aron/wohnbot
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

gewobag fix

Browse source
This commit is contained in:
Aron Petau 2026-01-05 13:40:12 +01:00
parent cc40121e46
commit 4ea437e3e6
4 changed files with 312 additions and 124 deletions
Download patch file Download diff file Expand all files Collapse all files

77
handlers/gesobau_handler.py
View file

@ -1,6 +1,7 @@
from .base_handler import BaseHandler
import logging
import asyncio
import os
from pathlib import Path
logger = logging.getLogger(__name__)
@ -22,33 +23,47 @@ class GesobauHandler(BaseHandler):
# 404 detection
status = response.status if response else None
page_title = await page.title()
if status == 404 or (page_title and "404" in page_title):
logger.warning(f"[GESOBAU] Listing is down (404): {listing['link']}")
page_content = await page.content()
is_404 = (
status == 404 or
(page_title and "404" in page_title) or
(page_title and "nicht gefunden" in page_title.lower()) or
("Angebot nicht mehr verfügbar" in page_content)
)
if is_404:
logger.warning(f"[GESOBAU] Listing is down (404 or unavailable): {listing['link']}")
result["success"] = False
result["message"] = "Listing is no longer available (404). Application impossible. Will not retry."
result["deactivated"] = True
await page.close()
return result
# Always handle cookies and consent before anything else
await self.handle_cookies(page)
await self.handle_consent(page)
# Save HTML after modal handling for debugging and check for deactivation
html_content = await page.content()
# Dismiss cookie banner
try:
with open(DATA_DIR / "gesobau_debug.html", "w", encoding="utf-8") as f:
f.write(html_content)
except Exception as e:
logger.debug(f"[GESOBAU] Debug HTML not saved: {e}")
cookie_btn = await page.query_selector('#CybotCookiebotDialogBodyLevelButtonLevelOptinAllowAll, button:has-text("Alle akzeptieren")')
if cookie_btn and await cookie_btn.is_visible():
await cookie_btn.click()
logger.info("[GESOBAU] Dismissed cookie banner")
await asyncio.sleep(1)
except:
pass
# Tailored 404 detection: Angebot nicht mehr verfügbar
if "Angebot nicht mehr verfügbar" in html_content:
logger.warning("[GESOBAU] Permanent fail: Angebot nicht mehr verfügbar")
result["deactivated"] = True
result["message"] = "Listing is no longer available (Angebot nicht mehr verfügbar). Marked as deactivated."
await page.close()
return result
# Save debug HTML and screenshot
try:
html_content = await page.content()
with open(DATA_DIR / f"gesobau_debug_{listing['id']}.html", "w", encoding="utf-8") as f:
f.write(html_content)
logger.info(f"[GESOBAU] Saved debug HTML: gesobau_debug_{listing['id']}.html")
except Exception as e:
logger.warning(f"[GESOBAU] Could not save debug HTML: {e}")
try:
await page.screenshot(path=DATA_DIR / f"gesobau_page_{listing['id']}.png", full_page=True)
logger.info(f"[GESOBAU] Saved page screenshot: gesobau_page_{listing['id']}.png")
except Exception as e:
logger.warning(f"[GESOBAU] Could not save screenshot: {e}")
# Log listing details
await self.log_listing_details(listing)
# Look for application button
logger.info("[GESOBAU] Searching for application button...")
@ -56,6 +71,8 @@ class GesobauHandler(BaseHandler):
'a[href*="bewerben"]',
'button:has-text("Bewerben")',
'a:has-text("Bewerben")',
'button:has-text("Interesse")',
'a:has-text("Kontakt")',
'button.btn',
]
@ -77,22 +94,25 @@ class GesobauHandler(BaseHandler):
if apply_btn:
await apply_btn.scroll_into_view_if_needed()
await asyncio.sleep(0.5)
logger.info("[GESOBAU] Clicking application button...")
await apply_btn.click()
await asyncio.sleep(2)
# --- Post-click confirmation logic ---
logger.info("[GESOBAU] Clicked application button, checking for confirmation...")
# Save screenshot and HTML after click
logger.info("[GESOBAU] Checking for confirmation...")
try:
await page.screenshot(path="data/gesobau_after_apply.png")
logger.info("[GESOBAU] Saved screenshot after application click.")
await page.screenshot(path=DATA_DIR / f"gesobau_after_apply_{listing['id']}.png", full_page=True)
logger.info(f"[GESOBAU] Saved after-apply screenshot: gesobau_after_apply_{listing['id']}.png")
except Exception as e:
logger.warning(f"[GESOBAU] Could not save screenshot: {e}")
logger.warning(f"[GESOBAU] Could not save after-apply screenshot: {e}")
try:
html_after = await page.content()
with open("data/gesobau_after_apply.html", "w", encoding="utf-8") as f:
with open(DATA_DIR / f"gesobau_after_apply_{listing['id']}.html", "w", encoding="utf-8") as f:
f.write(html_after)
logger.info("[GESOBAU] Saved HTML after application click.")
logger.info(f"[GESOBAU] Saved after-apply HTML: gesobau_after_apply_{listing['id']}.html")
except Exception as e:
logger.warning(f"[GESOBAU] Could not save after-apply HTML: {e}")
logger.warning(f"[GESOBAU] Could not save HTML after apply: {e}")
# Look for confirmation message on the page
@ -123,6 +143,9 @@ class GesobauHandler(BaseHandler):
else:
logger.warning("[GESOBAU] No application button found.")
result["message"] = "No application button found."
screenshot_path = DATA_DIR / f"gesobau_nobtn_{listing['id']}.png"
await page.screenshot(path=str(screenshot_path))
logger.info(f"[GESOBAU] Saved no-button screenshot: {screenshot_path}")
except Exception as e:
result["message"] = f"Error during application: {e}"
logger.error(f"[GESOBAU] Application error: {e}")