简历智能匹配系统 – 颠覆海投没回音
一、实际应用场景描述
目标用户：数字文化艺术创新创业课程的毕业生、文化创意行业求职者、转型期职场人
典型场景：
– 小李学完《数字艺术IP孵化》课程，想投递"文化科技公司产品经理"岗位，但不知道自己的简历哪些关键词匹配
– 小王有3年传统广告经验，想转行做"数字文创运营"，海投50份简历只有2个面试，不知道问题在哪
– 小张应届毕业，面对"元宇宙内容策划"、"NFT运营专员"等新岗位，不知道如何包装自己的课程项目经历
传统痛点：
– 关键词盲区：不知道HR用ATS系统筛选哪些关键词
– 匹配度模糊：无法量化简历与目标岗位的契合程度
– 针对性不足：一份简历海投所有岗位，缺乏定制化
– 反馈缺失：投递后石沉大海，不知道哪里出了问题
二、核心痛点分析
痛点 影响 解决思路
关键词识别不准 ATS系统直接淘汰 多维度关键词提取+行业词典
匹配度无法量化 不知道改进方向 余弦相似度算法+技能矩阵
简历内容不聚焦 核心竞争力被埋没 智能高亮+缺失提醒
缺乏行业洞察 错失新兴岗位机会 动态行业词库+趋势分析
三、核心逻辑讲解
┌─────────────┐ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐
│ 简历解析 │ –> │ 技能提取 │ –> │ JD解析 │ –> │ 智能匹配 │ –> │ 报告生成 │
│ (PyPDF2/ │ │ (NER+分词) │ │ (关键词挖掘) │ │ (相似度计算) │ │ (可视化) │
│ python-docx)│ │ │ │ │ │ │ │ │
└─────────────┘ └─────────────┘ └─────────────┘ └─────────────┘ └─────────────┘
│ │ │ │ │
▼ ▼ ▼ ▼ ▼
支持PDF/DOCX 识别硬技能+软技能 提取岗位核心要求 计算匹配分数 生成优化建议
提取文本+格式 构建个人技能画像 分析公司业务方向 生成技能矩阵 可视化展示
关键技术点：
1. 简历解析：多格式支持，提取文本同时保留基础结构
2. 技能提取：结合命名实体识别(NER)和领域词典，精准识别技能
3. JD解析：基于职位描述的语义分析，提取硬性要求和加分项
4. 智能匹配：TF-IDF向量化+余弦相似度，量化匹配程度
5. 报告生成：可视化匹配结果，提供具体优化建议
四、代码实现
项目结构
resume_matcher/
├── resume_matcher.py # 主程序入口
├── resume_parser.py # 简历解析模块
├── skill_extractor.py # 技能提取模块
├── jd_analyzer.py # 职位描述分析模块
├── matcher_engine.py # 智能匹配引擎
├── report_generator.py # 报告生成模块
├── industry_dicts.py # 行业词典模块
├── config.py # 配置文件
├── requirements.txt # 依赖清单
└── README.md # 使用说明
1. config.py – 配置文件
"""
配置参数模块
包含文件路径、匹配规则、行业词典配置等
"""
# 文件路径配置
RESUME_PATH = "input/resume.pdf" # 输入简历路径
JD_PATH = "input/job_description.txt" # 职位描述文件路径
OUTPUT_DIR = "output" # 输出目录
# 匹配算法配置
MATCH_THRESHOLD = 0.6 # 基础匹配阈值
HIGH_WEIGHT_SKILLS = 2.0 # 高权重技能倍数
MEDIUM_WEIGHT_SKILLS = 1.5 # 中权重技能倍数
LOW_WEIGHT_SKILLS = 1.0 # 低权重技能倍数
# 行业配置
CURRENT_INDUSTRY = "digital_culture" # 当前行业
SUPPORTED_INDUSTRIES = [
"digital_culture", # 数字文化
"tech_product", # 科技产品
"creative_design", # 创意设计
"marketing_ops" # 营销运营
]
# 输出配置
GENERATE_REPORT = True # 是否生成详细报告
GENERATE_VISUALIZATION = True # 是否生成可视化
REPORT_FORMATS = ["html", "json"] # 报告格式
# 行业词典配置
INDUSTRY_DICT_PATHS = {
"digital_culture": "dicts/digital_culture_skills.txt",
"tech_product": "dicts/tech_product_skills.txt",
"creative_design": "dicts/creative_design_skills.txt",
"marketing_ops": "dicts/marketing_ops_skills.txt"
}
2. industry_dicts.py – 行业词典模块
"""
行业词典模块
包含各行业的技能词库、职位词库、工具词库
"""
import os
from typing import Dict, List, Set
import logging
# 配置日志
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)
class IndustryDictionaries:
"""
行业词典管理类
负责加载和管理各行业的专业词汇
"""
def __init__(self):
"""初始化行业词典"""
self.dicts = {}
self._load_default_dicts()
def _load_default_dicts(self):
"""加载默认行业词典"""
# 数字文化行业词典
self.dicts["digital_culture"] = {
"hard_skills": [
# 核心技术
"数字艺术", "IP孵化", "虚拟现实", "增强现实", "混合现实",
"NFT", "区块链", "元宇宙", "数字孪生", "3D建模",
