feat: 全功能前后端联调完成，47/47 测试通过

前端： - 新增 DietCapturePage 独立拍照识别页 - 5种消息卡片类型完整实现（数据确认/用药/饮食/报告/快捷选项） - 任务卡片区：异常警告+数据摘要+自动折叠 - 侧滑抽屉：历史对话列表+对话管理 - 运动计划：进度卡片+创建计划+每日打卡 - 报告页：拍照/相册/PDF上传+分析 - 面板按钮补全血氧/体重录入 - UI 升级：紫色主题+动画+气泡样式 - 全部迁移 Riverpod 3.x API 后端： - 新增 _UpdateMessageTypeAndMetadata，Tool Calling 自动映射消息类型 - SSE answer 事件携带 type 字段 - 提示词优化（移除"阿福"，语气规则归位） - 运动计划支持 AI 创建和打卡测试： - 新增 full_e2e_test.py 全流程测试（认证/数据CRUD/6个Agent对话/VLM/报告）
2026-06-02 20:31:22 +08:00
parent 498708e568
commit c6395ea9b4
12 changed files with 2631 additions and 126 deletions
--- a/backend/src/Health.Infrastructure/AI/prompt_manager.cs
+++ b/backend/src/Health.Infrastructure/AI/prompt_manager.cs
@@ -21,8 +21,7 @@ public sealed class PromptManager
    };

    private const string DefaultPrompt = """
-        你是一个心脏术后康复患者的私人 AI 健康管家，名叫"阿福"。
-        语气温暖、专业、像朋友一样关怀患者。
+        你是一位专业的AI健康管家，专注于为心脏术后康复患者提供贴心的健康管理服务。

        职责：
        1. 理解用户的健康需求，解析健康数据
@@ -34,6 +33,7 @@ public sealed class PromptManager
        - 不要提供超出你能力范围的医疗建议
        - 遇到紧急症状（剧烈胸痛、呼吸困难）立即建议就医
        - 饮食/运动建议要结合患者档案中的疾病和限制
+        - 回复语气温暖、专业、像朋友一样关怀患者
        """;

    private const string ConsultationPrompt = """
--- a/backend/src/Health.WebApi/Endpoints/ai_chat_endpoints.cs
+++ b/backend/src/Health.WebApi/Endpoints/ai_chat_endpoints.cs
@@ -107,6 +107,8 @@ public static class AiChatEndpoints
            var maxIterations = 5;
            var fullResponse = "";
            var completedNormally = false;
+            var messageType = "text";
+            var metadata = new Dictionary<string, object>();

            for (int i = 0; i < maxIterations; i++)
            {
@@ -129,7 +131,7 @@ public static class AiChatEndpoints
                            if (!string.IsNullOrEmpty(content))
                            {
                                fullResponse += content;
-                                await SseWriteAsync(http, new { action = "answer", data = content }, ct);
+                                await SseWriteAsync(http, new { action = "answer", data = content, type = messageType }, ct);
                            }
                        }
                        catch (JsonException) { /* 跳过解析失败的 chunk */ }
@@ -160,6 +162,8 @@ public static class AiChatEndpoints
                        }
                        await SseWriteAsync(http, new { action = "tool_result", tool = tc.Function.Name, data = toolResult }, ct);

+                        _UpdateMessageTypeAndMetadata(tc.Function.Name, toolResult, ref messageType, ref metadata);
+
                        messages.Add(new ChatMessage { Role = "tool", Content = JsonSerializer.Serialize(toolResult, JsonOpts), ToolCallId = tc.Id });
                    }
                }
@@ -597,6 +601,40 @@ public static class AiChatEndpoints
        }
    };

