feat(f8): CJK-safe search text normalization + matching

normalizeForIndex (lowercase + collapse whitespace runs incl. hard PDF/OCR
newlines + trim) and matchesNormalized so a query matches across the line breaks
in raw extracted text. Deliberately NO word-tokenization: Chinese has no
inter-word spaces, so substring match over normalized text is correct for both
Latin and CJK (段/word segmentation belongs in the DB FTS tokenizer). Verified
on CJK inputs (你好/笔记应用).

flutter analyze lib/editor clean; 181/181 tests (+9).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

lib/editor/search/search_text.dart

@@ -0,0 +1,30 @@
+// lib/editor/search/search_text.dart
+//
+// Pure text normalization + matching for full-text search (F8). PDF text layers
+// and OCR output are full of hard line breaks and irregular whitespace, so a
+// query like "hello world" won't substring-match raw extracted text that reads
+// "hello\nworld". Normalizing both sides (lowercase + collapse every whitespace
+// run to a single space + trim) fixes that.
+//
+// CJK NOTE: this user writes Chinese. We deliberately do NOT word-tokenize —
+// Chinese has no inter-word spaces, so a whitespace/punctuation tokenizer would
+// mangle it. Substring matching over normalized text is correct for both Latin
+// and CJK; word/段 segmentation belongs in the DB FTS tokenizer (trigram /
+// unicode61), not here.
+
+/// Matches any run of Unicode whitespace (spaces, tabs, newlines, NBSP, …).
+final RegExp _whitespaceRun = RegExp(r'\s+');
+
+/// Normalize [text] for indexing/matching: lowercase, collapse whitespace runs
+/// (incl. the hard newlines PDF/OCR insert mid-sentence) to single spaces, trim.
+String normalizeForIndex(String text) {
+  return text.toLowerCase().replaceAll(_whitespaceRun, ' ').trim();
+}
+
+/// Whether [source] contains [query] after both are normalized — so a match can
+/// span the line breaks present in the raw text. Empty query never matches.
+bool matchesNormalized(String source, String query) {
+  final q = normalizeForIndex(query);
+  if (q.isEmpty) return false;
+  return normalizeForIndex(source).contains(q);
+}