This commit is contained in:
2025-09-22 10:45:44 +02:00
parent 2d7204a9c0
commit fb9486d0af
6 changed files with 217 additions and 68 deletions

View File

@@ -1,42 +0,0 @@
# save_patient_detail_page.py
from pathlib import Path
from playwright.sync_api import sync_playwright
STATE_FILE = r"/medevio_storage.json"
BASE_URL = "https://my.medevio.cz/mudr-buzalkova/klinika/pacienti"
PATIENT_ID = "fcb2414b-067b-4ca2-91b2-6c36a86d4cbb" # <-- any valid patient UUID
def main():
out_dir = Path(f"capture_patient_{PATIENT_ID}")
out_dir.mkdir(exist_ok=True)
html_path = out_dir / "detail.html"
screenshot_path = out_dir / "detail.png"
with sync_playwright() as p:
browser = p.chromium.launch(headless=True) # set False if you want to watch
context = browser.new_context(storage_state=STATE_FILE)
page = context.new_page()
# Open the detail directly
target_url = f"{BASE_URL}?pacient={PATIENT_ID}"
page.goto(target_url, wait_until="domcontentloaded")
# Wait a bit for the detail drawer/dialog to render
try:
page.wait_for_selector("[role='dialog'], div.MuiDrawer-paper, div[aria-modal='true']", timeout=10000)
except:
print("Warning: did not detect a detail panel quickly")
# Save raw HTML and screenshot
html_path.write_text(page.content(), encoding="utf-8")
page.screenshot(path=str(screenshot_path), full_page=True)
browser.close()
print("Saved:")
print(" -", html_path.resolve())
print(" -", screenshot_path.resolve())
if __name__ == "__main__":
main()

111
Medevio3CapturePatients.py Normal file
View File

@@ -0,0 +1,111 @@
from playwright.sync_api import sync_playwright
import mysql.connector
import time
MYSQL_CFG = dict(
host="192.168.1.76",
port=3307,
user="root",
password="Vlado9674+",
database="medevio",
)
# --- load 3 patients from DB ---
conn = mysql.connector.connect(**MYSQL_CFG)
with conn.cursor() as cur:
cur.execute("""
SELECT rid, prijmeni, jmeno, rc
FROM patients_extracted
WHERE prijmeni IS NOT NULL
ORDER BY prijmeni ASC
LIMIT 3
""")
rows = cur.fetchall()
if not rows:
raise RuntimeError("No entries found in patients_extracted")
STATE_FILE = r"medevio_storage.json"
BASE_URL = "https://my.medevio.cz/mudr-buzalkova/klinika/pacienti"
NOT_FOUND_SEL = "div[role='alert']:has-text('Pacient nebyl nalezen'), div:has-text('Pacient nebyl nalezen')"
DIALOG_SEL = "[role='dialog'], div.MuiDrawer-paper, div[aria-modal='true']"
def close_dialog_if_open(page):
dlg = page.locator(DIALOG_SEL)
try:
if dlg.count():
# Try a close button; if not, press Escape
try:
dlg.locator("button:has-text('Zavřít'), [aria-label='Zavřít'], [aria-label='Close'], [data-testid='CloseIcon']").first.click(timeout=1000)
except:
page.keyboard.press("Escape")
page.wait_for_selector(DIALOG_SEL, state="detached", timeout=1500)
except:
pass # best-effort close
def main():
with sync_playwright() as p:
browser = p.chromium.launch(headless=False)
try:
context = browser.new_context(storage_state=STATE_FILE)
page = context.new_page()
for rid, surname, name, rc in rows:
# 0) close any previous dialog to avoid stale matches
close_dialog_if_open(page)
target_url = f"{BASE_URL}?pacient={rid}"
page.goto(target_url, wait_until="domcontentloaded")
page.wait_for_load_state("networkidle")
# 1) Not-found toast?
try:
page.wait_for_selector(NOT_FOUND_SEL, timeout=3000)
print(f"{surname} {name} {rc} ⚠️ pacient s RID {rid} nebyl nalezen, přeskočeno")
# (optional) set mamedevioucet=NULL for this rid here
continue
except:
pass
# 2) Detail panel
try:
page.wait_for_selector(DIALOG_SEL, timeout=6000)
except:
print(f"⚠️ {surname} {name} {rc}: detailový panel se nenačetl, přeskočeno")
continue
# 3) Verify dialog belongs to current patient (avoid stale dialog)
detail = page.locator(DIALOG_SEL).first
detail_text = detail.inner_text()
if (surname not in detail_text) and (rc not in detail_text):
# Still looks wrong; give UI a moment and re-check once
page.wait_for_timeout(500)
detail_text = detail.inner_text()
if (surname not in detail_text) and (rc not in detail_text):
print(f"⚠️ {surname} {name} {rc}: detail neodpovídá (stará karta?), přeskočeno")
continue
# 4) Check Medevio account text
if "zatím nemá Medevio účet" in detail_text:
has_account = 0
print(f"{surname} {name} {rc} ❌ zatím nemá Medevio účet")
else:
has_account = 1
print(f"{surname} {name} {rc} ✅ má Medevio účet")
# Update DB by RID (or swap to rc if you prefer)
with conn.cursor() as c:
c.execute(
"UPDATE patients_extracted SET mamedevioucet = %s WHERE rid = %s",
(has_account, rid),
)
conn.commit()
time.sleep(0.5) # gentle pacing
finally:
browser.close()
if __name__ == "__main__":
main()

