393 lines
14 KiB
Python
393 lines
14 KiB
Python
|
"""
|
|||
|
|
E-E-A-T 强化模块
|
|||
|
|
Expertise（专业性）、Experience（经验性）、Authoritativeness（权威性）、Trustworthiness（可信度）
|
|||
|
|
"""
|
|||
|
|
from typing import Dict, List, Optional
|
|||
|
|
from langchain_core.prompts import PromptTemplate
|
|||
|
|
from langchain_core.output_parsers import StrOutputParser
|
|||
|
|
import json
|
|||
|
|
import re
|
|||
|
|
|
|||
|
|
|
|||
|
|
class EEATEnhancer:
|
|||
|
|
"""E-E-A-T 强化器"""
|
|||
|
|
|
|||
|
|
def __init__(self):
|
|||
|
|
# E-E-A-T 评估 Prompt
|
|||
|
|
self.assessment_prompt_template = """
|
|||
|
|
你是一名内容质量评估专家，专门评估内容的 E-E-A-T（专业性、经验性、权威性、可信度）水平。
|
|||
|
|
|
|||
|
|
【内容】
|
|||
|
|
{content}
|
|||
|
|
|
|||
|
|
【品牌】{brand}
|
|||
|
|
【优势】{advantages}
|
|||
|
|
【平台】{platform}
|
|||
|
|
|
|||
|
|
【E-E-A-T 评估标准】
|
|||
|
|
|
|||
|
|
1. **Expertise（专业性）**（25分）
|
|||
|
|
- 内容是否展示深度的专业知识？
|
|||
|
|
- 是否使用专业术语和准确的技术描述？
|
|||
|
|
- 是否提供专业见解和分析？
|
|||
|
|
- 作者是否表现出对该领域的专业理解？
|
|||
|
|
|
|||
|
|
2. **Experience（经验性）**（25分）
|
|||
|
|
- 是否包含实际使用经验或案例？
|
|||
|
|
- 是否有第一手体验描述？
|
|||
|
|
- 是否分享实践中的洞察和教训？
|
|||
|
|
- 是否有"我使用过"、"实际测试"等经验性表述？
|
|||
|
|
|
|||
|
|
3. **Authoritativeness（权威性）**（25分）
|
|||
|
|
- 是否引用权威来源或数据？
|
|||
|
|
- 是否提及行业标准、研究报告或官方文档？
|
|||
|
|
- 是否有明确的来源占位建议（如"根据XX报告"、"参考XX标准"）？
|
|||
|
|
- 内容是否建立在该领域的权威知识基础上？
|
|||
|
|
|
|||
|
|
4. **Trustworthiness（可信度）**（25分）
|
|||
|
|
- 是否避免编造数据或虚假信息？
|
|||
|
|
- 是否明确标注不确定信息（如"据公开资料"、"建议参考"）？
|
|||
|
|
- 是否提供可验证的信息？
|
|||
|
|
- 内容是否诚实、透明、负责任？
|
|||
|
|
|
|||
|
|
【来源占位检查】
|
|||
|
|
请检查内容中是否包含以下来源占位元素：
|
|||
|
|
- 数据来源占位（如"根据XX行业报告"、"XX数据显示"）
|
|||
|
|
- 案例来源占位（如"某企业案例"、"参考XX实践"）
|
|||
|
|
- 标准来源占位（如"按照XX标准"、"参考XX规范"）
|
|||
|
|
- 专家观点占位（如"行业专家认为"、"XX机构指出"）
|
|||
|
|
|
|||
|
|
【输出格式】
|
|||
|
|
请严格按照以下 JSON 格式输出，不要添加任何其他内容：
|
|||
|
|
|
|||
|
|
{{
|
|||
|
|
"eeat_scores": {{
|
|||
|
|
"expertise": <专业性得分 0-25>,
|
|||
|
|
"experience": <经验性得分 0-25>,
|
|||
|
|
"authoritativeness": <权威性得分 0-25>,
|
|||
|
|
"trustworthiness": <可信度得分 0-25>,
|
|||
|
|
"total": <总分 0-100>
|
|||
|
|
}},
|
|||
|
|
"source_placeholders": {{
|
|||
|
|
"data_sources": ["<数据来源占位1>", "<数据来源占位2>"],
|
|||
|
|
"case_sources": ["<案例来源占位1>"],
|
|||
|
|
"standard_sources": ["<标准来源占位1>"],
|
|||
|
|
"expert_opinions": ["<专家观点占位1>"]
|
|||
|
|
}},
|
|||
|
|
"details": {{
|
|||
|
|
"expertise": "<专业性评估详情>",
|
|||
|
|
"experience": "<经验性评估详情>",
|
|||
|
|
