gbalabanli/protected-site-scraper

Protected Site Scraper

Overview

This skill provides tools to scrape websites that use anti-bot protection (Cloudflare, DataDome, PerimeterX, etc.) using undetected-chromedriver - a specialized Selenium wrapper designed to bypass detection.

Perfect for:

• Real estate sites (sahibinden.com, emlakjet, etc.)
• Classified listings
• E-commerce with bot protection
• Any site showing "Just a moment..." or CAPTCHA challenges

When to Use

Use this skill when:

1. Standard requests/httpx fail with 403 errors
2. The site shows Cloudflare challenge pages
3. Selenium without undetected-chromedriver gets blocked
4. You need to scrape JavaScript-heavy protected sites

Quick Start

# Install dependencies
pip install undetected-chromedriver selenium

# Run the sahibinden example
python examples/sahibinden_rentals.py --location istanbul-maltepe-altintepe

Usage Patterns

Pattern 1: Simple Protected Site Scraper

from protected_site_scraper import ProtectedSiteScraper

scraper = ProtectedSiteScraper()
driver = scraper.setup_driver()

# Navigate and scrape
driver.get("https://example.com/listings")
listings = scraper.extract_listings(driver)

scraper.save_results(listings, "output.json")
driver.quit()

Pattern 2: Custom Extraction

from protected_site_scraper import ProtectedSiteScraper
from selenium.webdriver.common.by import By

scraper = ProtectedSiteScraper()
driver = scraper.setup_driver()

driver.get("https://example.com")

# Wait for Cloudflare
scraper.wait_for_protection_to_clear(driver)

# Custom extraction
elements = driver.find_elements(By.CSS_SELECTOR, ".listing")
data = []
for elem in elements:
    data.append({
        'title': elem.find_element(By.CSS_SELECTOR, '.title').text,
        'price': elem.find_element(By.CSS_SELECTOR, '.price').text,
    })

scraper.save_results(data, "output.json")
driver.quit()

Pattern 3: Sahibinden.com Specific

from examples.sahibinden_rentals import SahibindenScraper

scraper = SahibindenScraper()
listings = scraper.scrape_rentals(
    city="istanbul",
    district="maltepe", 
    neighborhood="altintepe"
)

print(f"Found {len(listings)} listings")
scraper.save_to_csv(listings, "rentals.csv")

Core Components

1. ProtectedSiteScraper (Base Class)

Location: protected_site_scraper/core.py

Key Methods:

• setup_driver(headless=False) - Initialize undetected Chrome
• wait_for_protection_to_clear(driver, timeout=30) - Wait for Cloudflare/anti-bot
• safe_find_elements(driver, selectors) - Try multiple selectors
• extract_with_fallback(element, selectors) - Extract text with fallback
• save_results(data, filename) - Save to JSON/CSV
• safe_print(text) - Handle Turkish/special characters in console

2. Site-Specific Scrapers

Location: examples/

• sahibinden_rentals.py - Complete working example for sahibinden.com
• template_scraper.py - Template for new sites

3. Prompts

Location: prompts/

• find_rentals.md - Prompt to find rentals on sahibinden
• scrape_generic.md - Generic scraping prompt template

Configuration

Environment Variables

# Optional: Set Chrome binary path if not default
CHROME_BINARY_PATH="C:\Program Files\Google\Chrome\Application\chrome.exe"

# Optional: Set specific Chrome version
CHROME_VERSION="120"

Driver Options

from protected_site_scraper import ProtectedSiteScraper

scraper = ProtectedSiteScraper()

# Custom options
driver = scraper.setup_driver(
    headless=False,  # Show browser window
    window_size=(1920, 1080),
    user_agent="Custom User Agent",
    disable_images=True,  # Faster loading
    proxy="http://proxy:8080"  # Use proxy
)

