---
title: "04-ann_model_job"
created: 2026-04-02
tags:
  - 项目
aliases:
  - ann_model_job
---
# ann_model_job.py

### `ann_model_job.py` — 公告内单职位解析

由 [[03-ann_model|ann_model]] 分支调用：HTML 转纯文本 → Qwen 大模型提取 → 校验质量 → 与公告元数据（公司名、链接、申请方式、联系人）合并 → 写入标准职位 JSON。

## 代码

```python
# -*- coding: utf-8 -*-
"""
单个职位解析模块
功能：从招聘详情HTML中，使用大模型提取职位描述、任职要求等结构化信息
输出：标准职位JSON，用于入库
"""

import json
import sys
sys.path.append('../')
import re

from utils import ner_logger, getMD5Str, get_local_ip, QZ_VERISON

# 大模型调用接口
from api.doubao_api import call_gpt as doubao_call_gpt
from api.qwen_api import call_gpt as qwen_call_gpt

# 职位解析Prompt模板
from parsegpt.template import get_template_cjob_annotation

# 学历标准化工具
from utils_resume import fix_diploma

# HTML纯文本提取工具
from parsegpt.html_to_text import Html2txt

# HTML解析
from bs4 import BeautifulSoup

# ====================== 核心函数：解析单个招聘职位 ======================
def parse_cjob(_hfile, ann_json_data):
    """
    从HTML详情页提取单个职位的结构化信息
    :param _hfile: 本地HTML文件路径
    :param ann_json_data: 上层已解析的公告信息（公司、标题、链接等）
    :return: (状态, 职位JSON数据)
    """
    _ann_dict = {}

    try:
        # 1. 从HTML文件提取纯文本内容
        _text = get_cjob_html_content(_hfile)

        # 2. 构造大模型Prompt，调用千问API解析职位信息
        _t_text = get_template_cjob_annotation(_text)
        (_ok_flag, json_str) = qwen_call_gpt(_t_text, True)

        if not _ok_flag:
            return "", f"通过大模型获取公告里面的单个职位职位信息Error:{_ok_flag}\n{json_str}"

        ner_logger.info(f"通过大模型获取公告里面的单个职位信息json:{json_str}")

        # 3. 解析大模型返回的JSON
        json_data = json.loads(json_str, strict=False)

        # 4. 过滤：职位描述+任职要求过短，视为无效职位
        if len(json_data['JobDescribe']) + len(json_data['Jobreq']) < 30:
            return "", f"职位信息的描述太少:{_hfile}\n{json_str}"

        # 5. 学历信息标准化（本科/硕士/博士等）
        fix_diploma_data_map(json_data)

        # ====================== 复制上层公告的公共字段 ======================
        json_data['FileId'] = ann_json_data['FileId']
        json_data['JobLink'] = ann_json_data['JobLink']
        json_data['DocType'] = 'xiaozhao'                  # 文档类型：校招
        json_data['ComLogo'] = ""                          # 公司Logo
        json_data['ComName'] = ann_json_data['ComName']     # 公司名称
        json_data['ComShortName'] = ann_json_data['ComName']
        json_data['NoticeToJob'] = 1                       # 标记为招聘公告
        json_data['WxName'] = ann_json_data['WeixinName']  # 公众号名称
        json_data['ApplyTypeLink'] = ann_json_data['ApplyTypeLink']  # 网申链接
        json_data['GraduationYear'] = ann_json_data['GraduationYear']  # 招聘届别
        json_data['AnnouncementLabel'] = ann_json_data['AnnouncementLabel']
        json_data['JobTitle'] = ann_json_data['JobTitle']  # 职位标题
        json_data['EmailSubject'] = ann_json_data['EmailSubject']
        json_data['ApplyTypeText'] = ann_json_data['ApplyTypeText']
        json_data['GraduationTimeRequirement'] = ann_json_data['GraduationTimeRequirement']
        json_data['mdfile'] = ann_json_data['mdfile']
        json_data['HopeWorkType'] = ann_json_data['HopeWorkType']
        json_data['PublishTime'] = ann_json_data['PublishTime']

        # ====================== 复制联系人信息 ======================
        if 'ApplyContacts' in ann_json_data:
            acontacts = ann_json_data['ApplyContacts']
            if len(acontacts) > 0:
                # 联系人姓名（不含*才有效）
                if 'Name' in acontacts[0] and "*" not in acontacts[0]['Name']:
                    json_data['ContactPerson'] = acontacts[0]['Name']
                # 联系电话
                if 'Mobile' in acontacts[0] and "*" not in acontacts[0]['Mobile']:
                    json_data['Phone'] = acontacts[0]['Mobile']
                # 邮箱
                if 'Email' in acontacts[0] and "*" not in acontacts[0]['Email']:
                    json_data['Email'] = acontacts[0]['Email']

        # 如果没有联系人邮箱，从公告公共邮箱补充
        if 'Email' in json_data and json_data['Email'] == "" and 'ApplyTypeEmail' in ann_json_data and ann_json_data['ApplyTypeEmail'] != "":
            json_data['Email'] = ann_json_data['ApplyTypeEmail']

        # ====================== 组装最终输出结构 ======================
        _ann_dict['cjob'] = json_data  # 职位主体信息

        # 清理冗余字段（二维码）
        _n_ann_json_data = ann_json_data
        if 'ApplyTypeQrcode' in _n_ann_json_data:
            _n_ann_json_data.pop('ApplyTypeQrcode')
        _ann_dict['other'] = _n_ann_json_data

        # 固定前端展示用的字段名映射
        _ann_dict['cjob_o_field'] = {
            'RecruitProcess': "招聘流程",
            'Attention': '注意事项',
            'WelfareInfo': '薪酬福利',
            'JobDevelopment': '岗位发展',
            'ApplyTypeText': '应聘方式'
        }

    # JSON解析异常处理
    except json.JSONDecodeError as e:
        import traceback
        traceback.print_exc()
        ner_logger.error(f"Error:{json_str}")
        return "", f"通过大模型获取公告里面的单个职位信息Error:\n{e}"

    # 返回成功状态 + 职位结构化数据
    return "OK", _ann_dict


# ====================== 学历标准化 ======================
def fix_diploma_data_map(item):
    """
    统一学历格式：如“本科及以上”“硕士”等
    """
    if 'Degree' in item:
        need_fix = item['Degree']
        item['Degree'] = fix_diploma(need_fix)


# ====================== 从HTML提取职位纯文本 ======================
def get_cjob_html_content(_htmlfile):
    """
    读取HTML文件，清洗标签，返回干净的职位文本
    """
    with open(_htmlfile, "r", encoding="utf-8", errors='ignore') as f1:
        htmltext = f1.read()
        # 使用工具类清洗HTML，返回纯文本
        text = Html2txt().clean_html(htmltext)
        return text
```

---

**项目分区导航**：[[03-ann_model|ann_model]] ⬅️ | 04-ann_model_job | ➡️ [[05-cjob_model|cjob_model]]