修改打包配置

pom.xml

@@ -150,7 +150,13 @@
 	</dependencyManagement>
 
 	<build>
+		<finalName>ai-interview</finalName>
 		<plugins>
+			<plugin>
+				<groupId>org.springframework.boot</groupId>
+				<artifactId>spring-boot-maven-plugin</artifactId>
+				<version>3.5.0</version>
+			</plugin>
 			<!-- maven 打包时跳过测试 -->
 			<plugin>
 				<groupId>org.apache.maven.plugins</groupId>

src/main/java/com/qingqiu/interview/aspect/AiChatLogAspect.java

@@ -1,10 +1,17 @@
 package com.qingqiu.interview.aspect;
 
+import com.qingqiu.interview.dto.ChatDTO;
+import com.qingqiu.interview.entity.AiSessionLog;
+import com.qingqiu.interview.service.IAiSessionLogService;
+import com.qingqiu.interview.vo.ChatVO;
+import jakarta.annotation.Resource;
+import org.apache.commons.lang3.StringUtils;
 import org.aspectj.lang.ProceedingJoinPoint;
 import org.aspectj.lang.annotation.Around;
 import org.aspectj.lang.annotation.Aspect;
 import org.aspectj.lang.annotation.Pointcut;
 import org.springframework.stereotype.Component;
+import org.springframework.transaction.annotation.Transactional;
 
 /**
  * <h1>
@@ -18,6 +25,9 @@ import org.springframework.stereotype.Component;
 @Component
 public class AiChatLogAspect {
 
+    @Resource
+    private IAiSessionLogService aiSessionLogService;
+
     public AiChatLogAspect() {
 
     }
@@ -27,8 +37,35 @@ public class AiChatLogAspect {
     }
 
     @Around("logPointCut()")
+    @Transactional(rollbackFor = Exception.class)
     public Object around(ProceedingJoinPoint point) throws Throwable {
+
+        Object[] args = point.getArgs();
+        ChatDTO arg = (ChatDTO) args[0];
+        if (StringUtils.isNoneBlank(arg.getSessionId())) {
+            AiSessionLog userSessionLog = new AiSessionLog();
+            userSessionLog
+                    .setRole(arg.getRole())
+                    .setDataType(arg.getDataType())
+                    .setContent(arg.getContent())
+                    .setToken(arg.getSessionId())
+            ;
+            aiSessionLogService.save(userSessionLog);
+        }
+
+
         Object result = point.proceed();
+
+        ChatVO chatVO = (ChatVO) result;
+        if (StringUtils.isNotBlank(chatVO.getSessionId())) {
+            AiSessionLog aiSessionLog = new AiSessionLog();
+            aiSessionLog
+                    .setRole(chatVO.getRole())
+                    .setContent(chatVO.getContent())
+                    .setToken(chatVO.getSessionId())
+            ;
+            aiSessionLogService.save(aiSessionLog);
+        }
         return result;
     }
 }

src/main/java/com/qingqiu/interview/common/dto/PageBaseParams.java

@@ -1,4 +1,4 @@
-package com.qingqiu.interview.dto;
+package com.qingqiu.interview.common.dto;
 
 import lombok.Data;
 import lombok.EqualsAndHashCode;

src/main/java/com/qingqiu/interview/controller/AiSessionLogController.java

@@ -1,10 +1,22 @@
 package com.qingqiu.interview.controller;
 
 
+import com.alibaba.dashscope.common.Role;
+import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
+import com.qingqiu.interview.common.res.R;
+import com.qingqiu.interview.entity.AiSessionLog;
+import com.qingqiu.interview.service.IAiSessionLogService;
+import lombok.RequiredArgsConstructor;
+import lombok.extern.slf4j.Slf4j;
+import org.springframework.beans.factory.annotation.Autowired;
+import org.springframework.context.annotation.Lazy;
+import org.springframework.web.bind.annotation.GetMapping;
+import org.springframework.web.bind.annotation.PathVariable;
 import org.springframework.web.bind.annotation.RequestMapping;
-
 import org.springframework.web.bind.annotation.RestController;
 
+import java.util.List;
+
 /**
  * <p>
  * ai会话记录 前端控制器
@@ -13,8 +25,22 @@ import org.springframework.web.bind.annotation.RestController;
  * @author huangpeng
  * @since 2025-08-30
  */
+@Slf4j
 @RestController
 @RequestMapping("/ai-session-log")
+@RequiredArgsConstructor(onConstructor_ = {@Autowired, @Lazy})
 public class AiSessionLogController {
 
+    private final IAiSessionLogService service;
+
+    @GetMapping("/list-by-session-id/{sessionId}")
+    public R<List<AiSessionLog>> list(@PathVariable String sessionId) {
+        return R.success(service.list(
+                new LambdaQueryWrapper<AiSessionLog>()
+                        .eq(AiSessionLog::getToken, sessionId)
+                        .ne(AiSessionLog::getRole, Role.SYSTEM.getValue())
+        ));
+    }
+
+
 }

src/main/java/com/qingqiu/interview/controller/InterviewController.java

@@ -1,65 +1,125 @@
 package com.qingqiu.interview.controller;
 
-import com.alibaba.fastjson2.JSONObject;
+import com.qingqiu.interview.common.res.R;
 import com.qingqiu.interview.dto.*;
+import com.qingqiu.interview.entity.InterviewMessage;
+import com.qingqiu.interview.entity.InterviewQuestionProgress;
 import com.qingqiu.interview.entity.InterviewSession;
 import com.qingqiu.interview.service.InterviewService;
 import lombok.RequiredArgsConstructor;
 import lombok.extern.slf4j.Slf4j;
-import org.springframework.validation.annotation.Validated;
+import org.springframework.beans.factory.annotation.Autowired;
+import org.springframework.context.annotation.Lazy;
 import org.springframework.web.bind.annotation.*;
 import org.springframework.web.multipart.MultipartFile;
 
-import java.io.IOException;
+import java.util.List;
-import java.util.UUID;
 
 /**
- * 面试流程相关接口
+ * <h1></h1>
+ *
+ * @author qingqiu
+ * @date 2025/9/19 16:13
  */
 @Slf4j
 @RestController
 @RequestMapping("/interview")
-@RequiredArgsConstructor
+@RequiredArgsConstructor(onConstructor_ = {@Autowired, @Lazy})
 public class InterviewController {
 
     private final InterviewService interviewService;
 
+
     /**
-     * 开始新的面试会话
+     * 开始面试
+     *
+     * @return 包含会话ID的会话信息
      */
     @PostMapping("/start")
-    public ApiResponse<InterviewResponse> startInterview(
+    public R<InterviewSession> start(@RequestPart("resume") MultipartFile resume,
-            @RequestParam("resume") MultipartFile resume,
+                                     @RequestPart("interviewStartDto") InterviewStartRequest request) {
-            @Validated @ModelAttribute InterviewStartRequest request) throws IOException {
+//        log.info("接受的数据: {}", JSONObject.toJSONString(request));
-//        InterviewResponse response = interviewService.startInterview(resume, request);
+//        return R.success();
-        log.info("接收到的数据: {}", JSONObject.toJSONString(request));
+        try {
-        InterviewResponse interviewResponse = new InterviewResponse();
+            InterviewSession session = interviewService.startInterview(resume, request);
-        interviewResponse.setSessionId(UUID.randomUUID().toString().replace("-", ""));
+            return R.success(session);
-        return ApiResponse.success(interviewResponse);
+        } catch (Exception e) {
+            log.error("开始面试失败", e);
+            return R.error("开始面试失败：" + e.getMessage());
+        }
     }
 
     /**
-     * 继续面试会话（用户回答）
+     * 获取下一个问题
+     *
+     * @param sessionId 会话ID
+     * @return 下一个问题
      */
-    @PostMapping("/chat")
+    @GetMapping("/next-question/{sessionId}/{progressId}")
-    public ApiResponse<InterviewResponse> continueInterview(@Validated @RequestBody ChatRequest request) {
+    public R<InterviewMessage> getNextQuestion(@PathVariable String sessionId,
-        InterviewResponse response = interviewService.continueInterview(request);
+                                               @PathVariable Long progressId) {
-        return ApiResponse.success(response);
+        try {
+            InterviewMessage nextQuestion = interviewService.getNextQuestion(sessionId, progressId);
+            if (nextQuestion == null) {
+                return R.success(null, "所有问题已回答完毕！");
+            }
+            return R.success(nextQuestion);
+        } catch (Exception e) {
+            // log.error("获取下一题失败", e);
+            return R.error("获取下一题失败：" + e.getMessage());
+        }
     }
 
     /**
-     * 获取所有面试会话列表
+     * 提交答案
+     *
+     * @param submitDto 包含进度ID和答案
+     * @return 对当前问题的评估
      */
+    @PostMapping("/submit-answer")
+    public R<InterviewQuestionProgress> submitAnswer(@RequestBody SubmitAnswerDTO submitDto) {
+        try {
+            InterviewQuestionProgress result = interviewService.submitAnswer(submitDto);
+            return R.success(result);
+        } catch (Exception e) {
+            // log.error("提交答案失败", e);
+            return R.error("提交答案失败：" + e.getMessage());
+        }
+    }
+
+    /**
+     * 结束面试并获取最终报告
+     *
+     * @param sessionId 会话ID
+     * @return 包含最终报告的会话信息
+     */
+    @PostMapping("/{sessionId}/end")
+    public R<InterviewSession> endInterview(@PathVariable String sessionId) {
+        try {
+            InterviewSession finalSession = interviewService.endInterview(sessionId);
+            return R.success(finalSession);
+        } catch (Exception e) {
+            // log.error("结束面试失败", e);
+            return R.error("结束面试失败：" + e.getMessage());
+        }
+    }
+
     @PostMapping("/get-history-list")
-    public ApiResponse<java.util.List<InterviewSession>> getInterviewHistoryList() {
+    public R<List<InterviewSession>> getHistoryList() {
-        return ApiResponse.success(interviewService.getInterviewSessions());
+        try {
+            List<InterviewSession> historyList = interviewService.list();
+            return R.success(historyList);
+        } catch (Exception e) {
+            // log.error("获取面试历史列表失败", e);
+            return R.error("获取面试历史列表失败：" + e.getMessage());
+        }
     }
 
     /**
      * 获取单次面试的详细复盘报告
      */
-    @PostMapping("/get-report-detail")
+    @PostMapping("/get-report-detail/{sessionId}")
-    public ApiResponse<InterviewReportResponse> getInterviewReportDetail(@RequestBody SessionRequest request) {
+    public R<InterviewReportResponse> getInterviewReportDetail(@PathVariable String sessionId) {
-        return ApiResponse.success(interviewService.getInterviewReport(request.getSessionId()));
+        return R.success(interviewService.getInterviewReport(sessionId));
     }
 }

src/main/java/com/qingqiu/interview/controller/InterviewMessageController.java

@@ -0,0 +1,42 @@
+package com.qingqiu.interview.controller;
+
+import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
+import com.qingqiu.interview.common.res.R;
+import com.qingqiu.interview.entity.InterviewMessage;
+import com.qingqiu.interview.service.InterviewMessageService;
+import lombok.RequiredArgsConstructor;
+import lombok.extern.slf4j.Slf4j;
+import org.springframework.beans.factory.annotation.Autowired;
+import org.springframework.context.annotation.Lazy;
+import org.springframework.web.bind.annotation.GetMapping;
+import org.springframework.web.bind.annotation.PathVariable;
+import org.springframework.web.bind.annotation.RequestMapping;
+import org.springframework.web.bind.annotation.RestController;
+
+import java.util.List;
+
+/**
+ * <h1></h1>
+ *
+ * @author qingqiu
+ * @date 2025/9/21 11:59
+ */
+@Slf4j
+@RestController
+@RequestMapping("/interview-message")
+@RequiredArgsConstructor(onConstructor_ = {@Autowired, @Lazy})
+public class InterviewMessageController {
+
+    public final InterviewMessageService service;
+
+    @GetMapping("/list-by-session-id/{sessionId}")
+    public R<List<InterviewMessage>> listBySessionId(@PathVariable String sessionId) {
+        return R.success(
+                service.list(
+                        new LambdaQueryWrapper<InterviewMessage>()
+                                .eq(InterviewMessage::getSessionId, sessionId)
+                                .orderByAsc(InterviewMessage::getCreatedTime)
+                )
+        );
+    }
+}

src/main/java/com/qingqiu/interview/dto/ChatDTO.java

@@ -1,5 +1,6 @@
 package com.qingqiu.interview.dto;
 
