This commit is contained in:
2026-08-31 19:18:05 +08:00
parent 493287c99f
commit a9752b01a9
10 changed files with 167 additions and 27 deletions
+13 -11
View File
@@ -3,11 +3,11 @@ schema_version: 1
locale: zh-CN
branch: master
nodes_managed: true
exported_at: "2026-08-26T15:18:24Z"
exported_at: "2026-08-31T09:16:18Z"
modules:
"":
dir_name: jianli-boss
title: jianli-boss
dir_name: Boss 直聘 AI 投递系统(客户端-服务端协作)
title: Boss 直聘 AI 投递系统(客户端-服务端协作)
scope: []
source_files: []
children:
@@ -16,8 +16,8 @@ modules:
depends_on: []
related_to: []
client:
dir_name: 简历投递客户端(本地控制面板 + AdsPower + AI 评分)
title: 简历投递客户端(本地控制面板 + AdsPower + AI 评分)
dir_name: 客户端自动化控制服务(client)
title: 客户端自动化控制服务(client)
scope:
- client/
source_files: []
@@ -26,8 +26,8 @@ modules:
depends_on: []
related_to: []
client/shared_client_utils:
dir_name: 客户端主程序与云端_AI 能力封装
title: 客户端主程序与云端/AI 能力封装
dir_name: 客户端主进程与云端 API 封装
title: 客户端主进程与云端 API 封装
scope:
- client/main.py
- client/config.py
@@ -38,17 +38,19 @@ modules:
source_files: []
children: []
depends_on: []
related_to: []
related_to:
- path: server
server:
dir_name: Boss直聘投递服务端
title: Boss直聘投递服务端
dir_name: Boss 直聘投递服务(FastAPI 服务端)
title: Boss 直聘投递服务(FastAPI 服务端)
scope:
- server/
source_files: []
children:
- server/infrastructure
depends_on: []
related_to: []
related_to:
- path: client/shared_client_utils
server/infrastructure:
dir_name: FastAPI 服务端基础设施(应用启动、配置、数据库、ORM、认证)
title: FastAPI 服务端基础设施(应用启动、配置、数据库、ORM、认证)
File diff suppressed because one or more lines are too long
+4 -3
View File
@@ -248,12 +248,13 @@ class ServerAPI:
# ------------------------------------------------------------------
# 投递记录
# ------------------------------------------------------------------
# 上报字段长度上限(与服务端列长对齐;jd_text 上限同服务端 JD_TEXT_LIMIT,
# 防止超大包触发 MySQL max_allowed_packet=1153 导致记录反复上报失败)
# 上报字段长度上限(与服务端列长对齐);jd_text 需传完整全文供服务端按上限分流:
# 短文本入库,超长由服务端落盘归档到文件——这里 300000 仅是防病理值的安全上限,
# 不做业务截断,避免丢失归档所需的全文
_REPORT_TEXT_LIMITS = {
"job_title": 128, "company": 128, "salary": 32, "location": 64,
"ai_reason": 255, "skip_reason": 255, "greeting": 500,
"jd_text": 20000,
"jd_text": 300000,
}
def report_application(self, session_id: int | None, data: dict) -> dict:
+3
View File
@@ -55,6 +55,9 @@ services:
DEBUG: ${DEBUG:-false}
ports:
- "8000:8000"
volumes:
# 超长 JD 归档目录持久化(服务端写入 /app/jd_archive/,重建容器不丢文件)
- ./jd_archive:/app/jd_archive
restart: unless-stopped
volumes:
+5 -1
View File
@@ -93,7 +93,8 @@ CREATE TABLE IF NOT EXISTS job_applications (
ai_score INT NOT NULL DEFAULT -1 COMMENT 'AI匹配评分(-1=未评分)',
ai_reason VARCHAR(255) NULL COMMENT 'AI评分理由',
skip_reason VARCHAR(255) NULL COMMENT '跳过原因',
jd_text TEXT NULL COMMENT '职位JD全文(投递/跳过时归档,供复盘与面试准备)',
jd_text TEXT NULL COMMENT '职位JD全文(投递/跳过时归档;超长时本列留空、落盘到 jd_archive/,改存 jd_file 路径)',
jd_file VARCHAR(255) NULL COMMENT '超长JD归档文件路径(相对项目根目录,如 jd_archive/1_xxx.txt)',