"Unity3D", "Unreal Engine", "Blender", "Maya", "Cinema 4D",
"数字策展", "文化数据分析", "用户画像", "内容运营",
"AIGC", "生成式AI", "Midjourney", "Stable Diffusion", "GPT",
# 平台工具
"Steam", "Roblox", "Decentraland", "Sandbox", "OpenSea",
"幻核", "鲸探", "百度超级链", "蚂蚁链", "腾讯云区块链"
],
"soft_skills": [
"文化敏感度", "创意策划", "跨界思维", "用户洞察",
"内容敏感度", "审美能力", "故事叙述", "IP运营思维",
"团队协作", "项目管理", "市场嗅觉", "创新思维"
],
"tools": [
"Photoshop", "Illustrator", "After Effects", "Premiere",
"Figma", "Sketch", "Principle", "ProtoPie",
"Python", "JavaScript", "HTML5", "CSS3",
"SQL", "Tableau", "Power BI", "Excel"
],
"positions": [
"数字艺术设计师", "IP运营专员", "元宇宙内容策划",
"NFT项目经理", "数字策展人", "文创产品经理",
"虚拟空间设计师", "AIGC应用专家", "文化数据分析师"
]
}
# 科技产品行业词典
self.dicts["tech_product"] = {
"hard_skills": [
"产品规划", "需求分析", "原型设计", "PRD撰写",
"敏捷开发", "Scrum", "Kanban", "用户研究",
"数据分析", "AB测试", "漏斗分析", "留存分析",
"API设计", "数据库设计", "微服务架构", "云计算",
"机器学习", "深度学习", "推荐算法", "自然语言处理"
],
"soft_skills": [
"逻辑思维", "沟通协调", "决策能力", "抗压能力",
"同理心", "商业敏感度", "学习能力", "执行力"
],
"tools": [
"Axure", "墨刀", "Figma", "Sketch", "Jira",
"Confluence", "Slack", "Notion", "Postman",
"MySQL", "MongoDB", "Redis", "Docker", "Kubernetes"
],
"positions": [
"产品经理", "产品助理", "高级产品经理", "产品总监",
"技术产品经理", "数据产品经理", "B端产品经理", "C端产品经理"
]
}
# 创意设计行业词典
self.dicts["creative_design"] = {
"hard_skills": [
"品牌设计", "UI设计", "UX设计", "视觉设计",
"交互设计", "动效设计", "插画设计", "字体设计",
"版式设计", "色彩理论", "构图技巧", "设计系统",
"响应式设计", "移动端设计", "网页设计", "电商设计"
],
"soft_skills": [
"审美能力", "创意思维", "细节把控", "趋势敏感度",
"客户沟通", "提案能力", "时间管理", "自我驱动"
],
"tools": [
"Adobe Creative Suite", "Figma", "Sketch", "InVision",
"Zeplin", "Abstract", "Procreate", "C4D",
"Blender", "Cinema 4D", "Octane", "Redshift"
],
"positions": [
"UI设计师", "UX设计师", "视觉设计师", "品牌设计师",
"动效设计师", "插画师", "设计主管", "创意总监"
]
}
# 营销运营行业词典
self.dicts["marketing_ops"] = {
"hard_skills": [
"内容营销", "社交媒体运营", "社群运营", "私域运营",
"活动策划", "用户增长", "转化优化", "SEO/SEM",
"数据分析", "ROI分析", "LTV计算", "CAC分析",
"邮件营销", "短信营销", "短视频运营", "直播运营"
],
"soft_skills": [
"文案能力", "网感", "用户洞察", "数据分析思维",
"执行力", "创意策划", "危机公关", "跨部门协作"
],
"tools": [
"微信公众号", "微博", "抖音", "小红书", "B站",
"企业微信", "钉钉", "飞书", "神策数据",
"GrowingIO", "Google Analytics", "百度统计", "友盟+"
],
"positions": [
"内容运营", "社群运营", "用户运营", "活动运营",
"新媒体运营", "增长黑客", "营销经理", "运营总监"
]
}
logger.info(f"已加载{len(self.dicts)}个行业词典")
def get_dict(self, industry: str) -> Dict[str, List[str]]:
"""
获取指定行业的词典
Args:
industry: 行业名称
Returns:
dict: 行业词典
"""
if industry in self.dicts:
return self.dicts[industry]
else:
logger.warning(f"未找到行业词典: {industry}，使用默认词典")
return self.dicts["tech_product"]
def merge_dicts(self, industries: List[str]) -> Dict[str, List[str]]:
"""
合并多个行业词典
Args:
industries: 行业名称列表
Returns:
dict: 合并后的词典
"""
merged = {
"hard_skills": [],
"soft_skills": [],
"tools": [],
"positions": []
}
seen = {
"hard_skills": set(),
