#!/usr/bin/env python3 # -*- coding: utf-8 -*- """ 媒体叙事分析模块 - 人物画像技能 分析不同倾向媒体如何"框架"目标人物 创建时间：2026-04-03 开发者：瞰宇 (Kàn Yǔ) """ import json from typing import Dict, List, Tuple, Optional from dataclasses import dataclass, asdict from collections import Counter, defaultdict import re @dataclass class MediaFramework: """媒体框架""" media_name: str media_type: str # left, right, center, international dominant_frame: str # 主导框架 secondary_frame: Optional[str] # 次要框架 sentiment_distribution: Dict[str, float] # 情感分布 keywords: List[str] # 高频关键词 report_count: int # 报道数量 @dataclass class CrossMediaConsensus: """跨媒体共识""" consensus_labels: List[str] # 共识标签 divergence_points: List[str] # 分歧点 polarization_score: float # 极化分数（0-1） class MediaAnalyzer: """ 媒体叙事分析器 分析不同倾向媒体如何"框架"目标人物 """ # 媒体分类（示例） MEDIA_CLASSIFICATION = { "left": [ "纽约时报", "华盛顿邮报", "卫报", "CNN", "MSNBC", "New York Times", "Washington Post", "The Guardian" ], "right": [ "福克斯新闻", "华尔街日报(社论)", "布赖特巴特", "每日来电", "Fox News", "WSJ Opinion", "Breitbart", "Daily Caller" ], "center": [ "路透社", "美联社", "BBC", "华尔街日报(新闻)", "Reuters", "AP", "BBC News" ], "international": [ "BBC", "卫报", "每日电讯报", "金融时报", "BBC", "The Guardian", "Daily Telegraph", "Financial Times" ] } # 常见框架关键词（示例） FRAME_KEYWORDS = { "reform_leader": ["改革", "先锋", "创新", "变革", "改革者"], "trouble_maker": ["麻烦", "混乱", "破坏", "煽动", "极化"], "pragmatic": ["务实", "平衡", "稳健", "中间", "理性"], "populist": ["民粹", "人民", "反精英", "草根", "群众"] } def __init__(self): """初始化媒体分析器""" self.media_reports = [] self.frameworks = [] def load_media_reports(self, reports: List[Dict]): """ 加载媒体报道 Args: reports: 媒体报道列表 """ self.media_reports = reports def classify_media(self, media_name: str) -> str: """ 分类媒体倾向 Args: media_name: 媒体名称 Returns: 媒体类型（left/right/center/international） """ media_name_lower = media_name.lower() for media_type, media_list in self.MEDIA_CLASSIFICATION.items(): for media in media_list: if media.lower() in media_name_lower or media_name_lower in media.lower(): return media_type # 默认返回中间派 return "center" def analyze_media_framework(self, media_name: str, reports: List[Dict]) -> MediaFramework: """ 分析单个媒体的框架 Args: media_name: 媒体名称 reports: 该媒体的报道列表 Returns: MediaFramework对象 """ if not reports: return MediaFramework( media_name=media_name, media_type=self.classify_media(media_name), dominant_frame="无数据", secondary_frame=None, sentiment_distribution={}, keywords=[], report_count=0 ) # 识别主导框架 dominant_frame, secondary_frame = self._identify_dominant_frame(reports) # 分析情感分布 sentiment_dist = self._analyze_sentiment_distribution(reports) # 提取高频关键词 keywords = self._extract_keywords(reports) return MediaFramework( media_name=media_name, media_type=self.classify_media(media_name), dominant_frame=dominant_frame, secondary_frame=secondary_frame, sentiment_distribution=sentiment_dist, keywords=keywords[:10], # 取前10个 report_count=len(reports) ) def _identify_dominant_frame( self, reports: List[Dict] ) -> Tuple[str, (str, None)]: """ 识别主导框架 Args: reports: 报道列表 Returns: (主导框架, 次要框架) """ # 统计各框架的出现频率 frame_counts = Counter() for report in reports: content = report.get("content", "").lower() for frame, keywords in self.FRAME_KEYWORDS.items(): count = sum(1 for kw in keywords if kw in content) if count > 0: frame_counts[frame] += count if not frame_counts: return "无明确框架", None # 获取出现频率最高的框架 sorted_frames = frame_counts.most_common(2) dominant = sorted_frames[0][0] secondary = sorted_frames[1][0] if len(sorted_frames) > 1 else None return dominant, secondary def _analyze_sentiment_distribution( self, reports: List[Dict] ) -> Dict[str, float]: """ 分析情感分布 Args: reports: 报道列表 Returns: 情感分布字典 """ sentiment_counts = Counter() for