+    /// <summary>根据工具调用结果更新消息类型和元数据</summary>
+    private static void _UpdateMessageTypeAndMetadata(string toolName, object toolResult, ref string messageType, ref Dictionary<string, object> metadata)
+    {
+        switch (toolName)
+        {
+            case "record_health_data":
+                messageType = "data_confirm";
+                if (toolResult is IDictionary<string, object> resultDict)
+                {
+                    if (resultDict.TryGetValue("type", out var type))
+                        metadata["type"] = type.ToString();
+                    if (resultDict.TryGetValue("success", out var success) && success is bool b && b)
+                        metadata["success"] = true;
+                }
+                break;
+            case "manage_medication":
+                messageType = "medication_confirm";
+                if (toolResult is IDictionary<string, object> medDict)
+                {
+                    if (medDict.TryGetValue("name", out var name))
+                        metadata["name"] = name.ToString();
+                    if (medDict.TryGetValue("dosage", out var dosage))
+                        metadata["dosage"] = dosage.ToString();
+                }
+                break;
+            case "estimate_food_text":
+                messageType = "diet_analysis";
+                break;
+            case "analyze_report":
+                messageType = "report_analysis";
+                break;
+        }
+    }
+
    /// <summary>压缩图片到合理大小供 VLM API 使用</summary>
    private static void CompressImage(string inputPath, string outputPath, int maxWidth, long quality)
    {
--- a/backend/tests/full_e2e_test.py
+++ b/backend/tests/full_e2e_test.py
@@ -0,0 +1,283 @@
+"""
+健康管家 - 全流程端到端测试
+模拟真实用户操作：注册→登录→各Agent对话→数据录入→查询验证
+"""
+import urllib.request, urllib.parse, json, sys, time, os
+
+BASE = "http://localhost:5000"
+PASSED = 0
+FAILED = 0
+TOKEN = None
+
+def api(method, path, data=None, token=None, files=None):
+    """调用后端 API"""
+    url = f"{BASE}{path}"
+    headers = {"Content-Type": "application/json"}
+    if token:
+        headers["Authorization"] = f"Bearer {token}"
+
+    body = None
+    if data:
+        body = json.dumps(data, ensure_ascii=False).encode("utf-8")
+
+    req = urllib.request.Request(url, data=body, headers=headers, method=method)
+    try:
+        resp = urllib.request.urlopen(req, timeout=30)
+        return json.loads(resp.read().decode("utf-8"))
+    except Exception as e:
+        return {"error": str(e), "code": -1}
+
+def check(name, condition, detail=""):
+    global PASSED, FAILED
+    if condition:
+        PASSED += 1
+        print(f"  [PASS] {name}")
+    else:
+        FAILED += 1
+        print(f"  [FAIL] {name} {detail}")
+
+def login(phone="13800000001"):
+    """发送验证码 + 登录，返回 token"""
+    sms = api("POST", "/api/auth/send-sms", {"phone": phone})
+    code = sms.get("data", {}).get("devCode", "")
+    if not code:
+        return None
+    result = api("POST", "/api/auth/login", {"phone": phone, "smsCode": code})
+    return result.get("data", {}).get("accessToken")
+
+def sse_stream(token, agent_type, message):
+    """连接 SSE 端点，返回所有事件"""
+    url = f"{BASE}/api/ai/{agent_type}/chat?message={urllib.parse.quote(message)}&token={urllib.parse.quote(token or '')}"
+    req = urllib.request.Request(url)
+    events = []
+    try:
+        resp = urllib.request.urlopen(req, timeout=60)
+        for line_bytes in resp:
+            line = line_bytes.decode("utf-8").strip()
+            if line.startswith("data: "):
+                data = line[6:]
+                if data == "[DONE]":
+                    break
+                try:
+                    events.append(json.loads(data))
+                except:
+                    pass
+    except Exception as e:
+        events.append({"error": str(e)})
+    return events
+
+def section(title):
+    print(f"\n{'='*60}")
+    print(f"  {title}")
+    print(f"{'='*60}")
+
+# ============================================================
+section("1. 认证流程")
+# ============================================================
+print("  1.1 发送验证码...")
+sms = api("POST", "/api/auth/send-sms", {"phone": "13800000001"})