greeting VARCHAR(500) NULL COMMENT '实际发送的打招呼语(仅chat方式有值)',
applied_at DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP COMMENT '投递时间',
UNIQUE KEY uq_user_job (user_id, job_id) COMMENT '同一用户不重复投递同一职位',
@@ -244,6 +245,9 @@ CREATE TABLE IF NOT EXISTS ai_score_task (
-- v1.11 老库升级(AI 服务端代理:ai_config + ai_score_task 两表),
-- 完整可重复执行版本见 docs/upgrade_1.11.sql。
-- v1.18 老库升级(job_applications 新增 jd_file:超长 JD 落盘归档路径),
-- 完整可重复执行版本见 docs/upgrade_jd_file.sql。
-- 老库城市字段改名(城市代码 → 城市名称):
ALTER TABLE apply_sessions
CHANGE COLUMN city_code city VARCHAR(20) NULL COMMENT '城市名称';
+21
View File
@@ -0,0 +1,21 @@
-- ============================================================
-- v1.18 老库升级:job_applications 新增 jd_file 列
-- 背景:超长 JD(> 20000 字符)不再入库,落盘到项目根目录
-- jd_archive/ 文件,库内仅存相对路径,避免触发
-- MySQL max_allowed_packet(1153)。
-- 幂等:可重复执行(已存在该列时跳过)。
-- 执行:docker exec -i jianli-boss-mysql mysql -uroot -p<密码> <数据库名> < docs/upgrade_jd_file.sql
-- ============================================================
SET @col_exists := (
SELECT COUNT(*) FROM information_schema.COLUMNS
WHERE TABLE_SCHEMA = DATABASE()
AND TABLE_NAME = 'job_applications'
AND COLUMN_NAME = 'jd_file'
);
SET @ddl := IF(@col_exists = 0,
'ALTER TABLE job_applications ADD COLUMN jd_file VARCHAR(255) NULL COMMENT ''超长JD归档文件路径(相对项目根目录,如 jd_archive/1_xxx.txt)'' AFTER jd_text',
'SELECT ''jd_file 列已存在,跳过'' AS msg');
PREPARE stmt FROM @ddl;
EXECUTE stmt;
DEALLOCATE PREPARE stmt;
+4 -1
View File
@@ -131,8 +131,11 @@ class JobApplication(Base):
ai_reason = Column(String(255), nullable=True)
# 跳过原因
skip_reason = Column(String(255), nullable=True)
# 职位 JD 全文(投递/跳过时归档,供复盘与面试准备)
# 职位 JD 全文(投递/跳过时归档,供复盘与面试准备);
# 超过 JD_TEXT_LIMIT 的超长 JD 不入本列,落盘到项目根目录 jd_archive/,本列留空、改存 jd_file 路径
jd_text = Column(Text, nullable=True)
# 超长 JD 归档文件路径(相对项目根目录,如 jd_archive/1_0daaa9fd.txt)
jd_file = Column(String(255), nullable=True)
# 实际发送的打招呼语(仅 chat 方式有值)
greeting = Column(String(500), nullable=True)
applied_at = Column(DateTime, default=utcnow, nullable=False)
+66 -8
View File
@@ -1,11 +1,14 @@
"""
投递记录路由:上报投递结果、查询历史、统计
"""
import re
from datetime import datetime, timedelta
from typing import Optional
from pathlib import Path
from typing import Optional, Tuple
from fastapi import APIRouter, Depends, HTTPException, Query
from sqlalchemy.orm import Session
from sqlalchemy import func
from loguru import logger
from ..auth import get_current_user
from ..database import get_db
@@ -15,10 +18,33 @@ from ..schemas import ApplicationIn, ApplicationOut, Message, StatsOut
router = APIRouter(prefix="/api/jobs", tags=["投递记录"])
# jd_text 归档上限(字符):控制写入包体积,避免触发 MySQL max_allowed_packet(1153);
# 正常 JD 全文远低于此值,超限基本意味着抓取异常(如抓到整页文本)
# jd_text 入库上限(字符):超过则落盘为文件(库内只存路径),
# 控制写入包体积,避免触发 MySQL max_allowed_packet(1153)
JD_TEXT_LIMIT = 20000
# 归档目录:项目根目录下的 jd_archive/(本文件位于 server/routers/,上溯两级即项目根)
PROJECT_ROOT = Path(__file__).resolve().parents[2]
