studyCode": STUDY_CODE,
|
|
"type": cells["type"] or "Lab Result",
|
|
"site": rec["site"],
|
|
"subject": rec["subject"],
|
|
"accession": cells["accession"],
|
|
"visit": cells["visit"],
|
|
"collected": fmt_date(cells["collected"]),
|
|
"posted": cells["postedDisplay"],
|
|
"postedIso": rec["postedIso"],
|
|
"fileId": rec["fileId"],
|
|
"serverFileName": rec["serverFileName"],
|
|
"fields": fields,
|
|
"fileName": build_basename(meta),
|
|
"pdf": Binary(data),
|
|
"pdfSize": len(data),
|
|
"pdfSha256": sha,
|
|
}
|
|
rid = rec["reportId"]
|
|
existing = col.find_one({"record_id": rid}, {"pdfSha256": 1, "fields": 1, "lastSeen": 1})
|
|
if existing is None:
|
|
col.insert_one({"record_id": rid, **derived,
|
|
"firstSeen": now, "lastSeen": now, "history": []})
|
|
return "insert"
|
|
if existing.get("pdfSha256") != sha or existing.get("fields") != fields:
|
|
col.update_one(
|
|
{"_id": existing["_id"]},
|
|
{"$push": {"history": {"date": existing.get("lastSeen"),
|
|
"fields": existing.get("fields"),
|
|
"pdfSha256": existing.get("pdfSha256")}},
|
|
"$set": {**derived, "lastSeen": now}},
|
|
)
|
|
return "update"
|
|
col.update_one({"_id": existing["_id"]}, {"$set": {"lastSeen": now}})
|
|
return "same"
|
|
|
|
|
|
def main():
|
|
now = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
|
|
log(f"START: studie {STUDY_CODE} ({STUDY}) -> {DB_NAME}.{COLLECTION}, "
|
|
f"{'DRY-RUN' if ARGS.dry_run else 'ZAPIS'}"
|
|
f"{' [FULL]' if ARGS.full else ' [stop-at-known]'}"
|
|
f"{f', limit {ARGS.limit}' if ARGS.limit else ''}.")
|
|
|
|
col = None
|
|
existing_ids = set()
|
|
client = MongoClient(MONGO_URI, serverSelectionTimeoutMS=5000)
|
|
client.admin.command("ping")
|
|
if not ARGS.dry_run:
|
|
col = client[DB_NAME][COLLECTION]
|
|
col.create_index([("record_id", ASCENDING)], unique=True)
|
|
for f in ("study", "site", "subject", "accession", "postedIso", "fileId"):
|
|
col.create_index([(f, ASCENDING)])
|
|
existing_ids = {d["record_id"] for d in
|
|
client[DB_NAME][COLLECTION].find({"study": STUDY}, {"record_id": 1})}
|
|
log(f"START: v Mongo je {len(existing_ids)} reportu pro tuto studii.")
|
|
|
|
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)
|
|
rows, row_height = open_grid(page)
|
|
|
|
# POJISTKA: jen CZ centra. URL filtr by mel vratit jen 10 CZ sites,
|
|
# ale kdyby selhal (napr. ztrata filtru po nejakem reloadu), tahle
|
|
# kontrola zabrani stazeni cizich center.
|
|
non_cz = [r for r in rows if not str(r["site"]).startswith("CZ")]
|
|
if non_cz:
|
|
log(f"POZOR: {len(non_cz)} ne-CZ radku v gridu (napr. {non_cz[0]['site']}) "
|
|
f"-> filtruji jen CZ. Zkontroluj URL filtr center!")
|
|
rows = [r for r in rows if str(r["site"]).startswith("CZ")]
|
|
log(f"GRID: po CZ-pojistce {len(rows)} CZ radku.")
|
|
|
|
# vyber radky ke zpracovani: shora dolu, stop-at-known
|
|
todo = []
|
|
for rec in rows:
|
|
if rec["reportId"] in existing_ids:
|
|
if ARGS.full:
|
|
continue
|
|
log(f"STOP-AT-KNOWN: rowIndex {rec['rowIndex']} (reportId {rec['reportId'][:12]}…) "
|
|
f"uz v Mongo -> koncim vyber (zbytek je starsi).")
|
|
break
|
|
todo.append(rec)
|
|
if ARGS.limit:
|
|
todo = todo[:ARGS.limit]
|
|
log(f"PLAN: {len(todo)} novych radku ke stazeni "
|
|
f"(z {len(rows)} v gridu, {len(existing_ids)} uz v Mongo).")
|
|
|
|
new_cnt = upd_cnt = same_cnt = 0
|
|
failed = []
|
|
for k, rec in enumerate(todo, 1):
|
|
idx = rec["rowIndex"]
|
|
try:
|
|
page.evaluate(JS_SCROLL_TO, [idx, row_height])
|
|
page.wait_for_selector(f'.ag-body-container .ag-row[row-index="{idx}"]', timeout=15000)
|
|
page.wait_for_timeout(120)
|
|
cells = page.evaluate(JS_CELLS, idx)
|
|
if not cells:
|
|
raise RuntimeError("nepodarilo se precist bunky radku")
|
|
meta = {"site": rec["site"], "subject": rec["subject"],
|
|
"accession": cells["accession"], "visit": cells["visit"],
|
|
"collected": cells["collected"], "postedDisplay": cells["postedDisplay"]}
|
|
fname = build_basename(meta)
|
|
|
|
if ARGS.dry_run:
|
|
log(f">>> {k}/{len(todo)} [DRY] NOVY rowIdx {idx} ({rec['reportId'][:12]}…): {fname}")
|
|
new_cnt += 1
|
|
continue
|
|
|
|
data = download_pdf_bytes(page, idx)
|
|
action = upsert(col, rec, cells, data, now)
|
|
existing_ids.add(rec["reportId"])
|
|
if action == "insert":
|
|
new_cnt += 1
|
|
log(f">>> {k}/{len(todo)} INSERT rowIdx {idx} ({len(data)//1024} KB): {fname}")
|
|
elif action == "update":
|
|
upd_cnt += 1
|
|
log(f">>> {k}/{len(todo)} UPDATE rowIdx {idx}: {fname}")
|
|
else:
|
|
same_cnt += 1
|
|
log(f">>> {k}/{len(todo)} SAME rowIdx {idx}: {fname}")
|
|
except Exception as e:
|
|
failed.append(idx)
|
|
log(f"CHYBA rowIdx {idx} ({rec['reportId'][:12]}…): {e!r} — pokracuji.")
|
|
|
|
log(f"KONEC: nove={new_cnt}, update={upd_cnt}, beze zmeny={same_cnt}, chyby={len(failed)}.")
|
|
if failed:
|
|
log(f"KONEC: SELHALY rowIndexy: {failed}")
|
|
context.close()
|
|
|
|
|
|
if __name__ == "__main__":
|
|
try:
|
|
main()
|
|
except Exception as e:
|
|
log(f"FATAL: beh spadl: {e!r}")
|
|
traceback.print_exc()
|
|
finally:
|
|
try:
|
|
input("\n[Enter] pro zavreni tohoto okna...")
|
|
except EOFError:
|
|
pass
|