Advanced Features

Handling Pagination

def scrape_all_pages(scraper, driver, base_url):
    all_listings = []
    page = 1
    
    while True:
        url = f"{base_url}?page={page}"
        driver.get(url)
        scraper.wait_for_protection_to_clear(driver)
        
        listings = scraper.extract_listings(driver)
        if not listings:
            break
            
        all_listings.extend(listings)
        page += 1
        
        # Check for next button
        try:
            next_btn = driver.find_element(By.CSS_SELECTOR, ".next")
            if not next_btn.is_enabled():
                break
        except:
            break
    
    return all_listings

Custom Field Extraction

def extract_custom_fields(element):
    """Define custom extraction logic"""
    fields = {}
    
    # Try multiple selectors for each field
    field_selectors = {
        'title': ['.title', 'h1', 'h2', '.listing-title'],
        'price': ['.price', '.cost', '[data-price]'],
        'location': ['.location', '.address', '.place'],
        'rooms': ['.rooms', '.oda', '[data-rooms]'],
        'area': ['.area', '.square-meter', '[data-area]'],
    }
    
    for field, selectors in field_selectors.items():
        fields[field] = scraper.extract_with_fallback(element, selectors)
    
    return fields

Error Recovery

from selenium.common.exceptions import TimeoutException, WebDriverException

def scrape_with_retry(url, max_retries=3):
    for attempt in range(max_retries):
        try:
            driver.get(url)
            scraper.wait_for_protection_to_clear(driver, timeout=30)
            return scraper.extract_listings(driver)
        except TimeoutException:
            print(f"Attempt {attempt + 1} failed, retrying...")
            time.sleep(5)
        except WebDriverException as e:
            print(f"Browser error: {e}")
            # Restart driver
            driver.quit()
            driver = scraper.setup_driver()
    
    return []

Anti-Detection Tips

1. Don't use headless mode - It triggers more detection
2. Add realistic delays - Use time.sleep() between actions
3. Use realistic User-Agent - The scraper handles this automatically
4. Avoid rapid requests - Add delays between page loads
5. Rotate proxies - For high-volume scraping
6. Handle cookies - Accept or manage cookies appropriately

Troubleshooting

"Session not created" Error

# Chrome version mismatch - specify version
driver = uc.Chrome(options=options, version_main=120)

Cloudflare Challenge Not Passing

# Increase wait time
scraper.wait_for_protection_to_clear(driver, timeout=60)

# Or check for specific elements
WebDriverWait(driver, 60).until(
    EC.presence_of_element_located((By.CSS_SELECTOR, ".content"))
)

Elements Not Found

# Try multiple selectors
selectors = [
    "#searchResultsTable",
    ".listings",
    "[data-listings]"
]

for selector in selectors:
    elements = driver.find_elements(By.CSS_SELECTOR, selector)
    if elements:
        break

Examples

Sahibinden.com - Rental Listings

python examples/sahibinden_rentals.py \
    --city istanbul \
    --district maltepe \
    --neighborhood altintepe \
    --output rentals.json

Generic Site Template

# examples/template_scraper.py
from protected_site_scraper import ProtectedSiteScraper

class MySiteScraper(ProtectedSiteScraper):
    def scrape(self, url):
        driver = self.setup_driver()
        driver.get(url)
        
        self.wait_for_protection_to_clear(driver)
        
        # Your extraction logic here
        data = self.extract_listings(driver)
        
        driver.quit()
        return data

if __name__ == "__main__":
    scraper = MySiteScraper()
    results = scraper.scrape("https://example.com/listings")
    scraper.save_results(results, "output.json")

Best Practices

1. Always quit driver - Use try/finally or context managers
2. Save incrementally - Don't lose data on crashes
3. Respect robots.txt - Check if scraping is allowed
4. Rate limiting - Don't overwhelm servers
5. Legal compliance - Check website terms of service

Dependencies

undetected-chromedriver>=3.5.0
selenium>=4.15.0
webdriver-manager>=4.0.0

License

MIT - Use responsibly and comply with website terms of service.