feat: 优化口语防重复问答、课程推进流转、键盘输入流畅度与竖屏锁定

- AI口语约束：在对话系统提示词与单轮指导中增加防重复问答约束，避免重复寒暄及索取已提供信息，根据历史自然向前推进
- 课程流转：修复多段课程学完后今日任务卡片流转下一课，及退出重进定位逻辑
- 竖屏锁定：在 Flutter、Android 及 iOS 平台配置仅支持竖屏显示
- 对话交互：优化对话界面文本输入重绘与词汇查询，解决键盘输入卡顿
- 测试用例：补充并更新单轮约束、状态流转与返回键退出测试，212项测试全绿

kouyu_english/lib/widgets/lexicon_lookup.dart

@@ -9,15 +9,59 @@ import '../core/courses/courses.dart';
 import '../core/voice_service.dart';
 import 'app_widgets.dart';
 
+List<VocabularyItem>? _cachedEntries;
+Map<String, VocabularyItem>? _cachedExactMap;
+RegExp? _cachedExpression;
+
+/// Clears the lexicon cache when courses are reloaded or reset.
+void clearLexiconCache() {
+  _cachedEntries = null;
+  _cachedExactMap = null;
+  _cachedExpression = null;
+}
+
 List<VocabularyItem> get courseLexiconEntries {
+  if (_cachedEntries != null) return _cachedEntries!;
   // Segment words first: they carry the example and IPA a lesson shows.
   final entries = <VocabularyItem>[
     ...segmentVocabulary.values.expand((items) => items),
     ...allLessons.expand((lesson) => lesson.vocabulary),
   ];
   final seen = <String>{};
-  return entries.where((item) => seen.add(item.word.toLowerCase())).toList()
+  _cachedEntries = entries.where((item) => seen.add(item.word.toLowerCase())).toList()
     ..sort((a, b) => b.word.length.compareTo(a.word.length));
+  return _cachedEntries!;
+}
+
+Map<String, VocabularyItem> get _exactLexiconMap {
+  if (_cachedExactMap != null) return _cachedExactMap!;
+  final map = <String, VocabularyItem>{};
+  for (final item in courseLexiconEntries) {
+    final normalized = item.word
+        .toLowerCase()
+        .replaceAll('’', "'")
+        .replaceAll('‘', "'")
+        .replaceAll(RegExp(r'\s+'), ' ')
+        .trim();
+    map.putIfAbsent(normalized, () => item);
+  }
+  _cachedExactMap = map;
+  return map;
+}
+
+RegExp? get courseLexiconExpression {
+  if (_cachedExpression != null) return _cachedExpression;
+  final entries = courseLexiconEntries;
+  if (entries.isEmpty) return null;
+  _cachedExpression = RegExp(
+    r'(?<![a-zA-Z0-9])(?:' +
+        entries
+            .map((item) => RegExp.escape(item.word).replaceAll('’', "['’]"))
+            .join('|') +
+        r')(?![a-zA-Z0-9])',
+    caseSensitive: false,
+  );
+  return _cachedExpression;
 }
 
 /// Finds a course item by exact query, then by the longest known phrase in it.
@@ -29,20 +73,28 @@ VocabularyItem? findCourseLexicon(String text) {
       .replaceAll(RegExp(r'\s+'), ' ')
       .trim();
   if (normalized.isEmpty) return null;
-  return courseLexiconEntries.cast<VocabularyItem?>().firstWhere((entry) {
-    final word = entry!.word
+  final exact = _exactLexiconMap[normalized];
+  if (exact != null) return exact;
+
+  final entries = courseLexiconEntries;
+  for (var i = 0; i < entries.length; i++) {
+    final entry = entries[i];
+    final word = entry.word
         .toLowerCase()
         .replaceAll('’', "'")
         .replaceAll('‘', "'")
         .replaceAll(RegExp(r'\s+'), ' ')
         .trim();
-    if (normalized == word) return true;
-    final pattern = RegExp(
-      r'(?<![a-zA-Z0-9])' + RegExp.escape(word) + r'(?![a-zA-Z0-9])',
-      caseSensitive: false,
-    );
-    return pattern.hasMatch(normalized);
-  }, orElse: () => null);
+    if (normalized == word) return entry;
+    if (normalized.contains(word)) {
+      final pattern = RegExp(
+        r'(?<![a-zA-Z0-9])' + RegExp.escape(word) + r'(?![a-zA-Z0-9])',
+        caseSensitive: false,
+      );
+      if (pattern.hasMatch(normalized)) return entry;
+    }
+  }
+  return null;
 }
 