"authoritativeness": "<权威性评估详情>",
|
|||
|
|
"trustworthiness": "<可信度评估详情>"
|
|||
|
|
}},
|
|||
|
|
"improvements": [
|
|||
|
|
"<改进建议1>",
|
|||
|
|
"<改进建议2>",
|
|||
|
|
"<改进建议3>"
|
|||
|
|
],
|
|||
|
|
"source_suggestions": [
|
|||
|
|
"<来源占位建议1>",
|
|||
|
|
"<来源占位建议2>"
|
|||
|
|
]
|
|||
|
|
}}
|
|||
|
|
|
|||
|
|
【开始评估】
|
|||
|
|
"""
|
|||
|
|
|
|||
|
|
# E-E-A-T 强化 Prompt
|
|||
|
|
self.enhancement_prompt_template = """
|
|||
|
|
你是一名内容优化专家，专门提升内容的 E-E-A-T（专业性、经验性、权威性、可信度）水平。
|
|||
|
|
|
|||
|
|
【原内容】
|
|||
|
|
{content}
|
|||
|
|
|
|||
|
|
【品牌】{brand}
|
|||
|
|
【优势】{advantages}
|
|||
|
|
【平台】{platform}
|
|||
|
|
|
|||
|
|
【E-E-A-T 强化要求】
|
|||
|
|
|
|||
|
|
1. **Expertise（专业性）强化**
|
|||
|
|
- 增加专业术语和准确的技术描述
|
|||
|
|
- 提供深度的专业见解和分析
|
|||
|
|
- 展示对该领域的专业理解
|
|||
|
|
- 使用行业标准术语
|
|||
|
|
|
|||
|
|
2. **Experience（经验性）强化**
|
|||
|
|
- 添加实际使用经验或案例描述
|
|||
|
|
- 包含第一手体验（如"实际测试发现"、"使用中观察到"）
|
|||
|
|
- 分享实践中的洞察和教训
|
|||
|
|
- 增加"基于实际使用"、"经过验证"等表述
|
|||
|
|
|
|||
|
|
3. **Authoritativeness（权威性）强化**
|
|||
|
|
- 添加权威来源占位（如"根据XX行业报告"、"参考XX研究"）
|
|||
|
|
- 提及行业标准、规范或官方文档
|
|||
|
|
- 引用权威机构或专家的观点（用占位符）
|
|||
|
|
- 建立内容在权威知识基础上的表述
|
|||
|
|
|
|||
|
|
4. **Trustworthiness（可信度）强化**
|
|||
|
|
- 明确标注不确定信息（如"据公开资料显示"、"建议参考官方文档"）
|
|||
|
|
- 避免编造具体数据，使用占位建议
|
|||
|
|
- 提供可验证的信息来源占位
|
|||
|
|
- 保持诚实、透明、负责任的表述
|
|||
|
|
|
|||
|
|
【来源占位要求】
|
|||
|
|
必须在内容中添加以下类型的来源占位（用占位符形式，不要编造真实来源）：
|
|||
|
|
|
|||
|
|
1. **数据来源占位**（至少2处）
|
|||
|
|
- 格式："根据XX行业报告"、"XX数据显示"、"据XX统计"
|
|||
|
|
- 示例："根据2024年外贸软件行业报告显示"、"据公开市场调研数据显示"
|
|||
|
|
|
|||
|
|
2. **案例来源占位**（至少1处）
|
|||
|
|
- 格式："某企业案例"、"参考XX实践"、"XX公司案例"
|
|||
|
|
- 示例："参考某大型外贸企业的实际应用案例"、"某知名企业的成功实践表明"
|
|||
|
|
|
|||
|
|
3. **标准来源占位**（至少1处）
|
|||
|
|
- 格式："按照XX标准"、"参考XX规范"、"符合XX要求"
|
|||
|
|
- 示例："按照ISO质量管理体系标准"、"参考行业最佳实践规范"
|
|||
|
|
|
|||
|
|
4. **专家观点占位**（可选，1处）
|
|||
|
|
- 格式："行业专家认为"、"XX机构指出"、"权威分析显示"
|
|||
|
|
- 示例："行业专家普遍认为"、"权威机构分析指出"
|
|||
|
|
|
|||
|
|
【输出格式】
|
|||
|
|
请输出两部分：
|
|||
|
|
|
|||
|
|
【E-E-A-T 强化后内容】
|
|||
|
|
（完整的优化后内容，保持原意和结构，但增强 E-E-A-T 元素）
|
|||
|
|
|
|||
|
|
【来源占位清单】
|
|||
|
|
（列出所有添加的来源占位，格式：类型 - 占位内容）
|
|||
|
|
|
|||
|
|
【开始优化】
|
|||
|
|
"""
|
|||
|
|
|
|||
|
|
def assess_eeat(self, content: str, brand: str, advantages: str,
|
|||
|
|
platform: str, llm_chain) -> Dict:
|
|||
|
|
"""
|
|||
|
|
评估内容的 E-E-A-T 水平
|
|||
|
|
|
|||
|
|
Args:
|
|||
|
|
content: 要评估的内容
|
|||
|
|
brand: 品牌名称
|
|||
|
|
advantages: 品牌优势
|
|||