+import com.qingqiu.interview.common.enums.LLMProvider;
 import lombok.Data;
 import lombok.experimental.Accessors;
 
@@ -10,13 +11,14 @@ import lombok.experimental.Accessors;
  * @date 2025/9/18 12:54
  */
 @Data
+
 @Accessors(chain = true)
 public class ChatDTO {
 
     /** 会话id */
     private String sessionId;
     /** 调用模型 */
-    private String aiModel;
+    private String aiModel = LLMProvider.DEEPSEEK.getCode();
     /** 输入内容 */
     private String content;
     /** 0 普通会话 1 面试会话 */

src/main/java/com/qingqiu/interview/dto/InterviewStartRequest.java

@@ -1,5 +1,6 @@
 package com.qingqiu.interview.dto;
 
+import com.qingqiu.interview.common.enums.LLMProvider;
 import com.qingqiu.interview.vo.QuestionAndCategoryTreeListVO;
 import jakarta.validation.constraints.NotBlank;
 import lombok.Data;
@@ -18,6 +19,11 @@ public class InterviewStartRequest {
     @NotBlank(message = "面试类型不能为空")
     private String model;
 
+    /** 选择的AI模型 */
+    private String aiModel = LLMProvider.QWEN.getCode();
+
+    /** 生成的面试题目数量 */
+    private Integer totalQuestions = 10;
 
     // 简历文件通过MultipartFile单独传递
 }

src/main/java/com/qingqiu/interview/dto/QuestionCategoryPageParams.java

@@ -1,5 +1,6 @@
 package com.qingqiu.interview.dto;
 
+import com.qingqiu.interview.common.dto.PageBaseParams;
 import lombok.Data;
 import lombok.EqualsAndHashCode;
 import lombok.experimental.Accessors;
@@ -14,7 +15,7 @@ import lombok.experimental.Accessors;
 @EqualsAndHashCode(callSuper = true)
 @Data
 @Accessors(chain = true)
-public class QuestionCategoryPageParams extends PageBaseParams{
+public class QuestionCategoryPageParams extends PageBaseParams {
     /**
      * 分类名称（模糊查询）
      */

src/main/java/com/qingqiu/interview/dto/QuestionPageParams.java

@@ -1,5 +1,6 @@
 package com.qingqiu.interview.dto;
 
+import com.qingqiu.interview.common.dto.PageBaseParams;
 import lombok.Data;
 import lombok.EqualsAndHashCode;
 import lombok.experimental.Accessors;

src/main/java/com/qingqiu/interview/dto/QuestionProgressPageParams.java

@@ -1,5 +1,6 @@
 package com.qingqiu.interview.dto;
 
+import com.qingqiu.interview.common.dto.PageBaseParams;
 import lombok.Data;
 import lombok.EqualsAndHashCode;
 import lombok.experimental.Accessors;
@@ -7,6 +8,6 @@ import lombok.experimental.Accessors;
 @EqualsAndHashCode(callSuper = true)
 @Data
 @Accessors(chain = true)
-public class QuestionProgressPageParams extends PageBaseParams{
+public class QuestionProgressPageParams extends PageBaseParams {
     private String questionName;
 }

src/main/java/com/qingqiu/interview/dto/StartInterviewDTO.java

@@ -0,0 +1,43 @@
+package com.qingqiu.interview.dto;
+
+import lombok.Data;
+import lombok.experimental.Accessors;
+
+import java.io.Serial;
+import java.io.Serializable;
+
+/**
+ * <h1>
+ *     开始面试请求的数据传输对象
+ * </h1>
+ *
+ * @author qingqiu
+ * @date 2025/9/19 16:03
+ */
+@Data
+@Accessors(chain = true)
+public class StartInterviewDTO implements Serializable {
+
+    @Serial
+    private static final long serialVersionUID = 1L;
+
+    /**
+     * 候选人姓名
+     */
+    private String candidateName;
+
+    /**
+     * 简历完整内容（或简历文件URL）
+     */
+    private String resumeContent;
+
+    /**
+     * 指定使用的AI模型
+     */
+    private String aiModel;
+
+    /**
+     * 计划提问总数
+     */
+    private Integer totalQuestions;
+}

src/main/java/com/qingqiu/interview/dto/SubmitAnswerDTO.java

@@ -0,0 +1,33 @@
+package com.qingqiu.interview.dto;
+
+import lombok.Data;
+import lombok.experimental.Accessors;
+
+import java.io.Serial;
+import java.io.Serializable;
+
+/**
+ * <h1></h1>
+ *
+ * @author qingqiu
+ * @date 2025/9/19 16:04
+ */
+@Data
+@Accessors(chain = true)
+public class SubmitAnswerDTO implements Serializable {
+
+    @Serial
+    private static final long serialVersionUID = 1L;
+
+    private String sessionId;
+
+    /**
+     * 当前问题的进度ID (interview_question_progress.id)
+     */
+    private Long progressId;
+
+    /**
+     * 用户的回答内容
+     */
+    private String answer;
+}

src/main/java/com/qingqiu/interview/entity/InterviewMessage.java

@@ -28,8 +28,8 @@ public class InterviewMessage {
     @TableField("content")
     private String content;
 
-    @TableField("question_id")
+    @TableField("question_progress_id")
-    private Long questionId;
+    private Long questionProgressId;
 
     @TableField("message_order")
     private Integer messageOrder;

src/main/java/com/qingqiu/interview/entity/InterviewSession.java

@@ -44,6 +44,8 @@ public class InterviewSession implements Serializable {
 
     @TableField("ai_model")
     private String aiModel;
+    @TableField("model")
+    private String model;
 
     @TableField("status")
     private String status;

src/main/java/com/qingqiu/interview/service/IInterviewQuestionProgressService.java

@@ -15,4 +15,6 @@ import com.qingqiu.interview.entity.InterviewQuestionProgress;
  */
 public interface IInterviewQuestionProgressService extends IService<InterviewQuestionProgress> {
     Page<InterviewQuestionProgress> pageList(QuestionProgressPageParams params);
+
+    InterviewQuestionProgress getNextQuestion(String sessionId);
 }

src/main/java/com/qingqiu/interview/service/InterviewAiService.java

@@ -0,0 +1,61 @@
+package com.qingqiu.interview.service;
+
+import com.alibaba.fastjson2.JSONObject;
+import com.qingqiu.interview.entity.InterviewQuestionProgress;
+import com.qingqiu.interview.entity.InterviewSession;
+import com.qingqiu.interview.entity.Question;
+
+import java.util.List;
+
+/**
+ * <h1>
+ *     面试接入AI的接口
+ * </h1>
+ *
+ * @author qingqiu
+ * @date 2025/9/19 16:48
+ */
+public interface InterviewAiService {
+
+    /**
+     * 从简历内容中提取技能列表
+     *
+     * @param resumeContent 简历文本
+     * @return 包含技能列表的JSON对象
+     */
+    JSONObject extractSkillsFromResume(String resumeContent);
+
+    /**
+     * 根据技能动态生成面试题目
+     *
+     * @param skills      技能列表
+     * @param resumeContent 简历内容
+     * @param count       需要生成的题目数量
+     * @return 包含问题列表的JSON对象
+     */
+    JSONObject generateQuestionsOfAi(String sessionId, List<String> skills, String resumeContent, int count);
+
+    JSONObject generateQuestionOfLocal(String sessionId, List<Question> questions, List<String> skills, String resumeContent, int count);
+
+    /**
+     * 评估用户的回答
+     *
+     * @param question    问题内容
+     * @param userAnswer  用户的回答
+     * @param context     可选的上下文(之前的问答历史)
+     * @return 包含评估结果的JSON对象
+     */
+    JSONObject evaluateAnswer(String sessionId, String question, String userAnswer, List<InterviewQuestionProgress> context);
+
+    /**
+     * 生成最终的面试评估报告
+     *
+     * @param session        面试会话信息
+     * @param progressList   整个面试的问答记录
+     * @return 包含最终报告的JSON对象
+     */
+    JSONObject generateFinalReport(InterviewSession session, List<InterviewQuestionProgress> progressList);
+
+    String generateFirstQuestion(String sessionId, String candidateName, String questionContent);
+
+}

src/main/java/com/qingqiu/interview/service/InterviewMessageService.java

@@ -0,0 +1,13 @@
+package com.qingqiu.interview.service;
+
+import com.baomidou.mybatisplus.extension.service.IService;
+import com.qingqiu.interview.entity.InterviewMessage;
+
+/**
+ * <h1></h1>
+ *
+ * @author qingqiu
+ * @date 2025/9/21 12:00
+ */
+public interface InterviewMessageService extends IService<InterviewMessage> {
+}

src/main/java/com/qingqiu/interview/service/InterviewService.java

@@ -1,579 +1,61 @@
 package com.qingqiu.interview.service;
 
-import cn.hutool.core.collection.CollectionUtil;
+import com.baomidou.mybatisplus.extension.service.IService;
-import cn.hutool.core.util.StrUtil;
+import com.qingqiu.interview.dto.InterviewReportResponse;
-import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
+import com.qingqiu.interview.dto.InterviewStartRequest;
-import com.baomidou.mybatisplus.core.conditions.update.LambdaUpdateWrapper;
+import com.qingqiu.interview.dto.SubmitAnswerDTO;
-import com.fasterxml.jackson.core.JsonProcessingException;
+import com.qingqiu.interview.entity.InterviewMessage;
-import com.fasterxml.jackson.databind.JsonNode;
+import com.qingqiu.interview.entity.InterviewQuestionProgress;
-import com.fasterxml.jackson.databind.ObjectMapper;
+import com.qingqiu.interview.entity.InterviewSession;
-import com.qingqiu.interview.dto.*;
-import com.qingqiu.interview.entity.*;
-import com.qingqiu.interview.mapper.*;
-import com.qingqiu.interview.service.llm.LlmService;
-import com.qingqiu.interview.service.parser.DocumentParser;
-import jakarta.annotation.PostConstruct;
-import lombok.RequiredArgsConstructor;
-import lombok.extern.slf4j.Slf4j;
-import org.apache.commons.lang3.StringUtils;
-import org.springframework.stereotype.Service;
-import org.springframework.transaction.annotation.Transactional;
 import org.springframework.web.multipart.MultipartFile;
 
 import java.io.IOException;
-import java.math.BigDecimal;
-import java.util.*;
-import java.util.function.Function;
-import java.util.stream.Collectors;
 
