优化面试反馈功能：1. 修复问题检测逻辑，支持中英文问号 2. 增强错误处理和日志记录 3. 前端添加返回面试按钮
This commit is contained in:
parent
66821e2a32
commit
d8bc80cba0
@ -137,5 +137,5 @@ MIT License
|
||||
项目初期，我们没有急着用trae写代码，而是先琢磨“为什么做”。身边同学求职时总抱怨：通用题库的问题和自己的简历不沾边，练完也没人指点改进。这让我们意识到，工具再便捷，没有精准需求导向就是空谈。我们决定不做简单的题库搬运，而是聚焦“个性化”，让AI能读懂简历、给出针对性反馈——这一定位成了后续开发的核心方向。
|
||||
用trae开发的过程，更像是一场“想法落地的试炼”。这款软件降低了编程门槛，让我们不用纠结复杂的语法，能专注于功能逻辑的打磨。但真正的挑战不在代码编写，而在如何让工具“懂用户”。比如生成个性化面试问题时，我们反复调试prompt逻辑，引导AI从简历中抓取项目细节、技能亮点，甚至职业空白期等关键信息，再结合职位类型精准提问。一开始AI会给出通用问题，我们就不断优化指令，加入“追问数据成果”“聚焦技术难点”等具体要求，慢慢让回复更贴合实际面试场景。
|
||||
反馈功能的设计，让我们对“实用”有了更深理解。我们不想让反馈只停留在“回答不错”的表面，而是希望能真正帮用户改进。借助trae的便捷调试功能，我们设计了多维度评价体系，从结构逻辑、语言表达等方面给出具体建议。比如提醒用户用STAR法则梳理回答，或结合Python、SQL技能突出专业优势。这个过程让我们明白，工具是实现想法的手段，而对用户需求的深度拆解，才是产品好用的关键。
|
||||
测试时，同学说“AI提的问题他觉得十分符合当下面试的提问热点”，这句话让我们备受鼓舞。原来，trae这类工具能帮我们快速落地想法，但真正打动用户的，是我们站在求职者角度的思考。
|
||||
测试时，同学说“AI提的问题他觉得十分符合当下面试的提问热点”，这句话让我们备受鼓舞。原来，trae这类工具能帮我们快速落地想法，但真正打动用户的，是我们站在求职者角度的思考。
|
||||
这次开发让我们领悟到：技术工具再强大，也只是辅助。作为大数据专业学生，我们的核心价值在于发现问题、拆解需求、用逻辑让工具服务于人。未来，我们想继续用trae优化功能，加入面试表现分析、薪资参考等模块，让这个工具真正成为求职者的助力，也让我们在实践中更懂技术与需求的平衡之道。
|
||||
153
app.py
153
app.py
@ -1,10 +1,30 @@
|
||||
import logging
|
||||
import sys
|
||||
|
||||
logging.basicConfig(
|
||||
level=logging.DEBUG,
|
||||
format='%(asctime)s - %(levelname)s - %(message)s',
|
||||
handlers=[
|
||||
logging.FileHandler('app_debug.log'),
|
||||
logging.StreamHandler(sys.stdout)
|
||||
]
|
||||
)
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
from flask import Flask, render_template, request, jsonify, session
|
||||
from services.deepseek_service import deepseek_service
|
||||
from config import Config
|
||||
import uuid
|
||||
import datetime
|
||||
|
||||
app = Flask(__name__)
|
||||
app.secret_key = "interviewer-secret-key-change-in-production"
|
||||
# 设置session过期时间为1小时
|
||||
app.permanent_session_lifetime = datetime.timedelta(hours=1)
|
||||
|
||||
# 内存数据库存储面试数据，避免session过期导致的"面试不存在"错误
|
||||
interviews_db = {}
|
||||
|
||||
INTERVIEW_PHASES = {
|
||||
"intro": "自我介绍",
|
||||
@ -86,7 +106,7 @@ def start_interview():
|
||||
|
||||
interview_id = str(uuid.uuid4())
|
||||
|
||||
session[f"interview_{interview_id}"] = {
|
||||
interview_data = {
|
||||
"job_position": job_position,
|
||||
"difficulty": difficulty,
|
||||
"current_phase": "intro",
|
||||
@ -95,6 +115,9 @@ def start_interview():
|
||||
"is_active": True
|
||||
}
|
||||
|
||||
session[f"interview_{interview_id}"] = interview_data
|
||||
interviews_db[interview_id] = interview_data
|
||||
|
||||
first_question = deepseek_service.generate_interview_question(
|
||||
job_position=job_position,
|
||||
difficulty=difficulty,
|
||||
@ -125,11 +148,16 @@ def answer_question():
|
||||
if not interview_id:
|
||||
return jsonify({"error": "无效的面试ID"}), 400
|