|
|
platform: 发布平台
|
|||
|
|
llm_chain: LangChain 链对象
|
|||
|
|
|
|||
|
|
Returns:
|
|||
|
|
包含 E-E-A-T 评分、来源占位检查和改进建议的字典
|
|||
|
|
"""
|
|||
|
|
try:
|
|||
|
|
prompt = PromptTemplate.from_template(self.assessment_prompt_template)
|
|||
|
|
chain = prompt | llm_chain | StrOutputParser()
|
|||
|
|
|
|||
|
|
result = chain.invoke({
|
|||
|
|
"content": content,
|
|||
|
|
"brand": brand,
|
|||
|
|
"advantages": advantages,
|
|||
|
|
"platform": platform
|
|||
|
|
})
|
|||
|
|
|
|||
|
|
# 解析结果
|
|||
|
|
assessment_data = self._parse_assessment_result(result)
|
|||
|
|
|
|||
|
|
return assessment_data
|
|||
|
|
|
|||
|
|
except Exception as e:
|
|||
|
|
# 如果评估失败，返回默认数据
|
|||
|
|
return {
|
|||
|
|
"eeat_scores": {
|
|||
|
|
"expertise": 0,
|
|||
|
|
"experience": 0,
|
|||
|
|
"authoritativeness": 0,
|
|||
|
|
"trustworthiness": 0,
|
|||
|
|
"total": 0
|
|||
|
|
},
|
|||
|
|
"source_placeholders": {
|
|||
|
|
"data_sources": [],
|
|||
|
|
"case_sources": [],
|
|||
|
|
"standard_sources": [],
|
|||
|
|
"expert_opinions": []
|
|||
|
|
},
|
|||
|
|
"details": {
|
|||
|
|
"expertise": f"评估失败：{str(e)}",
|
|||
|
|
"experience": "",
|
|||
|
|
"authoritativeness": "",
|
|||
|
|
"trustworthiness": ""
|
|||
|
|
},
|
|||
|
|
"improvements": ["E-E-A-T 评估系统暂时无法评估此内容，请手动检查"],
|
|||
|
|
"source_suggestions": []
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
def enhance_eeat(self, content: str, brand: str, advantages: str,
|
|||
|
|
platform: str, llm_chain) -> Dict:
|
|||
|
|
"""
|
|||
|
|
强化内容的 E-E-A-T 水平
|
|||
|
|
|
|||
|
|
Args:
|
|||
|
|
content: 要强化的内容
|
|||
|
|
brand: 品牌名称
|
|||
|
|
advantages: 品牌优势
|
|||
|
|
platform: 发布平台
|
|||
|
|
llm_chain: LangChain 链对象
|
|||
|
|
|
|||
|
|
Returns:
|
|||
|
|
包含强化后内容和来源占位清单的字典
|
|||
|
|
"""
|
|||
|
|
try:
|
|||
|
|
prompt = PromptTemplate.from_template(self.enhancement_prompt_template)
|
|||
|
|
chain = prompt | llm_chain | StrOutputParser()
|
|||
|
|
|
|||
|
|
result = chain.invoke({
|
|||
|
|
"content": content,
|
|||
|
|
"brand": brand,
|
|||
|
|
"advantages": advantages,
|
|||
|
|
"platform": platform
|
|||
|
|
})
|
|||
|
|
|
|||
|
|
# 解析结果
|
|||
|
|
enhanced_data = self._parse_enhancement_result(result)
|
|||
|
|
|
|||
|
|
return enhanced_data
|
|||
|
|
|
|||
|
|
except Exception as e:
|
|||
|
|
return {
|
|||
|
|
"enhanced_content": content,
|
|||
|
|
"source_placeholders": [],
|
|||
|
|
"changes": f"E-E-A-T 强化失败：{str(e)}"