View File

@@ -1,10 +1,8 @@
import mysql.connector import mysql.connector
from bs4 import BeautifulSoup from bs4 import BeautifulSoup
import re import re
import time
# ---------- CONFIG ----------
# MySQL connection settings (fill in)
MYSQL_CFG = dict( MYSQL_CFG = dict(
host="192.168.1.76", host="192.168.1.76",
port=3307, port=3307,
@@ -13,30 +11,100 @@ MYSQL_CFG = dict(
database="medevio", database="medevio",
) )
#Helper functions
def is_valid_rc(rc: str) -> bool:
"""
Very basic RC check:
remove any slash
must be 9 or 10 digits
"""
rc_clean = rc.replace("/", "")
return bool(re.fullmatch(r"\d{9,10}", rc_clean))
conn = mysql.connector.connect(**MYSQL_CFG) conn = mysql.connector.connect(**MYSQL_CFG)
cur=conn.cursor()
cur.execute("select html from kartoteka_html where 'fetched-at'=(SELECT MAX('fetched-at') FROM kartoteka_html)")
html=cur.fetchone()
html=html[0]
# --- get latest HTML (single-row result) ---
with conn.cursor() as cur:
cur.execute("""
SELECT html
FROM kartoteka_html
where round=3
ORDER BY `fetched-at` DESC
""")
rows = cur.fetchall()
if not rows:
raise RuntimeError("No HTML found in kartoteka_html")
for row in rows:
html = row[0]
# html is the string containing the entire web page
soup = BeautifulSoup(html, "html.parser") soup = BeautifulSoup(html, "html.parser")
# Find every <button> that has that specific class sequence records = []
# (space-separated class names → match as a set) for row in soup.find_all("div", attrs={"role": "row", "data-id": True}):
buttons = soup.find_all( data_id = row["data-id"]
"button",
class_="MuiTypography-root MuiTypography-body2 "
"MuiLink-root MuiLink-underlineAlways "
"MuiLink-button css-xf7pf8"
)
names = []
for btn in buttons:
text = btn.get_text(strip=True)
print(text)
names.append(text)
print(names) # full name -> surname + rest
# names = [btn.get_text(strip=True) for btn in buttons] name_btn = row.find("button", class_="MuiTypography-root")
print(names) fullname = name_btn.get_text(strip=True) if name_btn else ""
parts = fullname.split()
surname = parts[0] if parts else ""
name = " ".join(parts[1:]) if len(parts) > 1 else ""
# RC
id_cell = row.find("div", attrs={"data-field": "IdentificationNumber"})
rc = (id_cell.get("title", "") if id_cell else "")
rc = rc.replace("/", "").replace("\\", "")
# Phone
ph_cell = row.find("div", attrs={"data-field": "Phone"})
raw_phone = ph_cell.get("title", "") if ph_cell else ""
raw_phone = raw_phone.replace("\u00A0", " ") # NBSP -> space
phone = re.sub(r"[^\d+]", "", raw_phone) # keep + and digits
# Insurance
ins_cell = row.find("div", attrs={"data-field": "InsuranceCompany"})
poj = ins_cell.get("title", "") if ins_cell else ""
# Skip rows with no name or no RC or not valid TC
if not fullname or not rc:
continue
if not is_valid_rc(rc):
continue
records.append((data_id, fullname, rc, phone, poj))
# --- per-patient lookup: use a fresh cursor each time (or buffered=True) ---
with conn.cursor(buffered=True) as cur2:
cur2.execute(
"""
SELECT *
FROM patients_extracted
WHERE rc=%s
""",
(rc,),
)
rows = cur2.fetchall()
# print(surname, name, rc, len(rows))
if len(rows) > 1:
print(f"Pacient {surname} {name} {rc} je v medeviu {len(rows)}x")
time.sleep(1)
if len(rows)==0:
print(f"Pacient {surname} {name} {rc} je v medeviu {len(rows)}x")
time.sleep(1)
if len(rows)==1 and rows[0][0]!=data_id:
print(f"Pacient {surname} {name} {rc} má v medeviu jiný id, v db je {rows[0][0]} and nyní je {data_id}")
time.sleep(.1)
if len(rows) == 1:
cur2.execute("""
Update patients_extracted set rid=%s where rc=%s""",(data_id,rc))
conn.commit()
# preview
# for r in records[:10]:
# print(f"ID: {r[0]} Name: {r[1]} RC: {r[2]} Phone: {r[3]} Pojistovna: {r[4]}")
#
# print("Total patients:", len(records))

View File

File diff suppressed because one or more lines are too long

Binary file not shown.

After

Width:  |  Height:  |  Size: 34 KiB