||||
|
||||
interview_key = f"interview_{interview_id}"
|
||||
if interview_key not in session:
|
||||
return jsonify({"error": "面试不存在或已结束"}), 400
|
||||
# 首先从内存数据库查找面试数据
|
||||
if interview_id not in interviews_db:
|
||||
# 如果内存数据库中没有，再检查session
|
||||
interview_key = f"interview_{interview_id}"
|
||||
if interview_key not in session:
|
||||
return jsonify({"error": "面试不存在或已结束"}), 400
|
||||
# 如果session中有，同步到内存数据库
|
||||
interviews_db[interview_id] = session[interview_key]
|
||||
|
||||
interview_data = session[interview_key]
|
||||
interview_data = interviews_db[interview_id]
|
||||
|
||||
if not interview_data["is_active"]:
|
||||
return jsonify({"error": "面试已结束"}), 400
|
||||
@ -141,9 +169,12 @@ def answer_question():
|
||||
|
||||
if request_feedback:
|
||||
last_question = ""
|
||||
logger.debug(f"查找最后一个问题，对话历史长度：{len(conversation_history)}")
|
||||
for msg in reversed(conversation_history):
|
||||
if msg["role"] == "assistant" and "？" in msg["content"]:
|
||||
logger.debug(f"检查消息：角色={msg['role']}, 内容={msg['content'][:50]}...")
|
||||
if msg["role"] == "assistant" and ("？" in msg["content"] or "?" in msg["content"]):
|
||||
last_question = msg["content"]
|
||||
logger.debug(f"找到最后一个问题：{last_question[:50]}...")
|
||||
break
|
||||
|
||||
if last_question:
|
||||
@ -161,7 +192,11 @@ def answer_question():
|
||||
"ended": False
|
||||
})
|
||||
except Exception as e:
|
||||
logger.error(f"生成反馈失败：{str(e)}", exc_info=True)
|
||||
return jsonify({"error": f"生成反馈失败：{str(e)}"}), 500
|
||||
else:
|
||||
logger.warning("没有找到最后一个问题")
|
||||
return jsonify({"error": "没有找到相关问题"}), 400
|
||||
|
||||
conversation_history.append({"role": "user", "content": user_answer})
|
||||
|
||||
@ -205,7 +240,7 @@ def answer_question():
|
||||
next_question = deepseek_service.generate_interview_question(
|
||||
job_position=interview_data["job_position"],
|
||||
difficulty=interview_data["difficulty"],
|
||||
conversation_history=conversation_history[:-1],
|
||||
conversation_history=conversation_history,
|
||||
phase=next_phase
|
||||
)
|
||||
conversation_history.append({
|
||||
@ -219,14 +254,16 @@ def answer_question():
|
||||
next_question = deepseek_service.generate_interview_question(
|
||||
job_position=interview_data["job_position"],
|
||||
difficulty=interview_data["difficulty"],
|
||||
conversation_history=conversation_history[:-1],
|
||||
conversation_history=conversation_history,
|
||||
phase=current_phase
|
||||
)
|
||||
conversation_history.append({"role": "assistant", "content": next_question})
|
||||
except Exception as e:
|
||||
return jsonify({"error": f"生成问题失败：{str(e)}"}), 500
|
||||
|
||||
session[interview_key] = interview_data
|
||||
# 同时更新内存数据库和session中的数据
|
||||
interviews_db[interview_id] = interview_data
|