JD_DIR_NAME = "jd_archive"
def _archive_jd(user_id: int, job_id: str, jd_text: Optional[str]) -> Tuple[Optional[str], Optional[str]]:
"""JD 存储分流:≤ 上限直接入库;超限写入 jd_archive/ 文件。
返回 (jd_text, jd_file):两者必有一空。文件名按 用户_职位 去重,重复上报覆盖"""
if not jd_text or len(jd_text) <= JD_TEXT_LIMIT:
return jd_text, None
safe_job = re.sub(r"[^\w\-]", "_", job_id or "")[:48] or "unknown"
rel = f"{JD_DIR_NAME}/{user_id}_{safe_job}.txt"
try:
target = PROJECT_ROOT / rel
target.parent.mkdir(parents=True, exist_ok=True)
target.write_text(jd_text, encoding="utf-8")
logger.info(f"超长 JD 已归档到文件({len(jd_text)} 字符):{rel}")
return None, rel
except OSError as e:
# 落盘失败也不能把巨包塞进数据库:截断到上限入库,保底不丢全部信息且不炸库
logger.error(f"JD 归档落盘失败,截断入库:{e}")
return jd_text[:JD_TEXT_LIMIT], None
def _clip(value: Optional[str], limit: int) -> Optional[str]:
"""按列长度截断文本,防止超长写入导致 1153 / Data too long"""
@@ -26,7 +52,8 @@ def _clip(value: Optional[str], limit: int) -> Optional[str]:
def _sanitize(req: ApplicationIn) -> ApplicationIn:
"""上报字段统一裁剪(与 models.py 列长度保持一致)"""
"""上报字段统一裁剪(与 models.py 列长度保持一致);
jd_text 不在此截断——由 _archive_jd 按上限分流到数据库/文件"""
req.job_title = _clip(req.job_title, 128)
req.company = _clip(req.company, 128)
req.salary = _clip(req.salary, 32)
@@ -35,7 +62,6 @@ def _sanitize(req: ApplicationIn) -> ApplicationIn:
req.ai_reason = _clip(req.ai_reason, 255)
req.skip_reason = _clip(req.skip_reason, 255)
req.greeting = _clip(req.greeting, 500)
req.jd_text = _clip(req.jd_text, JD_TEXT_LIMIT)
return req
@@ -51,6 +77,8 @@ def report_application(
同一用户 + 同一职位 ID 去重,已存在则更新。
"""
req = _sanitize(req)
# JD 分流:短文本入库,超长落盘为文件(库存路径)
jd_text, jd_file = _archive_jd(current_user.id, req.job_id, req.jd_text)
existing = db.query(JobApplication).filter(
JobApplication.user_id == current_user.id,
JobApplication.job_id == req.job_id,
@@ -65,7 +93,9 @@ def report_application(
# 否则记录停留在首次上报时间,看起来像“漏报”
existing.applied_at = datetime.utcnow()
if req.jd_text:
existing.jd_text = req.jd_text
# 新 JD 覆盖旧值;文本与归档文件二者只留其一,避免读到陈旧数据
existing.jd_text = jd_text
existing.jd_file = jd_file
if req.greeting:
existing.greeting = req.greeting
if session_id:
@@ -84,7 +114,8 @@ def report_application(
ai_score=req.ai_score,
ai_reason=req.ai_reason,
skip_reason=req.skip_reason,
jd_text=req.jd_text,
jd_text=jd_text,
jd_file=jd_file,
greeting=req.greeting,
)
db.add(record)
@@ -208,6 +239,7 @@ def page_applications(
rows = query.order_by(JobApplication.applied_at.desc()) \
.offset((page - 1) * page_size).limit(page_size).all()
items = [{
"id": r.id,
"job_title": r.job_title,
"company": r.company,
"salary": r.salary,
@@ -217,8 +249,10 @@ def page_applications(
"ai_reason": r.ai_reason,
"skip_reason": r.skip_reason,
"greeting": r.greeting,
# JD 全文截断,供客户端悬停摘要展示,控制分页响应体积
# JD 全文截断,供客户端悬停摘要展示,控制分页响应体积;
# 完整内容(含归档文件)通过 /{app_id}/jd 接口按需获取,has_jd 标记有无 JD 可看
"jd_text": (r.jd_text or "")[:500],