+check("发送验证码", sms.get("code") == 0, str(sms.get("message","")))
+code = sms.get("data", {}).get("devCode", "")
+
+print("  1.2 验证码登录...")
+login_result = api("POST", "/api/auth/login", {"phone": "13800000001", "smsCode": code})
+check("登录成功", login_result.get("code") == 0)
+TOKEN = login_result.get("data", {}).get("accessToken", "")
+REFRESH = login_result.get("data", {}).get("refreshToken", "")
+check("返回accessToken", len(TOKEN) > 50)
+check("返回refreshToken", len(REFRESH) > 20)
+
+print("  1.3 刷新Token...")
+refresh_result = api("POST", "/api/auth/refresh", {"refreshToken": REFRESH})
+check("刷新Token成功", refresh_result.get("code") == 0)
+check("下发新Token", len(refresh_result.get("data", {}).get("accessToken", "")) > 50)
+
+print("  1.4 登出...")
+logout_result = api("POST", "/api/auth/logout", {"refreshToken": REFRESH})
+check("登出成功", logout_result.get("code") == 0)
+
+# 重新登录获取token
+TOKEN = login()
+
+# ============================================================
+section("2. 用户与档案")
+# ============================================================
+profile = api("GET", "/api/user/profile", token=TOKEN)
+check("获取个人信息", profile.get("code") == 0)
+
+archive = api("GET", "/api/user/health-archive", token=TOKEN)
+check("获取健康档案", archive.get("code") == 0)
+check("档案有诊断信息", archive.get("data", {}).get("diagnosis") is not None if archive.get("data") else False,
+      "诊断=" + str(archive.get("data", {}).get("diagnosis", "无")))
+
+# ============================================================
+section("3. 健康数据 CRUD")
+# ============================================================
+print("  3.1 录入血压...")
+bp = api("POST", "/api/health-records", token=TOKEN, data={
+    "type": "BloodPressure", "systolic": 128, "diastolic": 82,
+    "unit": "mmHg", "source": "Manual"
+})
+check("录入血压", bp.get("code") == 0, str(bp.get("message","")))
+
+print("  3.2 录入心率...")
+hr = api("POST", "/api/health-records", token=TOKEN, data={
+    "type": "HeartRate", "value": 72, "unit": "次/分", "source": "Manual"
+})
+check("录入心率", hr.get("code") == 0, str(hr.get("message","")))
+
+print("  3.3 录入血糖...")
+glu = api("POST", "/api/health-records", token=TOKEN, data={
+    "type": "Glucose", "value": 5.5, "unit": "mmol/L", "source": "Manual"
+})
+check("录入血糖", glu.get("code") == 0, str(glu.get("message","")))
+
+print("  3.4 录入血氧...")
+spo2 = api("POST", "/api/health-records", token=TOKEN, data={
+    "type": "SpO2", "value": 98, "unit": "%", "source": "Manual"
+})
+check("录入血氧", spo2.get("code") == 0, str(spo2.get("message","")))
+
+print("  3.5 获取最新数据...")
+latest = api("GET", "/api/health-records/latest", token=TOKEN)
+check("获取最新数据", latest.get("code") == 0)
+check("血压存在", latest.get("data", {}).get("BloodPressure") is not None)
+
+print("  3.6 获取趋势数据...")
+trend = api("GET", "/api/health-records/trend?type=HeartRate&period=7", token=TOKEN)
+check("获取趋势数据", trend.get("code") == 0)
+
+# ============================================================
+section("4. 用药管理")
+# ============================================================
+print("  4.1 获取用药列表...")
+meds = api("GET", "/api/medications", token=TOKEN)
+check("获取用药列表", meds.get("code") == 0)
+
+print("  4.2 添加用药...")
+new_med = api("POST", "/api/medications", token=TOKEN, data={
+    "name": "阿司匹林", "dosage": "100mg", "frequency": "Daily",
+    "timeOfDay": ["08:00"], "source": "Manual", "startDate": "2026-06-02"
+})
+check("添加用药", new_med.get("code") == 0, str(new_med.get("message","")))
+med_id = new_med.get("data", {}).get("id", "")
+
+print("  4.3 服药打卡...")
+if med_id:
+    confirm = api("POST", f"/api/medications/{med_id}/confirm", token=TOKEN)
+    check("服药打卡", confirm.get("code") == 0, str(confirm.get("message","")))
+
+# ============================================================
+section("5. 饮食记录")
+# ============================================================
+diet = api("GET", "/api/diet-records?date=2026-06-02", token=TOKEN)
+check("查询饮食记录", diet.get("code") == 0)
+
+# ============================================================
+section("6. 运动计划")
+# ============================================================
+print("  6.1 获取当前计划...")
+plan = api("GET", "/api/exercise-plans/current", token=TOKEN)
