dev.altai / debug_crawler.py
prince1604
Enhanced alt text analysis: added detection for short, sparse, generic, and filename-like alt text
4ec2569
Raw
History Blame Contribute Delete
602 Bytes
from src.crawler import Crawler
from src.analyzer import ImageAnalyzer
import logging
import json
# Configure logging to see what's happening
logging.basicConfig(level=logging.INFO)
c = Crawler()
analyzer = ImageAnalyzer()
domain = "https://eminentcoders.com/"
print(f"DEBUG: Starting crawl for {domain}")
site_data, discovered, reason = c.crawl_domain(domain, max_pages=1)
print(f"DEBUG: Result - Pages Scanned: {len(site_data)}")
print(f"DEBUG: Discovered: {discovered}")
print(f"DEBUG: Blocked Reason: {reason}")
report = analyzer.analyze_site(site_data)
print(json.dumps(report, indent=2))