-import static com.qingqiu.interview.common.constants.QwenModelConstant.QWEN_MAX;
+/**
-
+ * <h1></h1>
-@Slf4j
+ *
-@Service
+ * @author qingqiu
-@RequiredArgsConstructor
+ * @date 2025/9/19 16:05
-public class InterviewService {
+ */
-
-    private final LlmService llmService; // Changed to a single service
-    private final List<DocumentParser> documentParserList;
-    private final QuestionMapper questionMapper;
-    private final InterviewSessionMapper sessionMapper;
-    private final InterviewMessageMapper messageMapper;
-    private final InterviewEvaluationMapper evaluationMapper;
-    private final InterviewQuestionProgressMapper questionProgressMapper;
-
-    private final ObjectMapper objectMapper;
-
-
-    private Map<String, DocumentParser> documentParsers;
-
-    private static final int MAX_QUESTIONS_PER_INTERVIEW = 10;
-
-    @PostConstruct
-    public void init() {
-        this.documentParsers = documentParserList.stream()
-                .collect(Collectors.toMap(DocumentParser::getSupportedType, Function.identity()));
-    }
 
+public interface InterviewService extends IService<InterviewSession> {
     /**
-     * 开始新的面试会话
+     * 开始一场新的面试
-     */
-    @Transactional(rollbackFor = Exception.class)
-    public InterviewResponse startInterview(MultipartFile resume, InterviewStartRequest request) throws IOException {
-        log.info("开始新面试会话，候选人: {}, AI模型: qwen-max", request.getCandidateName());
-
-        // 1. 解析简历
-        String resumeContent = parseResume(resume);
-        // 判断是否AI出题
-        if (request.getModel().equals("local")) {
-            if (CollectionUtil.isEmpty(request.getSelectedNodes())) {
-
-            }
-        }
-
-
-        // 2. 创建会话 并发送AI请求 让其从题库中智能抽题
-        String sessionId = UUID.randomUUID().toString();
-        List<Question> selectedQuestions = selectQuestionsByAi(resumeContent, sessionId);
-        if (selectedQuestions.isEmpty()) {
-            throw new IllegalStateException("AI未能成功选取题目，请检查AI服务或题库。");
-        }
-
-        // 生成面试问题进度数据
-        if (CollectionUtil.isNotEmpty(selectedQuestions)) {
-            for (Question question : selectedQuestions) {
-                InterviewQuestionProgress progress =
-                        new InterviewQuestionProgress()
-                                .setSessionId(sessionId)
-                                .setQuestionId(question.getId())
-                                .setQuestionContent(question.getContent())
-                                .setStatus(InterviewQuestionProgress.Status.DEFAULT.name())
-                                .setTotalQuestions(selectedQuestions.size())
-                                .setScore(BigDecimal.ZERO)
-                                .setAiModel(QWEN_MAX)
-                                .setCandidateName(request.getCandidateName());
-                questionProgressMapper.insert(progress);
-            }
-
-        }
-
-        // 3. 保存AI选择的题目ID列表
-        List<Long> selectedQuestionIds = selectedQuestions.stream().map(Question::getId).collect(Collectors.toList());
-        String selectedQuestionIdsJson = objectMapper.writeValueAsString(selectedQuestionIds);
-
-        InterviewSession session = createSession(sessionId, request, resumeContent, selectedQuestionIdsJson);
-        session.setTotalQuestions(selectedQuestions.size()); // 更新会话中的总问题数
-        sessionMapper.updateById(session); // 更新数据库
-
-        // 4. 生成第一个问题
-        Question firstQuestion = selectedQuestions.get(0);
-        String firstQuestionContent = generateFirstQuestion(session, firstQuestion, sessionId);
-        // 激活问题
-        questionProgressMapper.update(
-                new LambdaUpdateWrapper<InterviewQuestionProgress>()
-                        .set(InterviewQuestionProgress::getStatus, InterviewQuestionProgress.Status.ACTIVE.name())
-                        .eq(InterviewQuestionProgress::getQuestionId, firstQuestion.getId())
-                        .eq(InterviewQuestionProgress::getSessionId, sessionId)
-        );
-
-        // 5. 保存消息记录
-        saveMessage(sessionId, InterviewMessage.MessageType.QUESTION.name(),
-                InterviewMessage.Sender.AI.name(), firstQuestionContent, firstQuestion.getId(), 1);
-
-        // 6. 返回响应
-        return new InterviewResponse()
-                .setSessionId(sessionId)
-                .setMessage(firstQuestionContent)
-                .setMessageType(InterviewMessage.MessageType.QUESTION.name())
-                .setSender(InterviewMessage.Sender.AI.name())
-                .setCurrentQuestionIndex(1)
-                .setCurrentQuestionId(firstQuestion.getId())
-                .setTotalQuestions(selectedQuestions.size())
-                .setStatus(InterviewSession.Status.ACTIVE.name());
-    }
-
-    /**
-     * 处理用户回答并生成下一个问题
-     */
-    @Transactional(rollbackFor = Exception.class)
-    public InterviewResponse continueInterview(ChatRequest request) {
-        log.info("继续面试会话: {}", request.getSessionId());
-
-        InterviewSession session = sessionMapper.selectBySessionId(request.getSessionId());
-        if (session == null) {
-            throw new IllegalArgumentException("会话不存在: " + request.getSessionId());
-        }
-
-        if (!InterviewSession.Status.ACTIVE.name().equals(session.getStatus())) {
-            throw new IllegalStateException("会话已结束");
-        }
-
-
-        // 1. 保存用户回答
-        int nextOrder = messageMapper.selectMaxOrderBySessionId(request.getSessionId()) + 1;
-        saveMessage(request.getSessionId(), InterviewMessage.MessageType.ANSWER.name(),
-                InterviewMessage.Sender.USER.name(), request.getUserAnswer(), null, nextOrder);
-        // 检查是否结束面试
-        InterviewQuestionProgress progress = questionProgressMapper.selectOne(
-                new LambdaQueryWrapper<InterviewQuestionProgress>()
-                        .eq(InterviewQuestionProgress::getSessionId, request.getSessionId())
-                        .orderByDesc(InterviewQuestionProgress::getCreatedTime)
-                        .last("limit 1")
-        );
-        if (Objects.nonNull(progress) && Objects.equals(progress.getQuestionId(), request.getCurrentQuestionId())) {
-
-        }
-        // 2. 评估回答
-        Long currentQuestionId = evaluateAnswer(session, request.getUserAnswer());
-        // 比对返回的id是否与当前id一致
-        if (currentQuestionId.equals(0L)) {
-            return finishInterview(session);
-        }
-        InterviewQuestionProgress nextQuestionProgress = questionProgressMapper.selectOne(
-                new LambdaQueryWrapper<InterviewQuestionProgress>()
-                        .eq(InterviewQuestionProgress::getSessionId, request.getSessionId())
-                        .eq(InterviewQuestionProgress::getQuestionId, currentQuestionId)
-                        .orderByDesc(InterviewQuestionProgress::getCreatedTime)
-                        .last("limit 1")
-        );
-        // 将ai返回的内容拼装返回给页面
-        // 查询数据
-        InterviewQuestionProgress currentQuestionData = questionProgressMapper.selectOne(
-                new LambdaQueryWrapper<InterviewQuestionProgress>()
-                        .eq(InterviewQuestionProgress::getSessionId, request.getSessionId())
-                        .eq(InterviewQuestionProgress::getQuestionId, request.getCurrentQuestionId())
-                        .orderByDesc(InterviewQuestionProgress::getCreatedTime)
-                        .last("limit 1")
-        );
-        StringBuilder sb = new StringBuilder();
-        if (Objects.nonNull(currentQuestionData)) {
-            if (StringUtils.isNotBlank(currentQuestionData.getFeedback())) {
-                sb.append(currentQuestionData.getFeedback()).append("\n");
-            }
-            if (StringUtils.isNotBlank(currentQuestionData.getSuggestions())) {
-                sb.append(currentQuestionData.getSuggestions()).append("\n");
-            }
-            if (StringUtils.isNotBlank(currentQuestionData.getAiAnswer())) {
-                sb.append(currentQuestionData.getAiAnswer()).append("\n");
-            }
-        }
-
-        if (!currentQuestionId.equals(request.getCurrentQuestionId())) {
-            // 5. 生成并保存AI的提问消息
-            String nextQuestionContent = String.format("好的，下一个问题是：%s", nextQuestionProgress.getQuestionContent());
-            sb.append(nextQuestionContent);
-            int messageOrder = messageMapper.selectMaxOrderBySessionId(session.getSessionId()) + 1;
-            saveMessage(session.getSessionId(), InterviewMessage.MessageType.QUESTION.name(),
-                    InterviewMessage.Sender.AI.name(), nextQuestionContent, currentQuestionId, messageOrder);
-        }
-        // 6. 返回响应
-        return new InterviewResponse()
-                .setSessionId(session.getSessionId())
-                .setMessage(sb.toString())
-                .setMessageType(InterviewMessage.MessageType.QUESTION.name())
-                .setSender(InterviewMessage.Sender.AI.name())
-                .setCurrentQuestionIndex(session.getCurrentQuestionIndex())
-                .setTotalQuestions(session.getTotalQuestions())
-                .setCurrentQuestionId(currentQuestionId)
-                .setStatus(InterviewSession.Status.ACTIVE.name());
-
-    }
-
-
-
-
-    private String parseResume(MultipartFile resume) throws IOException {
-        String fileExtension = getFileExtension(resume.getOriginalFilename());
-        DocumentParser parser = documentParsers.get(fileExtension);
-        if (parser == null) {
-            throw new IllegalArgumentException("不支持的简历文件类型: " + fileExtension);
-        }
-        return parser.parse(resume.getInputStream());
-    }
-
-    private List<Question> selectQuestionsByAi(String resumeContent, String sessionId) throws JsonProcessingException {
-        // 1. 获取全部题库
-        List<Question> allQuestions = questionMapper.selectList(null);
-        String questionBankJson = objectMapper.writeValueAsString(allQuestions);
-
-        // 2. 构建发送给AI的提示
-        String prompt = String.format("""
-                你是一位专业的面试官。请根据以下候选人的简历内容，从提供的题库中，精心挑选出 %d 道最相关的题目进行面试。
-                
-                要求：
-                1. 题目必须严格从【题库JSON】中选择。
-                2. 挑选的题目应根据候选人的简历内容来抽取。
-                3. 返回一个只包含所选题目ID的JSON数组，格式为：{"question_ids": [1, 5, 23, ...]}。
-                4. 不要返回任何多余的代码，包括markdown形式的代码，我只需要JSON对象，请严格按照api接口形式返回
-                5. 不要返回任何额外的解释或文字，只返回JSON对象。
-                6. 严格按照前后端分离的接口形式返回JSON数据给我，不要返回"```json```"
-                7. 请保证返回数据的完整性，不要返回不完整的数据，否则我的JSON解析会报错！！！
-                
-                【候选人简历】:
-                %s
-                
-                【题库JSON】:
-                %s
-                """, MAX_QUESTIONS_PER_INTERVIEW, resumeContent, questionBankJson);
-
-        // 3. 调用AI服务
-        String aiResponse = llmService.chat(prompt);
-        log.info("AI抽题响应: {}", aiResponse);
-
-        // 4. 解析AI返回的题目ID
-        List<Long> selectedIds = new ArrayList<>();
-        try {
-            JsonNode rootNode = objectMapper.readTree(aiResponse);
-            JsonNode idsNode = rootNode.get("question_ids");
-            if (idsNode != null && idsNode.isArray()) {
-                for (JsonNode idNode : idsNode) {
-                    selectedIds.add(idNode.asLong());
-                }
-            }
-        } catch (JsonProcessingException e) {
