# retoor from devplacepy.services.jobs.deepsearch.extract import extract_html, relevant_links PAGE = """ Framework Trends - Blog

Trends that define web development

Server components became the default rendering model, cutting client bundles by forty percent according to the survey.

Signals-based reactivity landed in the standards pipeline; see the signals deep dive article.

""" def test_extract_prefers_article_content_and_drops_chrome(): page = extract_html(PAGE, base_url="https://blog.example.com/trends/") assert page.title == "Framework Trends - Blog" assert "Server components" in page.text assert "Signals-based reactivity" in page.text assert "cookies" not in page.text assert "Copyright" not in page.text assert "Home" not in page.text def test_extract_emits_blank_line_paragraphs(): page = extract_html(PAGE, base_url="https://blog.example.com/trends/") assert "\n\n" in page.text def test_extract_resolves_links_absolute_and_skips_nav(): page = extract_html(PAGE, base_url="https://blog.example.com/trends/") urls = [url for url, _text in page.links] assert "https://blog.example.com/signals-deep-dive" in urls assert "https://blog.example.com/blog" not in urls def test_extract_unescapes_entities(): page = extract_html( "

Ampersand & arrow → and more text to pass the length gate.

" ) assert "&" not in page.text assert "→" not in page.text assert "&" in page.text def test_extract_survives_malformed_html(): page = extract_html("

Unclosed paragraph with enough characters to be kept around.

") assert "Unclosed paragraph" in page.text def test_relevant_links_scores_by_query_overlap(): links = [ ("https://a.example/web-frameworks-2026", "web frameworks in 2026"), ("https://a.example/cookie-policy", "cookie policy"), ("https://a.example/logo.png", "frameworks logo"), ] picked = relevant_links(links, "latest web frameworks 2026", 2) assert picked == ["https://a.example/web-frameworks-2026"] def test_relevant_links_empty_query_returns_nothing(): assert relevant_links([("https://a.example/x", "text")], "", 3) == []