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
def _parse_assessment_result(self, result: str) -> Dict:
|
|||
|
|
"""解析评估结果"""
|
|||
|
|
# 尝试提取 JSON
|
|||
|
|
json_match = re.search(r'\{.*\}', result, re.DOTALL)
|
|||
|
|
if json_match:
|
|||
|
|
try:
|
|||
|
|
data = json.loads(json_match.group())
|
|||
|
|
# 验证数据结构
|
|||
|
|
if "eeat_scores" in data and "total" in data["eeat_scores"]:
|
|||
|
|
return data
|
|||
|
|
except json.JSONDecodeError:
|
|||
|
|
pass
|
|||
|
|
|
|||
|
|
# 如果无法解析 JSON，尝试从文本中提取信息
|
|||
|
|
return self._extract_assessment_from_text(result)
|
|||
|
|
|
|||
|
|
def _extract_assessment_from_text(self, text: str) -> Dict:
|
|||
|
|
"""从文本中提取评估信息（备用方案）"""
|
|||
|
|
# 尝试提取总分
|
|||
|
|
total_match = re.search(r'总分[：:]\s*(\d+)', text)
|
|||
|
|
total_score = int(total_match.group(1)) if total_match else 0
|
|||
|
|
|
|||
|
|
# 简单分配分数
|
|||
|
|
avg_score = total_score // 4 if total_score > 0 else 0
|
|||
|
|
|
|||
|
|
return {
|
|||
|
|
"eeat_scores": {
|
|||
|
|
"expertise": avg_score,
|
|||
|
|
"experience": avg_score,
|
|||
|
|
"authoritativeness": avg_score,
|
|||
|
|
"trustworthiness": avg_score,
|
|||
|
|
"total": total_score
|
|||
|
|
},
|
|||
|
|
"source_placeholders": {
|
|||
|
|
"data_sources": [],
|
|||
|
|
"case_sources": [],
|
|||
|
|
"standard_sources": [],
|
|||
|
|
"expert_opinions": []
|
|||
|
|
},
|
|||
|
|
"details": {
|
|||
|
|
"expertise": "无法解析详细评估",
|
|||
|
|
"experience": "无法解析详细评估",
|
|||
|
|
"authoritativeness": "无法解析详细评估",
|
|||
|
|
"trustworthiness": "无法解析详细评估"
|
|||
|
|
},
|
|||
|
|
"improvements": ["请检查内容是否符合 E-E-A-T 原则"],
|
|||
|
|
"source_suggestions": []
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
def _parse_enhancement_result(self, result: str) -> Dict:
|
|||
|
|
"""解析强化结果"""
|
|||
|
|
enhanced_content = ""
|
|||
|
|
source_placeholders = []
|
|||
|
|
changes = ""
|
|||
|
|
|
|||
|
|
# 提取强化后内容
|
|||
|
|
if "【E-E-A-T 强化后内容】" in result:
|
|||
|
|
parts = result.split("【E-E-A-T 强化后内容】", 1)
|
|||
|
|
if len(parts) > 1:
|
|||
|
|
content_part = parts[1]
|
|||
|
|
if "【来源占位清单】" in content_part:
|
|||
|
|
enhanced_content = content_part.split("【来源占位清单】", 1)[0].strip()
|
|||
|
|
else:
|
|||
|
|
enhanced_content = content_part.strip()
|
|||
|
|
|
|||
|
|
# 提取来源占位清单
|
|||
|
|
if "【来源占位清单】" in result:
|
|||
|
|
placeholder_part = result.split("【来源占位清单】", 1)[1].strip()
|
|||
|
|
# 按行解析占位清单
|