-            log.error("解析AI返回的题目ID列表失败", e);
-            return Collections.emptyList(); // 解析失败则返回空列表
-        }
-
-        if (selectedIds.isEmpty()) {
-            return Collections.emptyList();
-        }
-
-        // 5. 根据ID从数据库中获取完整的题目信息，并保持AI选择的顺序
-        List<Question> finalQuestions = questionMapper.selectBatchIds(selectedIds);
-        finalQuestions.sort(Comparator.comparing(q -> selectedIds.indexOf(q.getId()))); // 保持AI返回的顺序
-
-        return finalQuestions;
-    }
-
-    private InterviewSession createSession(String sessionId, InterviewStartRequest request,
-                                           String resumeContent, String selectedQuestionIdsJson) {
-        InterviewSession session = new InterviewSession()
-                .setSessionId(sessionId)
-                .setCandidateName(request.getCandidateName())
-                .setResumeContent(resumeContent)
-                .setSelectedQuestionIds(selectedQuestionIdsJson)
-                .setAiModel("qwen-max") // Hardcoded to qwen-max
-                .setStatus(InterviewSession.Status.ACTIVE.name())
-                .setCurrentQuestionIndex(0);
-
-        sessionMapper.insert(session);
-        return session;
-    }
-
-    private String generateFirstQuestion(InterviewSession session, Question question, String sessionId) {
-        String prompt = String.format("""
-                你是一位专业的技术面试官。现在要开始面试，候选人是 %s。
-                
-                第一个问题是：%s
-                
-                请以友好但专业的语气提出这个问题，可以适当添加一些引导性的话语。
-                """, session.getCandidateName(), question.getContent());
-
-        return this.llmService.chat(prompt, sessionId);
-    }
-
-    private void saveMessage(String sessionId, String messageType, String sender,
-                             String content, Long questionId, int order) {
-        InterviewMessage message = new InterviewMessage()
-                .setSessionId(sessionId)
-                .setMessageType(messageType)
-                .setSender(sender)
-                .setContent(content)
-                .setQuestionId(questionId)
-                .setMessageOrder(order);
-
-        messageMapper.insert(message);
-    }
-
-    /**
-     * 评估答案
      *
-     * @param session    会话数据
+     * @param file 简历文件
-     * @param userAnswer 用户回答
+     * @param dto  开始面试的请求参数
-     * @return 当前问题id
+     * @return 创建的面试会话
      */
-    private Long evaluateAnswer(InterviewSession session, String userAnswer) {
+    InterviewSession startInterview(MultipartFile file, InterviewStartRequest dto) throws IOException;
-        // 根据会话id查询当前会话所有问题
-        List<InterviewQuestionProgress> interviewQuestionProgresses = questionProgressMapper.selectList(
-                new LambdaQueryWrapper<InterviewQuestionProgress>()
-                        .eq(InterviewQuestionProgress::getSessionId, session.getSessionId())
-                        .orderByAsc(InterviewQuestionProgress::getCreatedTime)
-        );
-        if (CollectionUtil.isEmpty(interviewQuestionProgresses)) {
-            throw new RuntimeException("当前会话没有任何可询问的问题！");
-        }
-
-        // 1. 获取当前正在回答的问题
-        InterviewQuestionProgress currentQuestionProgress = null;
-        for (InterviewQuestionProgress interviewQuestionProgress : interviewQuestionProgresses) {
-            if (interviewQuestionProgress.getStatus().equals(InterviewQuestionProgress.Status.ACTIVE.name())) {
-                currentQuestionProgress = interviewQuestionProgress;
-                break;
-            }
-        }
-        if (Objects.isNull(currentQuestionProgress)) {
-            throw new RuntimeException("当前没有正在回答的问题");
-        }
-        Long currentQuestionId = currentQuestionProgress.getQuestionId();
-
-
-        List<String> questionIds = interviewQuestionProgresses.stream()
-                .map(data -> {
-                    return data.getQuestionId().toString();
-                })
-                .collect(Collectors.toList());
-        String join = String.join(",", questionIds);
-        // 2. 构建评估提示
-        String prompt = String.format("""
-                你是一位资深的技术面试官。请根据以下问题和候选人的回答，进行一次专业的评估。
-                
-                要求：
-                1. 对回答的质量进行打分，分数范围为1-5分。
-                2. 给出简洁、专业的评语。
-                3. 提出具体的改进建议以及你认为应该回答的答案。
-                4. 以严格的JSON格式返回，不要包含任何额外的解释文字。格式如下：
-                {
-                 "score": 4.5,
-                 "feedback": "回答基本正确，但可以更深入...",
-                 "suggestions": "可以补充关于XXX方面的知识点...",
-                 "answer": "关于当前问题，您应该这样回答xxx",
-                 "currentQuestionId": xxx
-                }
-                5. 不要返回任何多余字符，请严格按照api接口格式的JSON数据进行返回，不要包含"```json```"
-                6. 如果你认为面试人对当前问题回答不完美，可以继续对当前问题进行补充提问，但不要修改currentQuestionId
-                7. 如果你认为面试人对当前问题回答已经比较好了，或者面试人回答不上来了，请你根据questionIds数据顺序选择下一个问题，并修改currentQuestionId进行返回
-                8. 如果所有问题都已回答完成，请将currentQuestionId设置为0
-                {
-                "questionIds": %s,
-                "currentQuestionId": %s
-                }
-                【面试问题】:
-                %s
-                
-                【候选人回答】:
-                %s
-                """, join, currentQuestionProgress.getQuestionId(), currentQuestionProgress.getQuestionContent(), userAnswer);
-
-        // 3. 调用AI进行评估
-        String aiResponse = llmService.chat(prompt, session.getSessionId());
-        log.info("AI评估响应: {}", aiResponse);
-
-        // 4. 解析AI响应并存储评估结果
-        try {
-            JsonNode rootNode = objectMapper.readTree(aiResponse);
-            InterviewEvaluation evaluation = new InterviewEvaluation()
-                    .setSessionId(session.getSessionId())
-                    .setQuestionId(currentQuestionId)
-                    .setUserAnswer(userAnswer)
-                    .setScore(new java.math.BigDecimal(rootNode.get("score").asText()))
-                    .setAiFeedback(rootNode.get("feedback").asText())
-                    .setEvaluationCriteria(rootNode.get("suggestions").asText()); // 暂时复用这个字段存建议
-            JsonNode currentQuestionId1 = rootNode.get("currentQuestionId");
-            JsonNode aiAnswerNode = rootNode.get("answer");
-            if (Objects.nonNull(currentQuestionId1)) {
-                String text = currentQuestionId1.asText();
-                if (StringUtils.isNoneBlank(text)) {
-                    currentQuestionProgress
-                            .setScore(new BigDecimal(rootNode.get("score").asText()))
-                            .setSuggestions(rootNode.get("suggestions").asText())
-                            .setFeedback(rootNode.get("feedback").asText())
-                            .setAiAnswer(Objects.nonNull(aiAnswerNode) ? aiAnswerNode.asText() : null)
-                            .setUserAnswer(userAnswer)
-                    ;
-                    if (!StrUtil.equals(text, currentQuestionProgress.getQuestionId().toString())) {
-                        currentQuestionProgress.setStatus(InterviewQuestionProgress.Status.COMPLETED.name());
-                        questionProgressMapper.updateById(currentQuestionProgress);
-                        questionProgressMapper.update(
-                                new LambdaUpdateWrapper<InterviewQuestionProgress>()
-                                        .set(InterviewQuestionProgress::getStatus, InterviewQuestionProgress.Status.ACTIVE.name())
-                                        .eq(InterviewQuestionProgress::getSessionId, session.getSessionId())
-                                        .eq(InterviewQuestionProgress::getQuestionId, Long.valueOf(text))
-                        );
-                    } else if (text.equals("0")) {
-                        currentQuestionProgress.setStatus(InterviewQuestionProgress.Status.COMPLETED.name());
-                        questionProgressMapper.updateById(currentQuestionProgress);
-                    }
-                    currentQuestionId = Long.valueOf(text);
-                }
-            }
-            evaluationMapper.insert(evaluation);
-            log.info("成功存储对问题ID {} 的评估结果", currentQuestionId);
-            return currentQuestionId;
-        } catch (Exception e) {
-            log.error("解析或存储AI评估结果失败", e);
-            throw new RuntimeException("解析或存储AI评估结果失败");
-        }
-    }
-
-    private InterviewResponse finishInterview(InterviewSession session) {
-        // 1. 获取本次面试的所有评估数据
-        List<InterviewEvaluation> evaluations = evaluationMapper.selectBySessionId(session.getSessionId());
-
-        // 2. 构建生成最终报告的提示
-        String prompt = buildFinalReportPrompt(session, evaluations);
-
-        // 3. 调用AI生成报告
-        String finalReportJson = llmService.chat(prompt, session.getSessionId());
-        log.info("AI生成的最终面试报告: {}", finalReportJson);
-
-        // 4. 更新会话状态和最终报告
-        session.setStatus(InterviewSession.Status.COMPLETED.name());
-        session.setFinalReport(finalReportJson);
-        sessionMapper.updateById(session);
-
-        // 5. 返回结束信息
-        return new InterviewResponse()
-                .setSessionId(session.getSessionId())
-                .setMessage("面试已结束，感谢您的参与！AI正在生成您的面试报告，请稍后在面试历史中查看。")
-                .setMessageType(InterviewMessage.MessageType.SYSTEM.name())
-                .setSender(InterviewMessage.Sender.SYSTEM.name())
-                .setCurrentQuestionId(null)
-                .setStatus(InterviewSession.Status.COMPLETED.name());
-    }
-
-    private String buildFinalReportPrompt(InterviewSession session, List<InterviewEvaluation> evaluations) {
-        StringBuilder historyBuilder = new StringBuilder();
-        for (InterviewEvaluation eval : evaluations) {
-            Question q = questionMapper.selectById(eval.getQuestionId());
-            historyBuilder.append(String.format("\n【问题】: %s\n【回答】: %s\n【AI单题反馈】: %s\n【AI单题建议】: %s\n【AI单题评分】: %s/5.0\n",
-                    q.getContent(), eval.getUserAnswer(), eval.getAiFeedback(), eval.getEvaluationCriteria(), eval.getScore()));
-        }
-
-        return String.format("""
-                你是一位资深的HR和技术总监。请根据以下候选人的简历、完整的面试问答历史和AI对每一题的初步评估，给出一份全面、专业、有深度的最终面试报告。
-                
-                要求：
-                1. **综合评价**: 对候选人的整体表现给出一个总结性的评语，点出其核心亮点和主要不足。
-                2. **技术能力评估**: 分点阐述候选人在不同技术领域（如Java基础, Spring, 数据库等）的掌握程度。
-                3. **改进建议**: 给出3-5条具体的、可操作的学习和改进建议。
-                4. **综合得分**: 给出一个1-100分的最终综合得分。
-                5. **录用建议**: 给出明确的录用建议（如：强烈推荐、推荐、待考虑、不推荐）。
-                6. 以严格的JSON格式返回，不要包含任何额外的解释文字。格式如下：
-                {
-                  "overallScore": 85,
-                  "overallFeedback": "候选人Java基础扎实，但在高并发场景下的经验有所欠缺...",
-                  "technicalAssessment": {
-                    "Java基础": "掌握良好，对集合框架理解深入。",
-                    "Spring框架": "熟悉基本使用，但对底层原理理解不足。",
-                    "数据库": "能够编写常规SQL，但在索引优化方面知识欠缺。"
-                  },
-                  "suggestions": [
-                    "深入学习Spring AOP和事务管理的实现原理。",
-                    "系统学习MySQL索引优化和查询性能分析。",
-                    "通过实际项目积累高并发处理经验。"
-                  ],
-                  "hiringRecommendation": "推荐"
-                }
-                
-                【候选人简历摘要】:
-                %s
-                
-                【面试问答与评估历史】:
-                %s
-                """, session.getResumeContent(), historyBuilder.toString());
-    }
-
 
