Floke/Brancheneinstufung2
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

scrape_fotograf.py aktualisiert

Browse Source
This commit is contained in:
Floke
2025-07-16 19:07:44 +00:00
parent e790862664
commit ee6fe3982e
1 changed files with 7 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
scrape_fotograf.py
View File

@@ -14,10 +14,10 @@ from selenium.common.exceptions import TimeoutException, NoSuchElementException,
# --- Konfiguration & Konstanten --- # --- Konfiguration & Konstanten ---
CREDENTIALS_FILE = 'fotograf_credentials.json' CREDENTIALS_FILE = 'fotograf_credentials.json'
OUTPUT_DIR = 'output' OUTPUT_DIR = 'output'
OUTPUT_FILE = os.path.join(OUTPUT_DIR, 'nutzer_ohne_logins.csv') OUTPUT_FILE = os.path.join(OUTPUT_DIR, 'nutzer_mit_wenig_logins.csv') # Dateiname angepasst
LOGIN_URL = 'https://app.fotograf.de/login/login' LOGIN_URL = 'https://app.fotograf.de/login/login'
# --- Selektoren (FINALE, VOLLSTÄNDIGE VERSION) --- # --- Selektoren (unverändert) ---
SELECTORS = { SELECTORS = {
"cookie_accept_button": "#CybotCookiebotDialogBodyLevelButtonLevelOptinAllowAll", "cookie_accept_button": "#CybotCookiebotDialogBodyLevelButtonLevelOptinAllowAll",
"login_user": "#login-email", "login_user": "#login-email",
@@ -26,7 +26,6 @@ SELECTORS = {
"job_name": "h1", "job_name": "h1",
"album_overview_rows": "//table/tbody/tr", "album_overview_rows": "//table/tbody/tr",
"album_overview_link": ".//td[2]//a", "album_overview_link": ".//td[2]//a",
# NEU: Selector für die Gesamtzahl der Zugangscodes
"access_code_count": "//span[text()='Zugangscodes']/following-sibling::strong", "access_code_count": "//span[text()='Zugangscodes']/following-sibling::strong",
"person_rows": "//div[contains(@class, 'border-legacy-silver-550') and .//span[text()='Logins']]", "person_rows": "//div[contains(@class, 'border-legacy-silver-550') and .//span[text()='Logins']]",
"person_vorname": ".//span[text()='Vorname']/following-sibling::strong", "person_vorname": ".//span[text()='Vorname']/following-sibling::strong",
@@ -140,7 +139,6 @@ def process_full_job(driver, job_url):
print(f"\n--- Betrete Album: {album['name']} ---") print(f"\n--- Betrete Album: {album['name']} ---")
driver.get(album['url']) driver.get(album['url'])
try: try:
# NEU: Pagination-Logik
total_codes_text = wait.until(EC.visibility_of_element_located((By.XPATH, SELECTORS["access_code_count"]))).text total_codes_text = wait.until(EC.visibility_of_element_located((By.XPATH, SELECTORS["access_code_count"]))).text
num_pages = math.ceil(int(total_codes_text) / 20) num_pages = math.ceil(int(total_codes_text) / 20)
print(f"Album hat {total_codes_text} Zugangscodes auf {num_pages} Seite(n).") print(f"Album hat {total_codes_text} Zugangscodes auf {num_pages} Seite(n).")
@@ -161,9 +159,11 @@ def process_full_job(driver, job_url):
person_row = person_rows[i] person_row = person_rows[i]
login_count_text = person_row.find_element(By.XPATH, SELECTORS["person_logins"]).text login_count_text = person_row.find_element(By.XPATH, SELECTORS["person_logins"]).text
if int(login_count_text) == 0:
# --- HIER IST DIE GEÄNDERTE LOGIK ---
if int(login_count_text) <= 1:
vorname = person_row.find_element(By.XPATH, SELECTORS["person_vorname"]).text vorname = person_row.find_element(By.XPATH, SELECTORS["person_vorname"]).text
print(f" --> ERFOLG: '{vorname}' mit 0 Logins gefunden!") print(f" --> ERFOLG: '{vorname}' mit {login_count_text} Login(s) gefunden!")
access_code_page_url = person_row.find_element(By.XPATH, SELECTORS["person_access_code_link"]).get_attribute('href') access_code_page_url = person_row.find_element(By.XPATH, SELECTORS["person_access_code_link"]).get_attribute('href')
driver.get(access_code_page_url) driver.get(access_code_page_url)
@@ -240,7 +240,7 @@ def get_profile_choice():
except ValueError: print("Ungültige Eingabe.") except ValueError: print("Ungültige Eingabe.")
def main(): def main():
print("--- Fotograf.de Scraper für Nutzer ohne Logins (FINALE VERSION) ---") print("--- Fotograf.de Scraper für Nutzer mit wenig Logins (0 oder 1) ---")
credentials = get_profile_choice() credentials = get_profile_choice()
if not credentials: return if not credentials: return