||||
session[f"interview_{interview_id}"] = interview_data
|
||||
|
||||
return jsonify({
|
||||
"question": next_question,
|
||||
@ -238,47 +275,95 @@ def answer_question():
|
||||
|
||||
@app.route("/api/interview/feedback", methods=["POST"])
|
||||
def get_interview_feedback():
|
||||
data = request.json
|
||||
interview_id = data.get("interview_id", "").strip()
|
||||
conversation_history = data.get("conversation_history", [])
|
||||
logger.info("接收到面试反馈请求")
|
||||
|
||||
if not interview_id:
|
||||
return jsonify({"error": "无效的面试ID"}), 400
|
||||
|
||||
system_prompt = """作为一位专业的面试评估专家，请对整场面试进行全面评估。
|
||||
try:
|
||||
# 首先确保能够获取请求数据
|
||||
if not request.is_json:
|
||||
logger.warning("请求数据不是JSON格式")
|
||||
return jsonify({"error": "请求数据必须是JSON格式"}), 400
|
||||
|
||||
data = request.json
|
||||
logger.debug(f"请求数据：{data}")
|
||||
|
||||
# 检查必要参数
|
||||
interview_id = data.get("interview_id", "").strip()
|
||||
conversation_history = data.get("conversation_history", [])
|
||||
|
||||
if not interview_id:
|
||||
logger.warning("面试ID无效")
|
||||
return jsonify({"error": "无效的面试ID"}), 400
|
||||
|
||||
if not conversation_history:
|
||||
logger.warning("没有提供面试对话历史")
|
||||
return jsonify({"error": "没有提供面试对话历史"}), 400
|
||||
|
||||
# 获取岗位信息
|
||||
job_position = ""
|
||||
if interview_id in interviews_db:
|
||||
job_position = interviews_db[interview_id].get("job_position", "")
|
||||
|
||||
logger.debug(f"岗位信息：{job_position}")
|
||||
logger.debug(f"对话历史长度：{len(conversation_history)}")
|
||||
|
||||
# 构建系统提示
|
||||
system_prompt = """作为一位专业的面试评估专家，请对整场面试进行全面评估。
|
||||
|
||||
请提供：
|
||||
1. 整体表现评分（0-100分）和评级（优秀/良好/一般/需改进）
|
||||
2. 各轮回答的详细分析
|
||||
3. strengths（优势）
|
||||
4. areas_for_improvement（需要改进的方面）
|
||||
5. 具体的准备建议
|
||||
2. 各轮回答的详细分析（针对每个问题和回答给出具体评价）
|
||||
3. strengths（优势）- 列出至少3点
|
||||
4. areas_for_improvement（需要改进的方面）- 列出至少3点
|
||||
5. 具体的准备建议（针对改进点给出可操作的建议）
|
||||
|
||||
请用中文回复，格式清晰、结构化。"""
|
||||
请用中文回复，格式清晰、结构化，避免过于笼统的描述。"""
|
||||
|
||||
conversation_text = "\n\n".join([
|
||||
f"{msg['role']}：{msg['content']}"
|
||||
for msg in conversation_history[-30:]
|
||||
])
|
||||
|
||||
messages = [
|
||||
{"role": "system", "content": system_prompt},
|
||||
{"role": "user", "content": f"请分析以下面试对话并给出综合反馈：\n\n{conversation_text}"}
|
||||
]
|
||||
|
||||
try:
|
||||
# 构建对话文本
|
||||
conversation_text = "\n\n".join([
|
||||
f"{'面试官' if msg['role'] == 'assistant' else '候选人'}：{msg['content']}"
|
||||
for msg in conversation_history
|
||||
])
|
||||
|
||||
# 构建用户提示
|
||||
if job_position:
|
||||
user_prompt = f"请分析以下针对{job_position}岗位的面试对话并给出综合反馈：\n\n{conversation_text}"
|
||||
else:
|
||||
user_prompt = f"请分析以下面试对话并给出综合反馈：\n\n{conversation_text}"
|
||||
|
||||
# 构建完整的消息
|
||||
messages = [
|
||||