     /**
-     * 获取所有面试会话列表
+     * 获取下一个问题
+     *
+     * @param sessionId 会话ID
+     * @return 下一个问题 或 null（如果没有更多问题）
      */
-    public List<InterviewSession> getInterviewSessions() {
+    InterviewMessage getNextQuestion(String sessionId, Long progressId);
-        log.info("Fetching all interview sessions");
-        return sessionMapper.selectList(null); // 实际中可能需要分页
-    }
 
     /**
-     * 获取详细的面试复盘报告
+     * 提交答案并获取AI评估
+     *
+     * @param submitAnswerDTO 提交答案的请求参数
+     * @return 对当前问题的评估和反馈
      */
-    public InterviewReportResponse getInterviewReport(String sessionId) {
+    InterviewQuestionProgress submitAnswer(SubmitAnswerDTO submitAnswerDTO);
-        log.info("Fetching interview report for session id: {}", sessionId);
 
-        InterviewSession session = sessionMapper.selectBySessionId(sessionId);
+    /**
-        if (session == null) {
+     * 结束面试并生成最终报告
-            throw new IllegalArgumentException("找不到ID为 " + sessionId + " 的面试会话。");
+     *
-        }
+     * @param sessionId 会话ID
+     * @return 包含最终报告的面试会话信息
+     */
+    InterviewSession endInterview(String sessionId);
 
-        List<InterviewEvaluation> evaluations = evaluationMapper.selectBySessionId(sessionId);
+    /**
-
+     * 获取面试报告
-        List<InterviewReportResponse.QuestionDetail> questionDetails = evaluations.stream().map(eval -> {
+     * @param sessionId
-            Question question = questionMapper.selectById(eval.getQuestionId());
+     * @return
-            InterviewReportResponse.QuestionDetail detail = new InterviewReportResponse.QuestionDetail();
+     */
-            detail.setQuestionId(eval.getQuestionId());
+    InterviewReportResponse getInterviewReport(String sessionId);
-            detail.setQuestionContent(question != null ? question.getContent() : "题目已不存在");
-            detail.setUserAnswer(eval.getUserAnswer());
-            detail.setAiFeedback(eval.getAiFeedback());
-            detail.setSuggestions(eval.getEvaluationCriteria());
-            detail.setScore(eval.getScore());
-            return detail;
-        }).collect(Collectors.toList());
-
-        InterviewReportResponse report = new InterviewReportResponse();
-        report.setSessionDetails(session);
-        report.setQuestionDetails(questionDetails);
-        List<InterviewMessage> interviewMessages = messageMapper.selectList(
-                new LambdaQueryWrapper<InterviewMessage>()
-                        .eq(InterviewMessage::getSessionId, sessionId)
-        );
-        // 获取当前面试的 问题
-        InterviewQuestionProgress progress = questionProgressMapper.selectOne(
-                new LambdaQueryWrapper<InterviewQuestionProgress>()
-                        .eq(InterviewQuestionProgress::getSessionId, sessionId)
-                        .eq(InterviewQuestionProgress::getStatus, InterviewQuestionProgress.Status.ACTIVE.name())
-                        .last("LIMIT 1")
-        );
-        if (Objects.nonNull(progress)) {
-            report.setCurrentQuestionId(progress.getQuestionId());
-        }
-        report.setMessages(interviewMessages);
-
-        return report;
-    }
-
-    private String getFileExtension(String fileName) {
-        if (fileName == null || fileName.lastIndexOf('.') == -1) {
-            return "";
-        }
-        return fileName.substring(fileName.lastIndexOf('.') + 1).toLowerCase();
-    }
 }
-

src/main/java/com/qingqiu/interview/service/QuestionService.java

@@ -5,7 +5,6 @@ import com.baomidou.mybatisplus.extension.service.IService;
 import com.qingqiu.interview.dto.QuestionOptionsDTO;
 import com.qingqiu.interview.dto.QuestionPageParams;
 import com.qingqiu.interview.entity.Question;
-import com.qingqiu.interview.entity.QuestionCategory;
 import com.qingqiu.interview.vo.QuestionAndCategoryTreeListVO;
 import org.springframework.web.multipart.MultipartFile;
 
@@ -27,4 +26,14 @@ public interface QuestionService extends IService<Question> {
     void useAiCheckQuestionData();
 
     List<QuestionAndCategoryTreeListVO> getTreeListCategory(QuestionOptionsDTO dto);
+
+    /**
+     * 根据技能和难度从本地题库随机选择题目
+     *
+     * @param skills     技能列表
+     * @param difficulty 难度
+     * @param count      题目数量
+     * @return 题目列表
+     */
+    List<Question> selectLocalQuestions(List<String> skills, String difficulty, int count);
 }

src/main/java/com/qingqiu/interview/service/impl/ChatServiceImpl.java

@@ -4,8 +4,9 @@ import cn.hutool.core.collection.CollectionUtil;
 import com.alibaba.dashscope.common.Message;
 import com.alibaba.dashscope.common.Role;
 import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
-import com.qingqiu.interview.common.enums.LLMProvider;
 import com.qingqiu.interview.ai.factory.AIClientManager;
+import com.qingqiu.interview.annotation.AiChatLog;
+import com.qingqiu.interview.common.enums.LLMProvider;
 import com.qingqiu.interview.common.utils.AIUtils;
 import com.qingqiu.interview.dto.ChatDTO;
 import com.qingqiu.interview.dto.InterviewStartRequest;
@@ -41,9 +42,10 @@ public class ChatServiceImpl implements ChatService {
 
     private final AIClientManager aiClientManager;
 
-    private IAiSessionLogService aiSessionLogService;
+    private final IAiSessionLogService aiSessionLogService;
 
     @Override
+    @AiChatLog
     public ChatVO createChat(ChatDTO dto) {
         LLMProvider llmProvider = LLMProvider.fromCode(dto.getAiModel());
         List<Message> messages = new ArrayList<>();
@@ -57,16 +59,13 @@ public class ChatServiceImpl implements ChatService {
                             .orderByAsc(AiSessionLog::getCreatedTime)
             );
             if (CollectionUtil.isNotEmpty(list)) {
-                messages = list.stream().map(data -> {
+                messages.addAll(list.stream().map(data -> {
                     tokens.getAndAdd(AIUtils.getPromptTokens(data.getContent()));
                     return AIUtils.createMessage(data.getRole(), data.getContent());
-                }).toList();
+                }).toList());
             }
 
         }
-        if (CollectionUtil.isEmpty( messages)) {
-            messages = new ArrayList<>();
-        }
         messages.add(AIUtils.createMessage(dto.getRole(), dto.getContent()));
         List<Message> finalMessage = new ArrayList<>();
         // 剪切 10%的消息

