Extract title tags & per-row chevron expand

Add parsing and UX for bracketed title metadata and per-row title expansion. Introduces extractTitleTags (src/lib/utils/title-tags.ts) to pull bracketed tags from result titles (de-duplicated, slash-split) and useIsTruncated (src/lib/hooks/useIsTruncated.ts) to detect horizontal overflow. Refactors InteractiveTorrentSearchModal to a ResultRow component that renders title chips (slate chips) for parsed tags (filtered vs displayFormat), shows a chevron disclosure only when the title is truncated (or while expanded), toggles expansion per GUID, and resets expansion state when the modal closes. Tests added/updated for the component, hook, and parser; documentation updated to reflect behavior.

src/lib/hooks/useIsTruncated.ts

@@ -0,0 +1,34 @@
+/**
+ * Component: useIsTruncated Hook
+ * Documentation: documentation/frontend/components.md
+ *
+ * Returns whether the referenced element's content overflows horizontally
+ * (i.e. is being clipped by `truncate` / `overflow: hidden`). Used by the
+ * Interactive Search modal to render an expand-disclosure chevron only when
+ * the title is actually being cut off — keeping the row clean when there's
+ * nothing to disclose.
+ */
+
+import { useLayoutEffect, useState, type RefObject } from 'react';
+
+export function useIsTruncated(ref: RefObject<HTMLElement | null>): boolean {
+  const [isTruncated, setIsTruncated] = useState(false);
+
+  useLayoutEffect(() => {
+    const el = ref.current;
+    if (!el) return;
+
+    const measure = () => {
+      setIsTruncated(el.scrollWidth > el.clientWidth);
+    };
+
+    measure();
+
+    if (typeof ResizeObserver === 'undefined') return;
+    const observer = new ResizeObserver(measure);
+    observer.observe(el);
+    return () => observer.disconnect();
+  }, [ref]);
+
+  return isTruncated;
+}

src/lib/utils/title-tags.ts

@@ -0,0 +1,41 @@
+/**
+ * Component: Title Tag Extraction Utility
+ * Documentation: documentation/frontend/components.md
+ *
+ * Pure parser used by the Interactive Search modal to split a result title
+ * into a "residual" string and a list of bracketed metadata tags. Brackets
+ * are ASCII `[`/`]` only (full-width `【】` is intentionally unsupported —
+ * Audible/indexer titles use ASCII in practice). Inner content is split on
+ * `/`, trimmed, empty segments dropped, then de-duplicated case-insensitively
+ * while preserving first-seen casing. The regex is non-nested by design:
+ * `Foundation [Edition [Deluxe]]` extracts `Deluxe` and leaves the outer
+ * `[Edition ]` in the residual. Accepted trade-off for v1 (rare).
+ */
+export interface TitleTags {
+  residual: string;
+  tags: string[];
+}
+
+// Why: character class excludes brackets so the regex never spans a nested
+// pair — this is what makes the inner `[Deluxe]` win over the outer group.
+const BRACKET_GROUP = /\[([^\[\]]*)\]/g;
+
+export function extractTitleTags(title: string): TitleTags {
+  if (!title) return { residual: '', tags: [] };
+
+  const tags: string[] = [];
+  const seen = new Set<string>();
+  for (const match of title.matchAll(BRACKET_GROUP)) {
+    for (const segment of match[1].split('/')) {
+      const trimmed = segment.trim();
+      if (!trimmed) continue;
+      const key = trimmed.toLowerCase();
+      if (seen.has(key)) continue;
+      seen.add(key);
+      tags.push(trimmed);
+    }
+  }
+
+  const residual = title.replace(BRACKET_GROUP, ' ').replace(/\s+/g, ' ').trim();
+  return { residual, tags };
+}