已完成用户APP首期功能开发

交付用户端首期页面、配套接口、后台资源及测试文档。用户APP构建、静态分析和三个管理后台构建通过；完整测试仍有2项失败，后端模型注释检查未通过，详见交付记录。

apps/user_app/lib/data/services/repair_speech.dart

@@ -0,0 +1,77 @@
+// 功能描述：故障描述语音识别适配，应用只复用一个系统识别实例，不保存录音。
+// 版本：1.0.0。
+import 'package:speech_to_text/speech_to_text.dart';
+
+/// 识别结果为本轮完整短句；调用方负责替换临时结果，不能逐次追加。
+abstract interface class RepairSpeech {
+  Future<bool> start({
+    required void Function(String) onWords,
+    required void Function(String) onError,
+    required void Function() onDone,
+  });
+  Future<void> stop();
+  Future<void> cancel();
+}
+
+/// 系统回调只初始化一次，每轮重新绑定当前界面，离页后忽略迟到结果。
+class SystemRepairSpeech implements RepairSpeech {
+  SystemRepairSpeech._();
+  static final SystemRepairSpeech instance = SystemRepairSpeech._();
+  final SpeechToText _speech = SpeechToText();
+  void Function(String)? _words, _error;
+  void Function()? _done;
+  int _generation = 0;
+  @override
+  Future<bool> start({
+    required void Function(String) onWords,
+    required void Function(String) onError,
+    required void Function() onDone,
+  }) async {
+    final generation = ++_generation;
+    _words = onWords;
+    _error = onError;
+    _done = onDone;
+    final available = await _speech.initialize(
+      onError: (error) => _error?.call(error.errorMsg),
+      onStatus: (status) {
+        if (status == SpeechToText.doneStatus) _done?.call();
+      },
+      options: [SpeechToText.androidNoBluetooth],
+    );
+    if (generation != _generation || !available) return false;
+    final locales = await _speech.locales();
+    if (generation != _generation) return false;
+    String? locale;
+    for (final item in locales) {
+      if (item.localeId.toLowerCase().replaceAll('_', '-') == 'zh-cn') {
+        locale = item.localeId;
+        break;
+      }
+    }
+    await _speech.listen(
+      onResult: (result) {
+        if (generation == _generation) _words?.call(result.recognizedWords);
+      },
+      listenOptions: SpeechListenOptions(
+        localeId: locale,
+        partialResults: true,
+        cancelOnError: true,
+        listenMode: ListenMode.dictation,
+        listenFor: const Duration(seconds: 45),
+        pauseFor: const Duration(seconds: 5),
+      ),
+    );
+    return generation == _generation;
+  }
+
+  @override
+  Future<void> stop() => _speech.stop();
+  @override
+  Future<void> cancel() async {
+    ++_generation;
+    _words = null;
+    _error = null;
+    _done = null;
+    await _speech.cancel();
+  }
+}