report in reports: sentiment = report.get("sentiment", "neutral") sentiment_counts[sentiment] += 1 total = sum(sentiment_counts.values()) if total == 0: return {"positive": 0.0, "neutral": 0.0, "negative": 0.0} return { "positive": sentiment_counts["positive"] / total, "neutral": sentiment_counts["neutral"] / total, "negative": sentiment_counts["negative"] / total } def _extract_keywords(self, reports: List[Dict], top_n: int = 10) -> List[str]: """ 提取高频关键词 Args: reports: 报道列表 top_n: 返回前N个关键词 Returns: 关键词列表 """ # 合并所有报道内容 all_text = " ".join([r.get("content", "") for r in reports]) # 简单分词（实际应该使用更复杂的NLP方法） words = re.findall(r'\b[\w\u4e00-\u9fff]{2,}\b', all_text) # 统计词频 word_counts = Counter(words) # 过滤停用词（示例） stop_words = {"的", "了", "在", "是", "和", "有", "说", "the", "and", "of", "to", "in"} for stop_word in stop_words: word_counts.pop(stop_word, None) # 返回高频词 return [word for word, count in word_counts.most_common(top_n)] def analyze_all_media(self) -> List[MediaFramework]: """ 分析所有媒体的框架 Returns: MediaFramework对象列表 """ # 按媒体分组 media_grouped = defaultdict(list) for report in self.media_reports: media_name = report.get("media_name", "未知媒体") media_grouped[media_name].append(report) # 分析每个媒体的框架 self.frameworks = [] for media_name, reports in media_grouped.items(): framework = self.analyze_media_framework(media_name, reports) self.frameworks.append(framework) return self.frameworks def analyze_cross_media_consensus(self) -> CrossMediaConsensus: """ 分析跨媒体共识与分歧 Returns: CrossMediaConsensus对象 """ if not self.frameworks: return CrossMediaConsensus([], [], 0.0) # 识别所有媒体都提及的框架 frame_mentions = defaultdict(int) for framework in self.frameworks: if framework.dominant_frame != "无数据": frame_mentions[framework.dominant_frame] += 1 total_media = len(self.frameworks) consensus_labels = [ frame for frame, count in frame_mentions.items() if count >= total_media * 0.7 # 至少70%媒体提及 ] # 识别分歧点 unique_frames = set(f.dominant_frame for f in self.frameworks if f.dominant_frame != "无数据") divergence_points = list(unique_frames) # 计算极化分数（简化版：基于框架多样性） polarization_score = min(1.0, len(unique_frames) / total_media) return CrossMediaConsensus( consensus_labels=consensus_labels, divergence_points=divergence_points, polarization_score=polarization_score ) def generate_media_report(self) -> Dict: """ 生成媒体叙事分析报告 Returns: 分析报告字典 """ frameworks = self.analyze_all_media() consensus = self.analyze_cross_media_consensus() # 按媒体类型分组 by_type = defaultdict(list) for framework in frameworks: by_type[framework.media_type].append({ "媒体名称": framework.media_name, "主导框架": framework.dominant_frame, "情感分布": framework.sentiment_distribution, "高频关键词": framework.keywords[:5], "报道数量": framework.report_count }) report = { "媒体框架分析": { "左翼媒体": by_type["left"], "右翼媒体": by_type["right"], "中间派媒体": by_type["center"], "国际媒体": by_type["international"] }, "跨媒体共识": { "共识标签": consensus.consensus_labels, "分歧点": consensus.divergence_points, "极化分数": f"{consensus.polarization_score:.2f}" } } return report def main(): """测试媒体分析器""" analyzer = MediaAnalyzer() # 示例媒体报道数据 reports = [ { "media_name": "纽约时报", "title": "民粹主义的兴起", "content": "极右翼煽动者正在破坏民主制度，制造社会分裂。", "date": "2024-01-01", "sentiment": "negative" }, { "media_name": "福克斯新闻", "title": "爱国反叛者", "content": "捍卫自由的爱国者，人民的代言人。", "date": "2024-01-02", "sentiment": "positive" }, { "media_name": "路透社", "title": "争议政治人物", "content": "一位极具争议的政治人物，引发了广泛讨论。", "date": "2024-01-03", "sentiment": "neutral" } ] analyzer.load_media_reports(reports) # 生成报告 print("=== 媒体叙事分析报告 ===") report = analyzer.generate_media_report() print(json.dumps(report, ensure_ascii=False, indent=2)) if __name__ == "__main__": main()