 /// Extracts all distinct course lexicon phrases/words that appear within [text].
@@ -92,6 +144,12 @@ bool isSentenceQuery(String text) {
       trimmed.length > 25;
 }
 
+class _LexiconChunk {
+  const _LexiconChunk({required this.text, this.entry});
+  final String text;
+  final VocabularyItem? entry;
+}
+
 /// Inline course text that lets a learner tap a known word or phrase without
 /// leaving the current task. Longest phrases are matched before their words.
 class LexiconText extends StatefulWidget {
@@ -117,6 +175,46 @@ class LexiconText extends StatefulWidget {
 class _LexiconTextState extends State<LexiconText> {
   final List<TapGestureRecognizer> _recognizers = [];
   String selectedText = '';
+  List<_LexiconChunk>? _chunks;
+
+  @override
+  void initState() {
+    super.initState();
+    _parseChunks();
+  }
+
+  @override
+  void didUpdateWidget(LexiconText oldWidget) {
+    super.didUpdateWidget(oldWidget);
+    if (oldWidget.text != widget.text) {
+      _parseChunks();
+    }
+  }
+
+  void _parseChunks() {
+    final expression = courseLexiconExpression;
+    if (expression == null) {
+      _chunks = [_LexiconChunk(text: widget.text)];
+      return;
+    }
+    final chunks = <_LexiconChunk>[];
+    var cursor = 0;
+    for (final match in expression.allMatches(widget.text)) {
+      if (match.start > cursor) {
+        chunks.add(
+          _LexiconChunk(text: widget.text.substring(cursor, match.start)),
+        );
+      }
+      final matched = widget.text.substring(match.start, match.end);
+      final entry = findCourseLexicon(matched);
+      chunks.add(_LexiconChunk(text: matched, entry: entry));
+      cursor = match.end;
+    }
+    if (cursor < widget.text.length) {
+      chunks.add(_LexiconChunk(text: widget.text.substring(cursor)));
+    }
+    _chunks = chunks;
+  }
 
   @override
   void dispose() {
@@ -132,26 +230,16 @@ class _LexiconTextState extends State<LexiconText> {
       recognizer.dispose();
     }
     _recognizers.clear();
-    final entries = courseLexiconEntries;
-    if (entries.isEmpty) return Text(widget.text, style: widget.style);
-    final expression = RegExp(
-      r'(?<![a-zA-Z0-9])(?:' +
-          entries
-              .map((item) => RegExp.escape(item.word).replaceAll('’', "['’]"))
-              .join('|') +
-          r')(?![a-zA-Z0-9])',
-      caseSensitive: false,
-    );
+
+    if (_chunks == null || _chunks!.isEmpty) {
+      return Text(widget.text, style: widget.style);
+    }
+
     final spans = <InlineSpan>[];
-    var cursor = 0;
-    for (final match in expression.allMatches(widget.text)) {
-      if (match.start > cursor) {
-        spans.add(TextSpan(text: widget.text.substring(cursor, match.start)));
-      }
-      final matched = widget.text.substring(match.start, match.end);
-      final entry = findCourseLexicon(matched);
+    for (final chunk in _chunks!) {
+      final entry = chunk.entry;
       if (entry == null) {
-        spans.add(TextSpan(text: matched));
+        spans.add(TextSpan(text: chunk.text));
       } else {
         final recognizer = TapGestureRecognizer()
           ..onTap = () => showLexiconLookup(
@@ -162,7 +250,7 @@ class _LexiconTextState extends State<LexiconText> {
         _recognizers.add(recognizer);
         spans.add(
           TextSpan(
-            text: matched,
+            text: chunk.text,
             recognizer: recognizer,
             style: TextStyle(
               color: AppColors.green,
@@ -172,10 +260,6 @@ class _LexiconTextState extends State<LexiconText> {
           ),
         );
       }
-      cursor = match.end;
-    }
-    if (cursor < widget.text.length) {
-      spans.add(TextSpan(text: widget.text.substring(cursor)));
     }
     return Column(
       crossAxisAlignment: CrossAxisAlignment.start,