"soft_skills": set(),
"tools": set(),
"positions": set()
}
for industry in industries:
industry_dict = self.get_dict(industry)
for category in merged.keys():
for item in industry_dict.get(category, []):
if item not in seen[category]:
merged[category].append(item)
seen[category].add(item)
logger.info(f"合并词典完成，共{sum(len(v) for v in merged.values())}个词条")
return merged
def add_custom_words(self, industry: str, category: str, words: List[str]):
"""
添加自定义词汇
Args:
industry: 行业名称
category: 类别（hard_skills/soft_skills/tools/positions）
words: 词汇列表
"""
if industry not in self.dicts:
self.dicts[industry] = {
"hard_skills": [], "soft_skills": [], "tools": [], "positions": []
}
seen = set(self.dicts[industry].get(category, []))
for word in words:
if word not in seen:
self.dicts[industry][category].append(word)
seen.add(word)
logger.info(f"已为{industry}行业添加{len(words)}个{category}词汇")
def demo_industry_dicts():
"""演示函数"""
dict_manager = IndustryDictionaries()
print("=== 数字文化行业硬技能 ===")
for skill in dict_manager.dicts["digital_culture"]["hard_skills"][:10]:
print(f" • {skill}")
print("\\n=== 科技产品行业工具 ===")
for tool in dict_manager.dicts["tech_product"]["tools"][:8]:
print(f" • {tool}")
print("\\n=== 合并数字文化+科技产品词典 ===")
merged = dict_manager.merge_dicts(["digital_culture", "tech_product"])
print(f" 硬技能总数: {len(merged['hard_skills'])}")
print(f" 工具总数: {len(merged['tools'])}")
if __name__ == "__main__":
demo_industry_dicts()
3. resume_parser.py – 简历解析模块
"""
简历解析模块
功能：支持PDF和DOCX格式，提取文本内容并保留基础结构
依赖：PyPDF2, python-docx
"""
import os
import re
from abc import ABC, abstractmethod
from typing import Dict, List, Optional, Tuple
import PyPDF2
from docx import Document
import logging
# 配置日志
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)
class ResumeParser(ABC):
"""
简历解析器抽象基类
定义解析器的标准接口
"""
@abstractmethod
def parse(self, file_path: str) -> Dict:
"""解析简历文件"""
pass
@abstractmethod
def extract_text(self, file_path: str) -> str:
"""提取纯文本"""
pass
class PDFResumeParser(ResumeParser):
"""
PDF格式简历解析器
"""
def __init__(self):
"""初始化PDF解析器"""
self.supported_extensions = ['.pdf']
def parse(self, file_path: str) -> Dict:
"""
解析PDF简历
Args:
file_path: PDF文件路径
Returns:
dict: 解析结果，包含文本、结构等信息
"""
if not self._validate_file(file_path):
return {"success": False, "error": "不支持的文件格式"}
try:
text = self.extract_text(file_path)
structured_data = self._extract_structured_data(text)
return {
"success": True,
"text": text,
"structured_data": structured_data,
"word_count": len(text),
"format": "PDF"
}
except Exception as e:
logger.error(f"PDF解析失败: {str(e)}")
return {"success": False, "error": str(e)}
def extract_text(self, file_path: str) -> str:
"""
从PDF提取文本
Args:
file_path: PDF文件路径
Returns:
str: 提取的文本内容
"""
text = ""
with open(file_path, 'rb') as file:
reader = PyPDF2.PdfReader(file)
for page_num, page in enumerate(reader.pages):
try:
page_text = page.extract_text()
if page_text:
text += f"\\n— 第{page_num + 1}页 —\\n"
text += page_text
except Exception as e:
logger.warning(f"第{page_num + 1}页提取失败: {str(e)}")
return text.strip()
def _validate_file(self, file_path: str) -> bool:
"""验证文件格式"""
_, ext = os.path.splitext(file_path)
return ext.lower() in self.supported_extensions
def _extract_structured_data(self, text: str) -> Dict:
"""
从文本中提取结构化数据
Args:
text: 纯文本
Returns:
dict: 结构化数据
"""
structured = {
"contact_info": self._extract_contact_info(text),
"sections": self._extract_sections(text),
"education": self._extract_education(text),
"experience": self._extract_experience(text)
}
return structured
def _extract_contact_info(self, text: str) -> Dict:
"""提取联系信息"""
contact = {
"name": "",
"email": "",
"phone": "",
"location": "",
"linkedin": ""
}
# 提取邮箱
email_pattern = r'[\\w\\.-]+@[\\w\\.-]+\\.\\w+'
emails = re.findall(email_pattern, text)
if emails:
contact["email"] = emails[0]
# 提取手机号
phone_pattern = r'1[3-9]\\d{9}|(\\+\\d{1,3}[- ]?)?\\(?\\d{1,4}\\)?[- ]?\\d{1,4}[- ]?\\d{1,9}'
phones = re.findall(phone_pattern, text)
if phones:
contact["phone"] = phones[0] if isinstance(phones[0], str) else phones[0][0]
# 提取姓名（假设在文档开头）
lines = text.split('\\n')
for line in lines[:5]:
line = line.strip()
if len(line) >= 2 and len(line) <= 20 and not re.search(r'[@\\d]', line):
contact["name"] = line
break
return contact
def _extract_sections(self, text: str) -> List[Dict]:
"""提取简历章节"""
sections = []
section_patterns = [
r'^(个人简介|自我介绍|Summary|Profile)',
r'^(工作经历|工作经验|Experience|Work History)',
r'^(教育背景|教育经历|Education|Academic)',
r'^(技能专长|专业技能|Skills|Competencies)',
r'^(项目经历|项目经验|Projects|Portfolio)',
r'^(获奖荣誉|Awards|Honors)',
r'^(自我评价|个人评价)'
]
compiled_patterns = [re.compile(p, re.I) for p in section_patterns]
lines = text.split('\\n')
current_section = None
for line in lines:
line = line.strip()
if not line:
continue
# 检查是否是章节标题
is_section = False
for pattern in compiled_patterns:
if pattern.match(line):
if current_section:
sections.append(current_section)
current_section = {
"title": line,
"content": []
}
is_section = True
break
if not is_section and current_section:
current_section["content"].append(line)
if current_section:
sections.append(current_section)
return sections
def _extract_education(self, text: str) -> List[Dict]:
"""提取教育背景"""
education = []
edu_pattern = r'([\\u4e00-\\u9fa5]+大学|[\\u4e00-\\u9fa5]+学院|University|College).*?(本科|硕士|博士|MBA|Bachelor|Master|PhD)'
matches = re.finditer(edu_pattern, text, re.I)
for match in matches:
edu_text = match.group(0)
education.append({
"raw_text": edu_text,
"institution": match.group(1),
"degree": match.group(2)
})
return education