src/main/java/com/qingqiu/interview/service/impl/InterviewAiServiceImpl.java

@@ -0,0 +1,227 @@
+package com.qingqiu.interview.service.impl;
+
+import com.alibaba.dashscope.common.Role;
+import com.alibaba.fastjson2.JSON;
+import com.alibaba.fastjson2.JSONObject;
+import com.qingqiu.interview.common.constants.CommonConstant;
+import com.qingqiu.interview.dto.ChatDTO;
+import com.qingqiu.interview.entity.InterviewQuestionProgress;
+import com.qingqiu.interview.entity.InterviewSession;
+import com.qingqiu.interview.entity.Question;
+import com.qingqiu.interview.service.ChatService;
+import com.qingqiu.interview.service.InterviewAiService;
+import com.qingqiu.interview.vo.ChatVO;
+import lombok.RequiredArgsConstructor;
+import lombok.extern.slf4j.Slf4j;
+import org.springframework.beans.factory.annotation.Autowired;
+import org.springframework.context.annotation.Lazy;
+import org.springframework.stereotype.Service;
+
+import java.util.List;
+import java.util.stream.Collectors;
+
+/**
+ * <h1></h1>
+ *
+ * @author qingqiu
+ * @date 2025/9/19 16:49
+ */
+@Slf4j
+@Service
+@RequiredArgsConstructor(onConstructor_ = {@Autowired, @Lazy})
+public class InterviewAiServiceImpl implements InterviewAiService {
+
+    private final ChatService chatService;
+
+    @Override
+    public JSONObject extractSkillsFromResume(String resumeContent) {
+        String prompt = "你是一位资深的IT技术招聘专家。" +
+                "请仔细阅读以下简历内容，并提取出其中所有的关键技术技能。" +
+                "请严格按照以下JSON格式返回，不要添加任何额外的解释或说明：\n" +
+                "{\"skills\": [\"技能1\", \"技能2\", \"...\"]}\n\n" +
+                "简历内容如下：\n" + resumeContent;
+
+        ChatDTO chatDTO = new ChatDTO()
+                .setContent(prompt)
+                .setRole(Role.SYSTEM.getValue())
+                .setDataType(CommonConstant.ONE);
+        ChatVO chatVO = chatService.createChat(chatDTO);
+
+        return JSONObject.parse(chatVO.getContent());
+    }
+
+    @Override
+    public JSONObject generateQuestionsOfAi(String sessionId, List<String> skills, String resumeContent, int count) {
+        String skillsStr = String.join(", ", skills);
+        String prompt = String.format(
+                "你是一位专业的软件开发岗位技术面试官。" +
+                        "请根据候选人的以下技术栈、项目经历、简历内容，生成 %d 道有深度和广度的面试题。" +
+                        "题目应覆盖候选人的主要技术领域，并能考察其解决问题的能力。" +
+                        "请严格按照以下JSON格式返回，question数组中必须包含 %d 个问题对象：\n" +
+                        "{\"questions\": [{\"id\": \"ai-gen-1\", \"content\": \"问题1内容...\"}, {\"id\": \"ai-gen-2\", \"content\": \"问题2内容...\"}]}\n\n" +
+                        "候选人技术栈：%s\n" +
+                        "候选人简历：%s",
+                count, count, skillsStr, resumeContent
+        );
+
+        ChatDTO chatDTO = new ChatDTO()
+                .setSessionId(sessionId)
+                .setContent(prompt)
+                .setRole(Role.SYSTEM.getValue())
+                .setDataType(CommonConstant.ONE);
+        ChatVO chatVO = chatService.createChat(chatDTO);
+        return JSON.parseObject(chatVO.getContent());
+    }
+
+    @Override
+    public JSONObject generateQuestionOfLocal(String sessionId, List<Question> questions, List<String> skills, String resumeContent, int count) {
+        String skillsStr = String.join(", ", skills);
+        // 2. 构建发送给AI的提示
+        String prompt = String.format("""
+                你是一位专业的面试官。请根据以下候选人的技术栈、项目经历、简历内容，从提供的题库中，精心挑选出 %d 道最相关的题目进行面试。
+                题目应覆盖候选人的主要技术领域，并能考察其解决问题的能力。
+                要求：
+                1. 题目必须严格从【题库JSON】中选择。
+                2. 挑选的题目应根据候选人的简历内容来抽取。
+                3. 返回一个只包含所选题目ID的JSON数组，格式为：{"question_ids": [1, 5, 23, ...]}。
+                4. 不要返回任何多余的代码，包括markdown形式的代码，我只需要JSON对象，请严格按照api接口形式返回
+                5. 不要返回任何额外的解释或文字，只返回JSON对象。
+                6. 严格按照前后端分离的接口形式返回JSON数据给我，不要返回"```json```"
+                7. 请保证返回数据的完整性，不要返回不完整的数据，否则我的JSON解析会报错！！！
+                
+                【候选人技术栈】：
+                %s
+                【候选人简历】:
+                [%s]
+                【题库JSON】:
+                %s
+                """, count, skillsStr, resumeContent, JSONObject.toJSONString(questions));
+        ChatDTO chatDTO = new ChatDTO()
+                .setSessionId(sessionId)
+                .setContent(prompt)
+                .setRole(Role.SYSTEM.getValue())
+                .setDataType(CommonConstant.ONE);
+        ChatVO chatVO = chatService.createChat(chatDTO);
+        return JSON.parseObject(chatVO.getContent());
+    }
+
+    @Override
+    public JSONObject evaluateAnswer(String sessionId, String question, String userAnswer, List<InterviewQuestionProgress> context) {
+        // 构建上下文历史
+        String history = context.stream()
+                .map(p -> String.format("Q: %s\nA: %s", p.getQuestionContent(), p.getUserAnswer()))
+                .collect(Collectors.joining("\n---\n"));
+
+        String prompt = "你是一位资深的技术面试官，以严格和深入著称。" +
+                "你需要评估候选人对以下问题的回答。请注意：\n" +
+                "1. 如果回答模糊、不完整或有错误，你可以提出一个具体的追问问题（followUpQuestion）来深入考察，此时'continueAsking'应为true。\n" +
+                "2. 如果回答得很好，则'continueAsking'为false，'followUpQuestion'为空字符串。\n" +
+                "3. 'score'范围为0-100分。\n" +
+                "4. 'feedback'和'suggestions'需要给出专业、有建设性的意见。\n" +
+                "5. 追问最好有限制，不要无限制的向下追问，注意追问是支线而非主线！追问至多3个问题，之后必须切回主线\n" +
+                "请严格按照以下JSON格式返回，不要有任何额外说明：\n" +
+                "{\"feedback\": \"...\", \"suggestions\": \"...\", \"aiAnswer\": \"...\", \"score\": 85.5, \"continueAsking\": false, \"followUpQuestion\": \"...\"}\n\n" +
+                "面试历史上下文：\n" + history + "\n\n" +
+                "当前问题：\n" + question + "\n\n" +
+                "候选人回答：\n" + userAnswer;
+
+        ChatDTO chatDTO = new ChatDTO()
+                .setSessionId(sessionId)
+                .setContent(prompt)
+                .setRole(Role.SYSTEM.getValue())
+                .setDataType(CommonConstant.ONE);
+        ChatVO chatVO = chatService.createChat(chatDTO);
+        return JSON.parseObject(chatVO.getContent());
+    }
+
+    @Override
+    public JSONObject generateFinalReport(InterviewSession session, List<InterviewQuestionProgress> progressList) {
+//        String transcript = progressList.stream()
+//                .map(p -> String.format("问题: %s\n回答: %s\nAI评分: %.1f\nAI反馈: %s\n",
+//                        p.getQuestionContent(), p.getUserAnswer(), p.getScore(), p.getFeedback()))
+//                .collect(Collectors.joining("\n-----------------\n"));
+
+//        String prompt = "你是一位经验丰富的招聘经理。" +
+//                "请根据以下完整的面试记录，为候选人生成一份综合评估报告。" +
+//                "报告需要包括一个总分（overallScore），简明扼要的总结（summary），以及候选人的优点（strengths）和待提升点（weaknesses）。" +
+//                "请严格按照以下JSON格式返回：\n" +
+//                "{\"overallScore\": 88.0, \"summary\": \"...\", \"strengths\": [\"...\"], \"weaknesses\": [\"...\"]}\n\n" +
+//                "候选人姓名：" + session.getCandidateName() + "\n" +
+//                "面试完整记录：\n" + transcript;
+
+        String prompt = buildFinalReportPrompt(session, progressList);
+
+        ChatDTO chatDTO = new ChatDTO()
+                .setRole(Role.SYSTEM.getValue())
+                .setDataType(CommonConstant.ONE)
+                .setContent(prompt);
+        ChatVO chatVO = chatService.createChat(chatDTO);
+        return JSON.parseObject(chatVO.getContent());
+    }
+
+    private String buildFinalReportPrompt(InterviewSession session, List<InterviewQuestionProgress> progressList) {
+        StringBuilder historyBuilder = new StringBuilder();
+        for (InterviewQuestionProgress progress : progressList) {
+            historyBuilder.append(
+                    String.format("\n【问题】: %s\n【回答】: %s\n【AI单题反馈】: %s\n【AI单题建议】: %s\n【AI单题评分】: %s/5.0\n",
+                    progress.getQuestionContent(),
+                            progress.getUserAnswer(),
+                            progress.getFeedback(),
+                            progress.getSuggestions(),
+                            progress.getScore()
+                    )
+            );
+        }
+
+        return String.format("""
+                你是一位资深的HR和技术总监。请根据以下候选人的简历、完整的面试问答历史和AI对每一题的初步评估，给出一份全面、专业、有深度的最终面试报告。
+                
+                要求：
+                1. **综合评价**: 对候选人的整体表现给出一个总结性的评语，点出其核心亮点和主要不足。
+                2. **技术能力评估**: 分点阐述候选人在不同技术领域（如Java基础, Spring, 数据库等）的掌握程度。
+                3. **改进建议**: 给出3-5条具体的、可操作的学习和改进建议。
+                4. **综合得分**: 给出一个1-100分的最终综合得分。
+                5. **录用建议**: 给出明确的录用建议（如：强烈推荐、推荐、待考虑、不推荐）。
+                6. 以严格的JSON格式返回，不要包含任何额外的解释文字。格式如下：
+                {
+                  "overallScore": 85,
+                  "overallFeedback": "候选人Java基础扎实，但在高并发场景下的经验有所欠缺...",
+                  "technicalAssessment": {
+                    "Java基础": "掌握良好，对集合框架理解深入。",
+                    "Spring框架": "熟悉基本使用，但对底层原理理解不足。",
+                    "数据库": "能够编写常规SQL，但在索引优化方面知识欠缺。"
+                  },
+                  "suggestions": [
+                    "深入学习Spring AOP和事务管理的实现原理。",
+                    "系统学习MySQL索引优化和查询性能分析。",
+                    "通过实际项目积累高并发处理经验。"
+                  ],
+                  "hiringRecommendation": "推荐"
+                }
+                
+                【候选人简历摘要】:
+                %s
+                
+                【面试问答与评估历史】:
+                %s
+                """, session.getResumeContent(), historyBuilder.toString());
+    }
+
+    @Override
+    public String generateFirstQuestion(String sessionId, String candidateName, String questionContent) {
+        String prompt = String.format("""
+                你是一位专业的技术面试官。现在要开始面试，候选人是 %s。
+                
+                第一个问题是：%s
+                
+                请以友好但专业的语气提出这个问题，可以适当添加一些引导性的话语。
+                """, candidateName, questionContent);
+        ChatDTO chatDTO = new ChatDTO()
+                .setSessionId(sessionId)
+                .setRole(Role.SYSTEM.getValue())
+                .setDataType(CommonConstant.ONE)
+                .setContent(prompt);
+        ChatVO chatVO = chatService.createChat(chatDTO);
+        return chatVO.getContent();
+    }
+}

src/main/java/com/qingqiu/interview/service/impl/InterviewMessageServiceImpl.java

@@ -0,0 +1,23 @@
+package com.qingqiu.interview.service.impl;
+
+import com.baomidou.mybatisplus.extension.service.impl.ServiceImpl;
+import com.qingqiu.interview.entity.InterviewMessage;
+import com.qingqiu.interview.mapper.InterviewMessageMapper;
+import com.qingqiu.interview.service.InterviewMessageService;
+import lombok.RequiredArgsConstructor;
+import lombok.extern.slf4j.Slf4j;
+import org.springframework.beans.factory.annotation.Autowired;
+import org.springframework.context.annotation.Lazy;
+import org.springframework.stereotype.Service;
+
+/**
+ * <h1></h1>
+ *
+ * @author qingqiu
+ * @date 2025/9/21 12:00
+ */
+@Slf4j
+@Service
+@RequiredArgsConstructor(onConstructor_ = {@Autowired, @Lazy})
+public class InterviewMessageServiceImpl extends ServiceImpl<InterviewMessageMapper, InterviewMessage> implements InterviewMessageService {
+}

src/main/java/com/qingqiu/interview/service/impl/InterviewQuestionProgressServiceImpl.java

@@ -9,8 +9,10 @@ import com.qingqiu.interview.mapper.InterviewQuestionProgressMapper;
 import com.qingqiu.interview.service.IInterviewQuestionProgressService;
 import org.apache.commons.lang3.StringUtils;
 import org.springframework.stereotype.Service;
+import org.springframework.transaction.annotation.Transactional;
 
 import java.util.Arrays;
+import java.util.Objects;
 
 /**
  * <p>
@@ -40,4 +42,37 @@ public class InterviewQuestionProgressServiceImpl extends ServiceImpl<InterviewQ
                         .orderByDesc(InterviewQuestionProgress::getCreatedTime)
                 );
     }
+
+    @Override
+    @Transactional(rollbackFor = Exception.class)
+    public InterviewQuestionProgress getNextQuestion(String sessionId) {
+        // 查找状态为“进行中”的问题
+        InterviewQuestionProgress activeQuestion = baseMapper.selectOne(
+                new LambdaQueryWrapper<InterviewQuestionProgress>()
+                        .eq(InterviewQuestionProgress::getSessionId, sessionId)
+                        .eq(InterviewQuestionProgress::getStatus, InterviewQuestionProgress.Status.ACTIVE.name())
+                        .orderByAsc(InterviewQuestionProgress::getId)
+                        .last("LIMIT 1")
+        );
+        if (Objects.nonNull(activeQuestion)) {
+            return activeQuestion;
+        }
+        // 1. 查找第一个处于“默认”状态的问题
+        LambdaQueryWrapper<InterviewQuestionProgress> queryWrapper = new LambdaQueryWrapper<>();
+        queryWrapper.eq(InterviewQuestionProgress::getSessionId, sessionId)
+                .eq(InterviewQuestionProgress::getStatus, InterviewQuestionProgress.Status.DEFAULT.name())
+                .orderByAsc(InterviewQuestionProgress::getId) // 按插入顺序
+                .last("LIMIT 1");
+        InterviewQuestionProgress nextQuestion = baseMapper.selectOne(queryWrapper);
+
+        if (nextQuestion == null) {
+            // 没有更多的问题了
+            return null;
+        }
+
+        // 2. 将问题状态更新为“进行中”
+        nextQuestion.setStatus(InterviewQuestionProgress.Status.ACTIVE.name());
+        baseMapper.updateById(nextQuestion);
+        return nextQuestion;
+    }
 }