"has_jd": bool(r.jd_text or r.jd_file),
"applied_at": (r.applied_at + timedelta(hours=8)).strftime("%Y-%m-%d %H:%M:%S"),
} for r in rows]
return {
@@ -256,3 +290,27 @@ def get_stats(
total_skipped=count(apply_type="skipped"),
today_applied=today_applied,
)
@router.get("/{app_id}/jd", response_model=dict)
def get_application_jd(
app_id: int,
current_user: User = Depends(get_current_user),
db: Session = Depends(get_db),
):
"""查看单条投递记录的完整 JD:先取 jd_text 字段,为空则读 jd_file 归档文件"""
record = db.query(JobApplication).filter(
JobApplication.id == app_id,
JobApplication.user_id == current_user.id,
).first()
if not record:
raise HTTPException(status_code=404, detail="投递记录不存在")
if record.jd_text:
return {"id": app_id, "source": "db", "jd_text": record.jd_text}
if record.jd_file:
target = (PROJECT_ROOT / record.jd_file).resolve()
# 防路径穿越:归档文件必须位于项目根目录内,防止库内路径被篡改后读到任意文件
if target.is_file() and str(target).startswith(str(PROJECT_ROOT)):
return {"id": app_id, "source": "file", "jd_text": target.read_text(encoding="utf-8")}
raise HTTPException(status_code=410, detail="JD 归档文件不存在或已丢失")
raise HTTPException(status_code=404, detail="该记录未包含 JD")
+2
View File
@@ -151,6 +151,8 @@ class ApplicationOut(BaseModel):
ai_reason: Optional[str]
skip_reason: Optional[str]
jd_text: Optional[str]
# 超长 JD 归档文件路径(相对项目根目录);与 jd_text 二者只有一非空
jd_file: Optional[str] = None
greeting: Optional[str]
applied_at: datetime
+48 -2
View File
@@ -72,6 +72,10 @@
.bl-btn { border: 1px solid #ffd6d6; background: #fff; color: #e64340; border-radius: 4px; font-size: 12px; padding: 1px 8px; cursor: pointer; margin-left: 6px; }
.bl-btn:hover { background: #fff1f0; }
/* JD 查看按钮 */
.jd-btn { border: 1px solid #91d5ff; background: #fff; color: #1890ff; border-radius: 4px; font-size: 12px; padding: 1px 8px; cursor: pointer; }
.jd-btn:hover { background: #e6f7ff; }
/* 黑名单页 */
.bl-add { display: flex; gap: 10px; background: #fff; border-radius: 10px; padding: 16px 18px; box-shadow: 0 2px 8px rgba(0,0,0,.05); margin-bottom: 16px; }
.bl-add input { flex: 1; padding: 9px 12px; border: 1px solid #d9d9d9; border-radius: 6px; font-size: 14px; }
@@ -156,7 +160,7 @@
<table>
<thead><tr>
<th>时间</th><th>职位</th><th>公司</th><th>薪资</th><th>结果</th><th>AI评分</th><th>备注</th>
<th>时间</th><th>职位</th><th>公司</th><th>薪资</th><th>结果</th><th>AI评分</th><th>备注</th><th style="width:60px">JD</th>
</tr></thead>
<tbody id="tbody"></tbody>
</table>
@@ -371,9 +375,12 @@ async function load() {
const company = r.company || '';
const blBtn = (company && !blSet.has(company))
? '<button class="bl-btn" data-company="' + esc(company) + '">拉黑</button>' : '';
// JD 列:有 JD(库内字段或归档文件)才可点开查看全文,走 /api/jobs/{id}/jd 按需拉取
const jdBtn = r.has_jd
? '<button class="jd-btn" data-id="' + r.id + '">查看</button>' : '<span style="color:#ccc;">-</span>';
return '<tr><td>' + esc(r.applied_at) + '</td><td>' + esc(r.job_title) + '</td><td>' + esc(company) + blBtn +