{"role": "system", "content": system_prompt},
|
||||
{"role": "user", "content": user_prompt}
|
||||
]
|
||||
|
||||
logger.debug("准备调用DeepSeek API生成反馈")
|
||||
logger.debug(f"API请求消息数量：{len(messages)}")
|
||||
|
||||
# 调用DeepSeek API
|
||||
response = deepseek_service._call_api(messages)
|
||||
logger.debug("DeepSeek API调用成功")
|
||||
|
||||
# 处理API响应
|
||||
if not response or "choices" not in response or not response["choices"]:
|
||||
logger.error("API响应格式错误：缺少choices字段")
|
||||
return jsonify({"error": "生成反馈失败：API响应格式错误"}), 500
|
||||
|
||||
feedback = response["choices"][0]["message"]["content"]
|
||||
logger.info("反馈生成成功")
|
||||
logger.debug(f"生成的反馈内容长度：{len(feedback)}字符")
|
||||
logger.debug(f"生成的反馈内容开头：{feedback[:100]}...")
|
||||
|
||||
return jsonify({"feedback": feedback})
|
||||
|
||||
except Exception as e:
|
||||
return jsonify({"error": str(e)}), 500
|
||||
logger.error(f"生成面试反馈失败：{str(e)}", exc_info=True)
|
||||
# 返回更具体的错误信息
|
||||
return jsonify({"error": f"生成面试反馈失败：{str(e)}", "details": str(type(e).__name__)}), 500
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
app.run(
|
||||
host=Config.APP_HOST,
|
||||
port=Config.APP_PORT,
|
||||
debug=Config.DEBUG
|
||||
debug=False
|
||||
)
|
||||
|
||||
@ -34,11 +34,14 @@ class DeepSeekService:
|
||||
"stream": stream
|
||||
}
|
||||
|
||||
response = requests.post(
|
||||
# 创建会话对象并设置超时
|
||||
session = requests.Session()
|
||||
session.timeout = 120 # 会话级别的超时设置
|
||||
|
||||
response = session.post(
|
||||
f"{self.api_base}/chat/completions",
|
||||
headers=self._get_headers(),
|
||||
json=payload,
|
||||
timeout=30
|
||||
json=payload
|
||||
)
|
||||
|
||||
if response.status_code != 200:
|
||||
|
||||
@ -1,5 +1,5 @@
|
||||
let currentFeature = null;
|
||||
let interviewId = null;
|
||||
let interviewId = localStorage.getItem('interviewId') || null;
|
||||
let conversationKey = 'default';
|
||||
|
||||
function showFeature(feature) {
|
||||
@ -35,9 +35,7 @@ function showFeature(feature) {
|
||||
if (targetSection) {
|
||||
document.getElementById(targetSection).style.display = 'block';
|
||||
|
||||
if (feature === 'interview' || feature === 'feedback') {
|
||||
resetInterview();
|
||||
} else if (feature === 'resume') {
|
||||
if (feature === 'resume') {
|
||||
document.getElementById('resume-result').style.display = 'none';
|
||||
}
|
||||
}
|
||||
@ -130,6 +128,7 @@ async function startInterview() {
|
||||
}
|
||||
|
||||
interviewId = data.interview_id;
|
||||
localStorage.setItem('interviewId', interviewId);
|
||||
|
||||
document.getElementById('interview-setup').style.display = 'none';
|
||||
document.getElementById('interview-active').style.display = 'flex';
|
||||
@ -207,7 +206,8 @@ async function submitAnswer() {
|
||||
}
|
||||
|
||||
if (data.ended) {
