Text Search II

Requirements

• Export highlightTerms(text, terms) as named and default. text is a plain string and terms is a readonly array of strings. Return HTML with each matched region wrapped in <mark>...</mark>. Do not edit the DOM or author an HTML/CSS file.
• Match terms literally, not as regular expressions. Ignore empty terms and duplicate terms. Matching is case-insensitive for ASCII A–Z only; other characters retain exact code-unit comparison. Preserve the original text’s case and character sequence in the output.
• Find all occurrences, including overlapping occurrences of a term. Combine overlapping or directly adjacent matches into one marked region. Output regions in text order regardless of term order. No nested or empty mark elements.
• Escape &, <, >, double quotes, and apostrophes everywhere in the original text to &amp;, &lt;, &gt;, &quot;, and &#39;. Add markup only after determining ranges on the unescaped input, so entity expansion does not shift offsets.
• Return the escaped original when no term matches or all terms are empty. Treat entity-looking text as literal characters, and never trust input text or terms as HTML. Indexes use UTF-16 offsets.
• Do not mutate terms. There is no fuzzy matching, stemming, normalization, locale-specific case folding, whole-word requirement, or cross-node searching.

Examples

Input: highlightTerms("banana", ["ana", "nan"]);
Output: "b<mark>anana</mark>"

Input: highlightTerms("<React & Vue>", ["react", "vue"]);
Output: "&lt;<mark>React</mark> &amp; <mark>Vue</mark>&gt;"