'</td><td>' + esc(r.salary) + '</td><td>' + type + '</td><td>' + score +
'</td><td class="remark" title="' + esc(remark) + '">' + esc(remark) + '</td></tr>';
'</td><td class="remark" title="' + esc(remark) + '">' + esc(remark) + '</td><td>' + jdBtn + '</td></tr>';
}).join('');
document.getElementById('empty').classList.toggle('hidden', (d.items || []).length > 0);
@@ -499,6 +506,28 @@ async function testAiConfig() {
} catch (e) { alert('测试失败:' + (e.message || '')); }
}
/* ---------- 岗位 JD 查看(先取库内字段,为空由服务端读归档文件) ---------- */
async function showJd(appId) {
const body = document.getElementById('jdBody');
const tip = document.getElementById('jdTip');
document.getElementById('jdModal').classList.remove('hidden');
tip.textContent = '加载中…';
body.textContent = '';
try {
const d = await api('/api/jobs/' + appId + '/jd');
tip.textContent = '来源:' + (d.source === 'file' ? '归档文件' : '数据库字段') +
' · ' + (d.jd_text || '').length + ' 字';
body.textContent = d.jd_text || '';
} catch (e) {
tip.textContent = '加载失败';
body.textContent = e.message || '无 JD 数据';
}
}
function closeJd() {
document.getElementById('jdModal').classList.add('hidden');
}
/* ---------- 页面切换 ---------- */
function switchPage(page) {
document.querySelectorAll('aside .menu-item').forEach(el =>
@@ -588,6 +617,8 @@ document.getElementById('blTbody').addEventListener('click', e => {
if (btn) removeBlacklist(btn.dataset.id);
});
document.getElementById('tbody').addEventListener('click', e => {
const jdBtn = e.target.closest('.jd-btn');
if (jdBtn) { showJd(jdBtn.dataset.id); return; }
const btn = e.target.closest('.bl-btn');
if (btn && confirm('确认拉黑公司「' + btn.dataset.company + '」?拉黑后将不再投递该公司。')) {
addBlacklist(btn.dataset.company);
@@ -720,5 +751,20 @@ setInterval(() => {
</div>
</div>
</div>
<!-- 岗位 JD 查看弹窗 -->
<div id="jdModal" class="hidden" style="position:fixed;inset:0;background:rgba(0,0,0,0.45);z-index:2000;display:flex;align-items:center;justify-content:center;">
<div style="background:#fff;border-radius:8px;padding:24px;width:760px;max-width:95vw;max-height:85vh;display:flex;flex-direction:column;">
<h3 style="margin:0 0 8px;">岗位 JD 全文</h3>
<div id="jdTip" style="font-size:12px;color:#999;margin-bottom:10px;"></div>
<pre id="jdBody" style="flex:1;min-height:0;overflow:auto;white-space:pre-wrap;word-break:break-all;background:#fafafa;border:1px solid #f0f0f0;border-radius:6px;padding:14px;font-size:13px;line-height:1.7;"></pre>
<div style="margin-top:16px;text-align:right;">
<button id="jdClose" style="padding:6px 14px;border-radius:4px;border:1px solid #d9d9d9;background:#fff;cursor:pointer;">关闭</button>
</div>
</div>
</div>
<script>
document.getElementById('jdClose').onclick = closeJd;
document.getElementById('jdModal').addEventListener('click', e => { if (e.target.id === 'jdModal') closeJd(); });
</script>
</body>
</html>