|||
|
|
for line in placeholder_part.split("\n"):
|
|||
|
|
line = line.strip()
|
|||
|
|
if line and "-" in line:
|
|||
|
|
source_placeholders.append(line)
|
|||
|
|
|
|||
|
|
# 如果没有找到明确的分隔符，尝试其他方式
|
|||
|
|
if not enhanced_content:
|
|||
|
|
# 尝试提取整个结果作为内容
|
|||
|
|
enhanced_content = result.strip()
|
|||
|
|
|
|||
|
|
return {
|
|||
|
|
"enhanced_content": enhanced_content,
|
|||
|
|
"source_placeholders": source_placeholders,
|
|||
|
|
"changes": f"已添加 {len(source_placeholders)} 个来源占位" if source_placeholders else "未检测到明确的来源占位"
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
def get_eeat_level(self, total_score: int) -> tuple:
|
|||
|
|
"""
|
|||
|
|
根据 E-E-A-T 总分返回等级和颜色
|
|||
|
|
|
|||
|
|
Returns:
|
|||
|
|
(等级名称, 颜色代码)
|
|||
|
|
"""
|
|||
|
|
if total_score >= 90:
|
|||
|
|
return ("优秀", "#10B981") # 绿色
|
|||
|
|
elif total_score >= 75:
|
|||
|
|
return ("良好", "#3B82F6") # 蓝色
|
|||
|
|
elif total_score >= 60:
|
|||
|
|
return ("中等", "#F59E0B") # 橙色
|
|||
|
|
else:
|
|||
|
|
return ("需改进", "#EF4444") # 红色
|
|||
|
|
|
|||
|
|
def get_quick_eeat_check(self, content: str) -> Dict:
|
|||
|
|
"""
|
|||
|
|
快速 E-E-A-T 检查（不调用 LLM，基于规则）
|
|||
|
|
用于在 LLM 评估前提供初步检查
|
|||
|
|
"""
|
|||
|
|
check = {
|
|||
|
|
"has_professional_terms": bool(re.search(r'(标准|规范|体系|框架|方法论|最佳实践)', content)),
|
|||
|
|
"has_experience_words": bool(re.search(r'(实际|使用|测试|验证|经验|实践|案例)', content)),
|
|||
|
|
"has_source_placeholders": bool(re.search(r'(根据|参考|据|按照|符合|显示|表明|指出)', content)),
|
|||
|
|
"has_data_placeholders": bool(re.search(r'(报告|数据|统计|调研|分析)', content)),
|
|||
|
|
"has_uncertainty_markers": bool(re.search(r'(建议|可能|通常|一般|据公开|参考)', content)),
|
|||
|
|
"source_placeholder_count": len(re.findall(r'(根据|参考|据|按照|显示|表明)', content))
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
# 计算初步分数
|
|||
|
|
quick_score = 0
|
|||
|
|
if check["has_professional_terms"]:
|
|||
|
|
quick_score += 5
|
|||
|
|
if check["has_experience_words"]:
|
|||
|
|
quick_score += 5
|
|||
|
|
if check["has_source_placeholders"]:
|
|||
|
|
quick_score += 5
|
|||
|
|
if check["has_data_placeholders"]:
|
|||
|
|
quick_score += 5
|
|||
|
|
if check["has_uncertainty_markers"]:
|
|||
|
|
quick_score += 5
|
|||
|
|
if check["source_placeholder_count"] >= 3:
|
|||
|
|
quick_score += 5
|
|||
|
|
|
|||
|
|
check["quick_score"] = min(quick_score, 30) # 最高30分（快速检查）
|
|||
|
|
|
|||
|
|
return check
|