+check("获取当前计划", plan.get("code") == 0)
+
+print("  6.2 创建运动计划...")
+new_plan = api("POST", "/api/exercise-plans", token=TOKEN, data={
+    "weekStartDate": "2026-06-02",
+    "items": [
+        {"dayOfWeek": 1, "exerciseType": "散步", "durationMinutes": 30, "isRestDay": False},
+        {"dayOfWeek": 3, "exerciseType": "太极", "durationMinutes": 40, "isRestDay": False},
+        {"dayOfWeek": 5, "exerciseType": "散步", "durationMinutes": 30, "isRestDay": False},
+    ]
+})
+check("创建运动计划", new_plan.get("code") == 0, str(new_plan.get("message","")))
+
+# ============================================================
+section("7. 医生与问诊")
+# ============================================================
+print("  7.1 医生列表...")
+docs = api("GET", "/api/doctors", token=TOKEN)
+check("获取医生列表", docs.get("code") == 0)
+check("有医生数据", len(docs.get("data", [])) > 0, f"共{len(docs.get('data',[]))}位医生")
+
+print("  7.2 问诊配额...")
+quota = api("GET", "/api/user/consultation-quota", token=TOKEN)
+check("获取问诊配额", quota.get("code") == 0)
+
+# ============================================================
+section("8. AI 智能体对话")
+# ============================================================
+agents_to_test = [
+    ("default", "你好，介绍一下你自己"),
+    ("health", "我血压128/82"),
+    ("medication", "我现在在吃什么药"),
+    ("consultation", "最近胸口有点不舒服"),
+    ("diet", "我中午吃了红烧肉和米饭"),
+    ("exercise", "帮我查询运动计划"),
+]
+for agent_name, msg in agents_to_test:
+    print(f"  8.{agents_to_test.index((agent_name,msg))+1} {agent_name} Agent: \"{msg[:30]}...\"")
+    TOKEN = login()  # fresh token
+    events = sse_stream(TOKEN, agent_name, msg)
+
+    has_answer = any(e.get("action") == "answer" for e in events)
+    has_tool = any(e.get("action") == "tool_result" for e in events)
+    has_conv_id = any(e.get("action") == "conversation_id" for e in events)
+    errors = [e for e in events if e.get("action") == "error"]
+
+    check(f"{agent_name}: 对话建立", has_conv_id)
+    check(f"{agent_name}: 有回复", has_answer or has_tool,
+          f"(events: {len(events)}, tools: {has_tool}, answer: {has_answer})")
+    check(f"{agent_name}: 无错误", len(errors) == 0,
+          f"errors: {[e.get('message','') for e in errors]}" if errors else "")
+
+# ============================================================
+section("9. 对话历史")
+# ============================================================
+convs = api("GET", "/api/ai/conversations", token=TOKEN)
+check("获取对话列表", convs.get("code") == 0)
+check("有对话记录", len(convs.get("data", [])) > 0, f"共{len(convs.get('data',[]))}条")
+
+# ============================================================
+section("10. VLM 食物识别")
+# ============================================================
+# 尝试上传测试图片
+test_img = "D:/health_project/食堂三菜一饭热量估算.png"
+if os.path.exists(test_img):
+    # Use subprocess for multipart upload
+    import subprocess
+    cmd = [
+        'curl', '-s', '--max-time', '30', '-X', 'POST',
+        f'{BASE}/api/ai/analyze-food-image',
+        '-H', f'Authorization: Bearer {TOKEN}',
+        '-F', f'images=@{test_img}'
+    ]
+    r = subprocess.run(cmd, capture_output=True, text=True)
+    try:
+        vlm = json.loads(r.stdout)
+        check("VLM食物识别调通", vlm.get("code") == 0, str(vlm.get("message","")))
+        has_data = bool(vlm.get("data", ""))
+        check("VLM返回数据", has_data, f"data长度: {len(str(vlm.get('data','')))}")
+    except:
+        check("VLM食物识别", False, "JSON解析失败")
+else:
+    check("VLM测试图片存在", False, f"{test_img} 不存在")
+
+# ============================================================
+section("11. 报告列表")
+# ============================================================
+reports = api("GET", "/api/reports", token=TOKEN)
+check("获取报告列表", reports.get("code") == 0)
+
+# ============================================================
+section("12. 通知偏好")
+# ============================================================
+notifs = api("GET", "/api/notifications/preferences", token=TOKEN)
+check("获取通知偏好", notifs.get("code") == 0)
+
+# ============================================================
+print(f"\n{'='*60}")
+print(f"  测试结果: PASS={PASSED} FAIL={FAILED} TOTAL={PASSED+FAILED}")
+print(f"{'='*60}")
+
+if FAILED > 0:
+    sys.exit(1)