src/main/java/com/qingqiu/interview/service/impl/InterviewServiceImpl.java

@@ -0,0 +1,414 @@
+package com.qingqiu.interview.service.impl;
+
+import cn.hutool.core.collection.CollectionUtil;
+import cn.hutool.core.io.file.FileNameUtil;
+import com.alibaba.fastjson2.JSONArray;
+import com.alibaba.fastjson2.JSONObject;
+import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
+import com.baomidou.mybatisplus.extension.service.impl.ServiceImpl;
+import com.qingqiu.interview.common.enums.DocumentParserProvider;
+import com.qingqiu.interview.common.ex.ApiException;
+import com.qingqiu.interview.dto.InterviewReportResponse;
+import com.qingqiu.interview.dto.InterviewStartRequest;
+import com.qingqiu.interview.dto.SubmitAnswerDTO;
+import com.qingqiu.interview.entity.*;
+import com.qingqiu.interview.mapper.InterviewEvaluationMapper;
+import com.qingqiu.interview.mapper.InterviewMessageMapper;
+import com.qingqiu.interview.mapper.InterviewSessionMapper;
+import com.qingqiu.interview.service.IInterviewQuestionProgressService;
+import com.qingqiu.interview.service.InterviewAiService;
+import com.qingqiu.interview.service.InterviewService;
+import com.qingqiu.interview.service.QuestionService;
+import com.qingqiu.interview.service.parser.DocumentParser;
+import com.qingqiu.interview.service.parser.DocumentParserManager;
+import com.qingqiu.interview.vo.QuestionAndCategoryTreeListVO;
+import lombok.RequiredArgsConstructor;
+import lombok.extern.slf4j.Slf4j;
+import org.apache.commons.lang3.StringUtils;
+import org.springframework.beans.factory.annotation.Autowired;
+import org.springframework.context.annotation.Lazy;
+import org.springframework.stereotype.Service;
+import org.springframework.transaction.annotation.Transactional;
+import org.springframework.web.multipart.MultipartFile;
+
+import java.io.IOException;
+import java.util.ArrayList;
+import java.util.List;
+import java.util.Objects;
+import java.util.UUID;
+import java.util.stream.Collectors;
+
+/**
+ * <h1></h1>
+ *
+ * @author qingqiu
+ * @date 2025/9/19 16:07
+ */
+@Slf4j
+@Service
+@RequiredArgsConstructor(onConstructor_ = {@Autowired, @Lazy})
+public class InterviewServiceImpl extends ServiceImpl<InterviewSessionMapper, InterviewSession> implements InterviewService {
+
+    private final QuestionService questionService;
+
+    private final IInterviewQuestionProgressService progressService;
+
+    private final InterviewEvaluationMapper evaluationMapper;
+
+    private final InterviewMessageMapper messageMapper;
+
+    private final InterviewAiService aiService;
+
+    private final DocumentParserManager documentParserManager;
+
+    @Override
+    @Transactional(rollbackFor = Exception.class)
+    public InterviewSession startInterview(MultipartFile file, InterviewStartRequest dto) throws IOException {
+        // 1. 创建并保存会话主记录
+        String sessionId = UUID.randomUUID().toString().replace("-", "");
+        String resumeContent = parseResume(file);
+        InterviewSession session = new InterviewSession();
+
+        session.setSessionId(sessionId);
+        session.setCandidateName(dto.getCandidateName());
+        session.setResumeContent(resumeContent);
+        session.setAiModel(dto.getAiModel());
+        session.setStatus(InterviewSession.Status.ACTIVE.name());
+        session.setTotalQuestions(dto.getTotalQuestions());
+        session.setModel(dto.getModel());
+        this.baseMapper.insert(session); // 先插入以获取ID
+
+        // 2. 调用AI服务从简历提取技能
+        JSONObject skillsJson = aiService.extractSkillsFromResume(resumeContent);
+        // ---> 解析AI返回的JSON数据，获取技能列表 <---
+        List<String> skills = skillsJson.getList("skills", String.class);
+        session.setExtractedSkills(skillsJson.toJSONString());
+
+        // 3. 准备面试问题（本地 + AI生成）
+        if (dto.getModel().equals("local")) {
+            localGenerateQuestions(session, skills, dto.getSelectedNodes());
+        } else {
+            aiGenerateQuestions(session, skills);
+        }
+
+        // 4. 更新会话信息
+        this.baseMapper.updateById(session);
+        InterviewQuestionProgress nextQuestion = progressService.getNextQuestion(sessionId);
+        aiService.generateFirstQuestion(session.getSessionId(), session.getCandidateName(), nextQuestion.getQuestionContent());
+        saveMessage(sessionId,
+                InterviewMessage.MessageType.QUESTION.name(),
+                InterviewMessage.Sender.AI.name(),
+                nextQuestion.getQuestionContent(),
+                nextQuestion.getId()
+        );
+        return session;
+    }
+
+
+    private void aiGenerateQuestions(InterviewSession session, List<String> skills) {
+        List<InterviewQuestionProgress> progressList = new ArrayList<>();
+        JSONObject aiQuestionsJson = aiService.generateQuestionsOfAi(
+                session.getSessionId(),
+                skills,
+                session.getResumeContent(),
+                session.getTotalQuestions()
+        );
+        // ---> 解析AI返回的JSON数据，获取问题列表 <---
+        JSONArray questions = aiQuestionsJson.getJSONArray("questions");
+        if (questions != null) {
+            questions.forEach(item -> {
+                JSONObject q = (JSONObject) item;
+                InterviewQuestionProgress progress = new InterviewQuestionProgress();
+                progress.setSessionId(session.getSessionId());
+                progress.setQuestionId(0L); // AI生成的问题没有本地ID
+                // ---> 解析单个问题内容 <---
+                progress.setQuestionContent(q.getString("content"));
+                progress.setStatus(InterviewQuestionProgress.Status.DEFAULT.name());
+                progressList.add(progress);
+            });
+        }
+        // 批量保存问题进度
+        if (CollectionUtil.isNotEmpty(progressList)) {
+            progressList.forEach(progressService::save);
+        }
+    }
+
+
+    private void localGenerateQuestions(InterviewSession session,
+                                        List<String> skills,
+                                        List<QuestionAndCategoryTreeListVO> selectedNodes) {
+        List<Question> localQuestionDataList = new ArrayList<>();
+        // 如果用户选择了题目 则使用用户选择的题目 否则直接使用全部的题目
+        if (CollectionUtil.isNotEmpty(selectedNodes)) {
+            List<QuestionAndCategoryTreeListVO> question = selectedNodes.stream()
+                    .filter(node -> node.getType().equals("question"))
+                    .toList();
+            if (CollectionUtil.isNotEmpty(question)) {
+                localQuestionDataList = question.stream()
+                        .map(node -> {
+                            return new Question().setId(node.getId()).setContent(node.getName());
+                        }).toList();
+
+            }
+        }
+        if (CollectionUtil.isEmpty(localQuestionDataList)) {
+            localQuestionDataList = questionService.list(
+                    new LambdaQueryWrapper<Question>()
+                            .select(Question::getId, Question::getContent)
+            );
+        }
+        // ai调用返回的内容进行提取
+        JSONObject jsonObject = aiService.generateQuestionOfLocal(
+                session.getSessionId(),
+                localQuestionDataList,
+                skills,
+                session.getResumeContent(),
+                session.getTotalQuestions()
+        );
+        JSONArray questionIds = jsonObject.getJSONArray("question_ids");
+        List<Long> list = questionIds.toList(Long.class);
+        // 查询返回的内容 并将其保存为问题进度的相关数据
+        List<Question> questionList = questionService.list(
+                new LambdaQueryWrapper<Question>()
+                        .in(Question::getId, list)
+        );
+        List<InterviewQuestionProgress> progressList = new ArrayList<>();
+        questionList.forEach(q -> {
+            InterviewQuestionProgress progress = new InterviewQuestionProgress();
+            progress.setSessionId(session.getSessionId());
+            progress.setQuestionId(q.getId());
+            progress.setQuestionContent(q.getContent());
+            progress.setStatus(InterviewQuestionProgress.Status.DEFAULT.name());
+            progressList.add(progress);
+        });
+        // 批量保存问题进度
+        if (CollectionUtil.isNotEmpty(progressList)) {
+            progressList.forEach(progressService::save);
+        }
+
+    }
+
+
+    @Override
+    @Transactional(rollbackFor = Exception.class)
+    public InterviewMessage getNextQuestion(String sessionId, Long progressId) {
+
+        // 获取下一个问题
+        InterviewQuestionProgress nextQuestion = progressService.getNextQuestion(sessionId);
+        if (Objects.isNull(nextQuestion)) {
+            return null;
+        }
+        // 判断是否在interview_message中存在
+        InterviewMessage interviewMessage = messageMapper.selectOne(
+                new LambdaQueryWrapper<InterviewMessage>()
+                        .eq(InterviewMessage::getQuestionProgressId, nextQuestion.getId())
+                        .orderByAsc(InterviewMessage::getId)
+                        .last("LIMIT 1")
+        );
+        if (Objects.isNull(interviewMessage)) {
+            InterviewQuestionProgress prevQuestion = progressService.getById(progressId);
+
+            // 格式化返回的内容
+            StringBuilder sb = new StringBuilder();
+            if (StringUtils.isNotBlank(prevQuestion.getFeedback())) {
+                sb.append(prevQuestion.getFeedback()).append("\n");
+            }
+            if (StringUtils.isNotBlank(prevQuestion.getSuggestions())) {
+                sb.append(prevQuestion.getSuggestions()).append("\n");
+            }
+            if (StringUtils.isNotBlank(prevQuestion.getAiAnswer())) {
+                sb.append(prevQuestion.getAiAnswer()).append("\n");
+            }
+            sb.append(nextQuestion.getQuestionContent());
+
+            interviewMessage = saveMessage(sessionId,
+                    InterviewMessage.MessageType.QUESTION.name(),
+                    InterviewMessage.Sender.AI.name(),
+                    sb.toString(),
+                    nextQuestion.getId()
+            );
+        }
+
+        return interviewMessage;
+    }
+
+    @Override
+    @Transactional(rollbackFor = Exception.class)
+    public InterviewQuestionProgress submitAnswer(SubmitAnswerDTO dto) {
+        // 1. 查询当前正在进行的这个问题
+        InterviewQuestionProgress currentProgress = progressService.getById(dto.getProgressId());
+        if (Objects.isNull(currentProgress) || !InterviewQuestionProgress.Status.ACTIVE.name().equals(currentProgress.getStatus())) {
+            throw new ApiException("问题进度不存在或已处理");
+        }
+        currentProgress.setUserAnswer(dto.getAnswer());
+        // 存储消息
+        saveMessage(dto.getSessionId(),
+                InterviewMessage.MessageType.ANSWER.name(),
+                InterviewMessage.Sender.USER.name(),
+                dto.getAnswer(),
+                currentProgress.getId()
+        );
+
+        // 2. 调用AI服务评估回答
+        List<InterviewQuestionProgress> context = progressService.list(
+                new LambdaQueryWrapper<InterviewQuestionProgress>()
+                        .eq(InterviewQuestionProgress::getSessionId, currentProgress.getSessionId())
+                        .eq(InterviewQuestionProgress::getStatus, InterviewQuestionProgress.Status.COMPLETED.name())
+                        .orderByAsc(InterviewQuestionProgress::getId)
+        );
+        JSONObject evalResult = aiService.evaluateAnswer(
+                currentProgress.getSessionId(),
+                currentProgress.getQuestionContent(),
+                dto.getAnswer(),
+                context
+        );
+
+        // 3. ---> 解析AI返回的JSON评估结果并存入数据库 <---
+        currentProgress.setFeedback(evalResult.getString("feedback"));
+        currentProgress.setSuggestions(evalResult.getString("suggestions"));
+        currentProgress.setAiAnswer(evalResult.getString("aiAnswer"));
+        currentProgress.setScore(evalResult.getBigDecimal("score"));
+        currentProgress.setStatus(InterviewQuestionProgress.Status.COMPLETED.name());
+        progressService.updateById(currentProgress);
+
+        // 4. 将单题评估结果存入 evaluation 表用于分析
+        saveEvaluationRecord(currentProgress, evalResult);
+
+
+        // 5. ---> 解析AI的是否追问判断，并处理追问逻辑 <---
+        if (evalResult.getBooleanValue("continueAsking", false)) {
+            // 创建一个新的、状态为ACTIVE的追问问题
+            InterviewQuestionProgress followUp = new InterviewQuestionProgress();
+            followUp.setSessionId(currentProgress.getSessionId());
+            followUp.setQuestionId(0L); // 追问问题没有本地ID
+            followUp.setQuestionContent(evalResult.getString("followUpQuestion"));
+            followUp.setStatus(InterviewQuestionProgress.Status.ACTIVE.name()); // 直接设为激活状态，作为下一个问题
+            progressService.save(followUp);
+            return followUp; // 将这个新的追问问题返回给前端
+        }
+
+        return currentProgress;
+    }
+
+    private void saveEvaluationRecord(InterviewQuestionProgress progress, JSONObject evalResult) {
+        InterviewEvaluation evaluation = new InterviewEvaluation();
+        evaluation.setSessionId(progress.getSessionId());
+        evaluation.setQuestionId(progress.getQuestionId());
+        evaluation.setUserAnswer(progress.getUserAnswer());
+        // ---> 解析AI评估结果并存入分析表 <---
+        evaluation.setAiFeedback(evalResult.getString("feedback"));
+        evaluation.setScore(evalResult.getBigDecimal("score"));
+        evaluationMapper.insert(evaluation);
+    }
+
+    @Override
+    public InterviewSession endInterview(String sessionId) {
+        InterviewSession session = this.getOne(new LambdaQueryWrapper<InterviewSession>()
+                .eq(InterviewSession::getSessionId, sessionId));
+        if (session == null) throw new RuntimeException("会话不存在");
+
+        List<InterviewQuestionProgress> completedProgresses = progressService.list(
+                new LambdaQueryWrapper<InterviewQuestionProgress>()
+                        .eq(InterviewQuestionProgress::getSessionId, sessionId)
+                        .eq(InterviewQuestionProgress::getStatus, InterviewQuestionProgress.Status.COMPLETED.name())
+        );
+        if (CollectionUtil.isEmpty(completedProgresses)) {
+            session.setStatus(InterviewSession.Status.COMPLETED.name());
+            this.baseMapper.updateById(session);
+            return session;
+        }
+
+        // 2. 调用AI服务生成最终报告
+        JSONObject finalReportJson = aiService.generateFinalReport(session, completedProgresses);
+
+        // 3. ---> 解析AI返回的最终报告JSON，更新会话状态 <---
+        session.setStatus(InterviewSession.Status.COMPLETED.name());
+        session.setScore(finalReportJson.getBigDecimal("overallScore"));
+        session.setFinalReport(finalReportJson.toJSONString());
+        this.baseMapper.updateById(session);
+
+        return session;
+    }
+
+
+    /**
+     * 获取详细的面试复盘报告
+     */
+    @Override
+    public InterviewReportResponse getInterviewReport(String sessionId) {
+        log.info("Fetching interview report for session id: {}", sessionId);
+
+        InterviewSession session = getOne(
+                new LambdaQueryWrapper<InterviewSession>()
+                        .eq(InterviewSession::getSessionId, sessionId)
+                        .last("LIMIT 1")
+        );
+        if (session == null) {
+            throw new IllegalArgumentException("找不到ID为 " + sessionId + " 的面试会话。");
+        }
+        List<InterviewQuestionProgress> progressList = progressService.list(
+                new LambdaQueryWrapper<InterviewQuestionProgress>()
+                        .eq(InterviewQuestionProgress::getSessionId, sessionId)
+                        .orderByAsc(InterviewQuestionProgress::getUpdatedTime)
+        );
+
+
+        List<InterviewReportResponse.QuestionDetail> questionDetails = progressList.stream().map(progress -> {
+            InterviewReportResponse.QuestionDetail detail = new InterviewReportResponse.QuestionDetail();
+            detail.setQuestionId(progress.getQuestionId());
+            detail.setQuestionContent(progress.getQuestionContent());
+            detail.setUserAnswer(progress.getUserAnswer());
+            detail.setAiFeedback(progress.getFeedback());
+            detail.setSuggestions(progress.getSuggestions());
+            detail.setScore(progress.getScore());
+            return detail;
+        }).collect(Collectors.toList());
+
+        InterviewReportResponse report = new InterviewReportResponse();
+        report.setSessionDetails(session);
+        report.setQuestionDetails(questionDetails);
+        List<InterviewMessage> interviewMessages = messageMapper.selectList(
+                new LambdaQueryWrapper<InterviewMessage>()
+                        .eq(InterviewMessage::getSessionId, sessionId)
+        );
+        // 获取当前面试的 问题
+        InterviewQuestionProgress progress = progressService.getOne(
+                new LambdaQueryWrapper<InterviewQuestionProgress>()
+                        .eq(InterviewQuestionProgress::getSessionId, sessionId)
+                        .eq(InterviewQuestionProgress::getStatus, InterviewQuestionProgress.Status.ACTIVE.name())
+                        .last("LIMIT 1")
+        );
+        if (Objects.nonNull(progress)) {
+            report.setCurrentQuestionId(progress.getQuestionId());
+        }
+        report.setMessages(interviewMessages);
+
+        return report;
+    }
+
+
+    private String parseResume(MultipartFile resume) throws IOException {
+        // 获取文件扩展名
+        String extName = FileNameUtil.extName(resume.getOriginalFilename());
+        // 1. 获取简历解析器
+        DocumentParser parser = documentParserManager.getParser(DocumentParserProvider.fromCode(extName));
+        // 2. 解析简历
+        return parser.parse(resume.getInputStream());
+    }
+
+
+    private InterviewMessage saveMessage(String sessionId, String messageType, String sender,
+                             String content, Long questionId) {
+        int nextOrder = messageMapper.selectMaxOrderBySessionId(sessionId) + 1;
+        InterviewMessage message = new InterviewMessage()
+                .setSessionId(sessionId)
+                .setMessageType(messageType)
+                .setSender(sender)
+                .setContent(content)
+                .setQuestionProgressId(questionId)
+                .setMessageOrder(nextOrder);
+
+        messageMapper.insert(message);
+        return message;
+    }
+}

