""" 人群包分析脚本：解析G1-G5 RTF/TXT文件，输出洞察JSON 用法：python3 analyze_crowd.py """ import json, sys, os def read_file(path): for enc in ['utf-8', 'gbk', 'gb2312', 'gb18030']: try: with open(path, 'r', encoding=enc) as f: return f.read() except (UnicodeDecodeError, LookupError): continue raise ValueError(f"无法解码: {path}") def extract_json(text): start = text.find('{') if start == -1: return None depth, end = 0, start for i in range(start, len(text)): if text[i] == '{': depth += 1 elif text[i] == '}': depth -= 1 if depth == 0: end = i + 1 break return json.loads(text[start:end]) def parse_report(data): """从DMP JSON中提取reportDetail""" if isinstance(data, dict): if 'data' in data and isinstance(data['data'], dict): inner = data['data'] if 'data' in inner and isinstance(inner['data'], dict): return inner['data'].get('reportDetail', []) return inner.get('reportDetail', []) return [] def get_tag_values(report, tag_en): """获取指定标签的值分布""" for item in report: if item.get('tagEn') == tag_en: return item.get('tagValueCoverResult', []) return [] def calc_tgi(g5_ratio, g1_ratio): if g1_ratio <= 0: return 0 return round(g5_ratio / g1_ratio * 100, 1) def analyze(files): stages = ['G1', 'G2', 'G3', 'G4', 'G5'] reports = {} total_users = {} for i, f in enumerate(files): text = read_file(f) data = extract_json(text) if not data: raise ValueError(f"无法解析JSON: {f}") report = parse_report(data) stage = stages[i] reports[stage] = report # 提取总人数 if isinstance(data, dict) and 'data' in data: inner = data['data'] if isinstance(inner, dict) and 'data' in inner: total_users[stage] = inner['data'].get('totalNum', 0) else: total_users[stage] = inner.get('totalNum', 0) else: total_users[stage] = 0 result = { 'total_users': total_users, 'funnel_rate': round(total_users.get('G5', 0) / total_users.get('G1', 1) * 100, 2), 'demographics': {}, 'strategy_crowd': {}, 'navigation_top': {}, 'key_insights': [] } # 性别分析 gender_tags = ['gender'] for tag in gender_tags: g1_vals = get_tag_values(reports['G1'], tag) g5_vals = get_tag_values(reports['G5'], tag) if g1_vals and g5_vals: gender_data = [] for v1 in g1_vals: name = v1.get('tagValueCn', v1.get('tagValue', '')) r1 = v1.get('coverRate', 0) r5 = next((v.get('coverRate', 0) for v in g5_vals if v.get('tagValueCn', v.get('tagValue', '')) == name), 0) gender_data.append({ 'name': name, 'g1': r1, 'g5': r5, 'tgi': calc_tgi(r5, r1), 'change': round(r5 - r1, 2) }) result['demographics']['gender'] = gender_data # 年龄分析 for tag in ['age', 'age_range']: g1_vals = get_tag_values(reports['G1'], tag) g5_vals = get_tag_values(reports['G5'], tag) if g1_vals and g5_vals: age_data = [] for v1 in g1_vals: name = v1.get('tagValueCn', v1.get('tagValue', '')) r1 = v1.get('coverRate', 0) r5 = next((v.get('coverRate', 0) for v in g5_vals if v.get('tagValueCn', v.get('tagValue', '')) == name), 0) age_data.append({ 'name': name, 'g1': r1, 'g5': r5, 'tgi': calc_tgi(r5, r1), 'change_pct': round((r5 - r1) / r1 * 100, 1) if r1 > 0 else 0 }) result['demographics']['age'] = sorted(age_data, key=lambda x: -x['tgi']) break # 人生阶段 for tag in ['life_stage', 'lifeStage']: g1_vals = get_tag_values(reports['G1'], tag) g5_vals = get_tag_values(reports['G5'], tag) if g1_vals and g5_vals: ls_data = [] for v1 in g1_vals: name = v1.get('tagValueCn', v1.get('tagValue', '')) r1 = v1.get('coverRate', 0) r5 = next((v.get('coverRate', 0) for v in g5_vals if v.get('tagValueCn', v.get('tagValue', '')) == name), 0) ls_data.append({ 'name': name, 'g1': r1, 'g5': r5, 'tgi': calc_tgi(r5, r1) }) result['demographics']['life_stage'] = sorted(ls_data, key=lambda x: -x['tgi']) break # 八大策略人群 for tag in ['strategy_crowd', 'strategyCrowd', 'crowd_strategy']: g1_vals = get_tag_values(reports['G1'], tag) g5_vals = get_tag_values(reports['G5'], tag) if g1_vals and g5_vals: sc_data = [] for v1 in g1_vals: name = v1.get('tagValueCn', v1.get('tagValue', '')) r1 = v1.get('coverRate', 0) r5 = next((v.get('coverRate', 0) for v in g5_vals if v.get('tagValueCn', v.get('tagValue', '')) == name), 0) sc_data.append({ 'name': name, 'g1': r1, 'g5': r5, 'tgi': calc_tgi(r5, r1) }) result['strategy_crowd'] = sorted(sc_data, key=lambda x: -x['tgi']) break # 导航TOP场所 for tag in ['navigation_top20', 'navigation_top', 'nav_top20']: g1_vals = get_tag_values(reports['G1'], tag) g5_vals = get_tag_values(reports['G5'], tag) if g1_vals and g5_vals: nav_data = [] for v1 in g1_vals: name = v1.get('tagValueCn', v1.get('tagValue', '')) r1 = v1.get('coverRate', 0) r5 = next((v.get('coverRate', 0) for v in g5_vals if v.get('tagValueCn', v.get('tagValue', '')) == name), 0) nav_data.append({ 'name': name, 'g1': r1, 'g5': r5, 'tgi': calc_tgi(r5, r1) }) result['navigation_top'] = sorted(nav_data, key=lambda x: -x['tgi']) break # 生成关键洞察 insights = [] if 'gender' in result['demographics']: for g in result['demographics']['gender']: if g['tgi'] > 100: insights.append(f"{g['name']}转化倾向更强(TGI={g['tgi']})") if 'age' in result['demographics']: top_age = result['demographics']['age'][0] insights.append(f"年龄段{top_age['name']}增幅最大(TGI={top_age['tgi']})") if 'life_stage' in result['demographics']: top_ls = result['demographics']['life_stage'][0] insights.append(f"'{top_ls['name']}'TGI={top_ls['tgi']}，G5过表达") if result.get('strategy_crowd'): high_sc = [s for s in result['strategy_crowd'] if s['tgi'] > 100] if high_sc: insights.append(f"策略人群'{high_sc[0]['name']}'G5富集(TGI={high_sc[0]['tgi']})") if result.get('navigation_top'): top_nav = result['navigation_top'][0] insights.append(f"导航'{top_nav['name']}'TGI={top_nav['tgi']}，最强转化信号") result['key_insights'] = insights return result if __name__ == '__main__': if len(sys.argv) < 7: print("Usage: python3 analyze_crowd.py g1.txt g2.txt g3.txt g4.txt g5.txt output.json") sys.exit(1) files = sys.argv[1:6] output = sys.argv[6] result = analyze(files) with open(output, 'w', encoding='utf-8') as f: json.dump(result, f, ensure_ascii=False, indent=2) print(f"分析完成，输出: {output}")