|
||||
showFinalFeedback(data.conversation_history);
|
||||
// 面试结束，调用endInterview获取生成的面试反馈
|
||||
endInterview();
|
||||
} else if (data.question) {
|
||||
addInterviewMessage(data.question, 'interviewer');
|
||||
updatePhaseBadge(data.phase);
|
||||
@ -256,6 +256,23 @@ async function requestFeedback() {
|
||||
|
||||
if (data.feedback) {
|
||||
addInterviewMessage(data.feedback, 'feedback');
|
||||
|
||||
// 添加返回面试的按钮
|
||||
const messagesContainer = document.getElementById('interview-messages');
|
||||
const returnButton = document.createElement('button');
|
||||
returnButton.className = 'return-interview-btn';
|
||||
returnButton.textContent = '返回面试';
|
||||
returnButton.onclick = function() {
|
||||
// 移除按钮
|
||||
this.remove();
|
||||
// 可以在这里添加额外的逻辑，比如重新启用输入框等
|
||||
};
|
||||
|
||||
const buttonContainer = document.createElement('div');
|
||||
buttonContainer.className = 'return-button-container';
|
||||
buttonContainer.appendChild(returnButton);
|
||||
messagesContainer.appendChild(buttonContainer);
|
||||
messagesContainer.scrollTop = messagesContainer.scrollHeight;
|
||||
}
|
||||
|
||||
} catch (error) {
|
||||
@ -298,7 +315,8 @@ async function endInterview() {
|
||||
showFinalFeedback(data.feedback);
|
||||
|
||||
} catch (error) {
|
||||
alert('获取反馈失败：' + error.message);
|
||||
console.error('获取反馈失败：', error);
|
||||
alert('获取反馈失败，请稍后重试');
|
||||
} finally {
|
||||
showLoading(false);
|
||||
}
|
||||
@ -311,7 +329,19 @@ function showFinalFeedback(feedbackData) {
|
||||
const feedbackContent = document.getElementById('feedback-content');
|
||||
|
||||
if (typeof feedbackData === 'string') {
|
||||
feedbackContent.innerHTML = formatSuggestions(feedbackData);
|
||||
// 将Markdown风格的列表和标题转换为HTML
|
||||
let formattedFeedback = feedbackData
|
||||
// 转换标题
|
||||
.replace(/^###\s+(.*)$/gm, '<h4>$1</h4>')
|
||||
.replace(/^##\s+(.*)$/gm, '<h3>$1</h3>')
|
||||
// 转换列表
|
||||
.replace(/^\*\s+(.*)$/gm, '<li>$1</li>')
|
||||
.replace(/(?:<li>.*<\/li>\n?)+/gm, '<ul>$&</ul>')
|
||||
// 转换换行
|
||||
.replace(/\n\n/g, '</p><p>')
|
||||
.replace(/\n/g, '<br>');
|
||||
|
||||
feedbackContent.innerHTML = `<div class="feedback-container">${formattedFeedback}</div>`;
|
||||
} else {
|
||||
let html = '<h4>面试综合评估</h4>';
|
||||
feedbackData.forEach(msg => {
|
||||
@ -323,6 +353,7 @@ function showFinalFeedback(feedbackData) {
|
||||
|
||||
function resetInterview() {
|
||||
interviewId = null;
|
||||
localStorage.removeItem('interviewId');
|
||||
document.getElementById('interview-setup').style.display = 'block';
|
||||
document.getElementById('interview-active').style.display = 'none';
|
||||
document.getElementById('interview-feedback').style.display = 'none';
|
||||
|
||||
Loading…
Reference in New Issue
Block a user