src/main/java/com/qingqiu/interview/service/impl/QuestionServiceImpl.java

@@ -31,10 +31,7 @@ import org.springframework.transaction.annotation.Transactional;
 import org.springframework.web.multipart.MultipartFile;
 
 import java.io.IOException;
-import java.util.ArrayList;
+import java.util.*;
-import java.util.List;
-import java.util.Map;
-import java.util.Objects;
 import java.util.stream.Collectors;
 
 @Service
@@ -242,6 +239,7 @@ public class QuestionServiceImpl extends ServiceImpl<QuestionMapper, Question> i
             log.info("根节点题目总数: {}", i);
             QuestionAndCategoryTreeListVO rootVO = new QuestionAndCategoryTreeListVO();
             rootVO.setId(0L);
+            rootVO.setNodeKey(UUID.randomUUID().toString().replace("-", ""));
             rootVO.setName("全部题目");
             rootVO.setType("root");
             rootVO.setChildren(voList);
@@ -252,6 +250,20 @@ public class QuestionServiceImpl extends ServiceImpl<QuestionMapper, Question> i
         return voList;
     }
 
+    @Override
+    public List<Question> selectLocalQuestions(List<String> skills, String difficulty, int count) {
+        // TODO: 实现更智能的选题逻辑，例如：
+        // 1. 根据技能(skills)匹配题目的`tags`或`category_name`。
+        // 2. 使用`difficulty`进行筛选。
+        // 3. 随机选取`count`道题目。
+        // 4. 此处仅为简单示例，随机获取指定数量的题目。
+
+        LambdaQueryWrapper<Question> queryWrapper = new LambdaQueryWrapper<>();
+        queryWrapper.last("ORDER BY RAND() LIMIT " + count);
+
+        return this.baseMapper.selectList(queryWrapper);
+    }
+
     /**
      * 将QuestionCategory列表转换为QuestionAndCategoryTreeListVO列表，并整合题目数据
      *
@@ -302,6 +314,7 @@ public class QuestionServiceImpl extends ServiceImpl<QuestionMapper, Question> i
         vo.setName(category.getName());
         vo.setType("category");
         vo.setCount(0);
+        vo.setNodeKey(UUID.randomUUID().toString().replace("-", ""));
 
         // 处理子节点（包括子分类和题目）
         List<QuestionAndCategoryTreeListVO> childrenVOs = new ArrayList<>();
@@ -355,6 +368,7 @@ public class QuestionServiceImpl extends ServiceImpl<QuestionMapper, Question> i
         vo.setChildren(List.of());
         vo.setType("question");
         vo.setCount(0); // 题目节点没有子节点，count设为0
+        vo.setNodeKey(UUID.randomUUID().toString().replace("-", ""));
 
         return vo;
     }

src/main/java/com/qingqiu/interview/vo/ChatVO.java

@@ -11,7 +11,7 @@ import lombok.experimental.Accessors;
  * @date 2025/9/18 12:56
  */
 @Data
-@Accessors
+@Accessors(chain = true)
 @Builder
 public class ChatVO {
 

src/main/java/com/qingqiu/interview/vo/QuestionAndCategoryTreeListVO.java

@@ -19,6 +19,8 @@ public class QuestionAndCategoryTreeListVO implements Serializable {
 
     private Long id;
 
+    private String nodeKey;
+
     private String name;
     /**
      * category:分类

src/main/resources/logback-spring.xml

@@ -0,0 +1,61 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<configuration>
+
+    <property name="LOG_PATH" value="./logs"/>
+    <property name="LOG_FILE_NAME" value="interview"/>
+
+    <appender name="CONSOLE" class="ch.qos.logback.core.ConsoleAppender">
+        <encoder>
+            <pattern>%d{yyyy-MM-dd HH:mm:ss.SSS} [%thread] %highlight(%-5level) %cyan(%logger{36}) - %msg%n</pattern>
+        </encoder>
+    </appender>
+
+    <appender name="FILE" class="ch.qos.logback.core.rolling.RollingFileAppender">
+        <file>${LOG_PATH}/${LOG_FILE_NAME}.log</file>
+        <rollingPolicy class="ch.qos.logback.core.rolling.TimeBasedRollingPolicy">
+            <fileNamePattern>${LOG_PATH}/${LOG_FILE_NAME}.%d{yyyy-MM-dd}.%i.log.gz</fileNamePattern>
+            <maxHistory>30</maxHistory>
+            <timeBasedFileNamingAndTriggeringPolicy class="ch.qos.logback.core.rolling.SizeAndTimeBasedFNATP">
+                <maxFileSize>100MB</maxFileSize>
+            </timeBasedFileNamingAndTriggeringPolicy>
+        </rollingPolicy>
+        <encoder>
+            <pattern>%d{yyyy-MM-dd HH:mm:ss.SSS} [%thread] %-5level %logger{36} - %msg%n</pattern>
+            <charset>UTF-8</charset>
+        </encoder>
+    </appender>
+
+    <appender name="ERROR_FILE" class="ch.qos.logback.core.rolling.RollingFileAppender">
+        <file>${LOG_PATH}/error.log</file>
+        <rollingPolicy class="ch.qos.logback.core.rolling.TimeBasedRollingPolicy">
+            <fileNamePattern>${LOG_PATH}/error.%d{yyyy-MM-dd}.log.gz</fileNamePattern>
+            <maxHistory>30</maxHistory>
+        </rollingPolicy>
+        <encoder>
+            <pattern>%d{yyyy-MM-dd HH:mm:ss.SSS} [%thread] %-5level %logger{36} - %msg%n</pattern>
+            <charset>UTF-8</charset>
+        </encoder>
+        <filter class="ch.qos.logback.classic.filter.ThresholdFilter">
+            <level>ERROR</level>
+        </filter>
+    </appender>
+
+    <appender name="ASYNC_FILE" class="ch.qos.logback.classic.AsyncAppender">
+        <queueSize>256</queueSize>
+        <discardingThreshold>0</discardingThreshold>
+        <appender-ref ref="FILE"/>
+    </appender>
+
+    <root level="INFO">
+        <appender-ref ref="CONSOLE"/>
+        <appender-ref ref="ASYNC_FILE"/>
+        <appender-ref ref="ERROR_FILE"/>
+    </root>
+
+    <logger name="com.qingqiu.interview" level="DEBUG" additivity="false">
+        <appender-ref ref="CONSOLE"/>
+        <appender-ref ref="ASYNC_FILE"/>
+        <appender-ref ref="ERROR_FILE"/>
+    </logger>
+
+</configuration>