def _extract_experience(self, text: str) -> List[Dict]:
"""提取工作经历"""
experience = []
exp_pattern = r'(\\d{4}[年/-]\\d{1,2})\\s*[至至-—]\\s*(\\d{4}[年/-]\\d{1,2}|至今|present)'
matches = list(re.finditer(exp_pattern, text))
for i, match in enumerate(matches):
start_date = match.group(1)
end_date = match.group(2)
# 提取时间段内的文本作为工作内容
start_pos = match.start()
end_pos = matches[i + 1].start() if i + 1 < len(matches) else len(text)
work_content = text[start_pos:end_pos]
experience.append({
"start_date": start_date,
"end_date": end_date,
"content": work_content.strip()
})
return experience
class DOCXResumeParser(ResumeParser):
"""
DOCX格式简历解析器
"""
def __init__(self):
"""初始化DOCX解析器"""
self.supported_extensions = ['.docx', '.doc']
def parse(self, file_path: str) -> Dict:
"""
解析DOCX简历
Args:
file_path: DOCX文件路径
Returns:
dict: 解析结果
"""
if not self._validate_file(file_path):
return {"success": False, "error": "不支持的文件格式"}
try:
doc = Document(file_path)
text = self.extract_text(file_path)
structured_data = self._extract_structured_data(doc, text)
return {
"success": True,
"text": text,
"structured_data": structured_data,
"word_count": len(text),
"format": "DOCX"
}
except Exception as e:
logger.error(f"DOCX解析失败: {str(e)}")
return {"success": False, "error": str(e)}
def extract_text(self, file_path: str) -> str:
"""
从DOCX提取文本
Args:
file_path: DOCX文件路径
Returns:
str: 提取的文本内容
"""
doc = Document(file_path)
text = ""
for para in doc.paragraphs:
text += para.text + "\\n"
# 提取表格中的文本
for table in doc.tables:
for row in table.rows:
for cell in row.cells:
text += cell.text + " "
text += "\\n"
return text.strip()
def _validate_file(self, file_path: str) -> bool:
"""验证文件格式"""
_, ext = os.path.splitext(file_path)
return ext.lower() in self.supported_extensions
def _extract_structured_data(self, doc, text: str) -> Dict:
"""从DOCX提取结构化数据"""
structured = {
"contact_info": self._extract_contact_info(text),
"sections": self._extract_sections(doc),
"education": [],
"experience": []
}
return structured
def _extract_contact_info(self, text: str) -> Dict:
"""提取联系信息（复用PDF解析器的方法）"""
contact = {
"name": "",
"email": "",
"phone": "",
"location": "",
"linkedin": ""
}
email_pattern = r'[\\w\\.-]+@[\\w\\.-]+\\.\\w+'
emails = re.findall(email_pattern, text)
if emails:
contact["email"] = emails[0]
phone_pattern = r'1[3-9]\\d{9}'
phones = re.findall(phone_pattern, text)
if phones:
contact["phone"] = phones[0]
lines = text.split('\\n')
for line in lines[:5]:
line = line.strip()
if len(line) >= 2 and len(line) <= 20 and not re.search(r'[@\\d]', line):
contact["name"] = line
break
return contact
def _extract_sections(self, doc) -> List[Dict]:
"""从DOCX提取章节"""
sections = []
current_section = None
for para in doc.paragraphs:
text = para.text.strip()
if not text:
continue
利用AI解决实际问题，如果你觉得这个工具好用，欢迎关注长安牧笛！
