| 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203 |
- // ==============================================================================
- // <品类名> · VOC 数据分析模块(模板)
- // ==============================================================================
- // 职责:
- // - 加载 _merged.json
- // - 提供 getEvidence() 主查询 API
- // - 提供 filterByPlatform/Hypothesis/Product 等辅助筛选
- // - 判断 "substantive" 内容(过滤纯表情/水字)
- // ==============================================================================
- const fs = require('fs');
- const path = require('path');
- const CATEGORY = '<品类>'; // TODO: 改为实际目录名
- // ==========================================================
- // 1. 常量
- // ==========================================================
- const HYPOTHESES = {
- H1: '<H1 主题>', // TODO: 按 2.VOC深度思路 填写
- H2: '<H2 主题>',
- H3: '<H3 主题>',
- H4: '<H4 主题>',
- H5: '<H5 主题>',
- H6: '<H6 主题>',
- H7: '<H7 主题>',
- H8: '<H8 主题>',
- };
- const PLATFORMS = ['xhs', 'douyin', 'weibo', 'jd', 'tmall', 'amazon'];
- // ==========================================================
- // 2. 数据加载
- // ==========================================================
- function loadMerged() {
- const fp = path.resolve(__dirname, '..', '..', 'docs', CATEGORY, 'raw', '_merged.json');
- if (!fs.existsSync(fp)) {
- throw new Error(`_merged.json 不存在。请先跑 collect --merge。路径:${fp}`);
- }
- const data = JSON.parse(fs.readFileSync(fp, 'utf8'));
- if (!data.items || !Array.isArray(data.items)) {
- throw new Error(`_merged.json 格式错误:缺少 items 数组`);
- }
- return data;
- }
- // ==========================================================
- // 3. 内容实质性判断
- // ==========================================================
- function isSubstantive(content, minChars = 10) {
- const s = String(content || '').trim();
- if (s.length < minChars) return false;
- // 剥离 [表情] 标记和 unicode 表情
- const stripped = s
- .replace(/\[[^\]]+\]/g, '')
- .replace(/[\s\p{P}\p{Emoji_Presentation}\p{Extended_Pictographic}]/gu, '');
- return stripped.length >= Math.max(4, Math.floor(minChars / 2));
- }
- // ==========================================================
- // 4. 基础筛选
- // ==========================================================
- const filterByPlatform = (items, p) => items.filter((it) => it.platform === p);
- const filterByHypothesis = (items, h) => items.filter((it) => (it.hypotheses || []).includes(h));
- const filterByProduct = (items, prod) => items.filter((it) => it.product === prod);
- const filterByKeyword = (items, kw) => items.filter((it) => it.keyword === kw);
- const filterBySentiment = (items, s) => items.filter((it) => it.sentiment === s);
- // ==========================================================
- // 5. 内容匹配
- // ==========================================================
- function filterByContentMatch(items, regex) {
- if (!regex) return items;
- const re = regex instanceof RegExp ? regex : new RegExp(regex);
- return items.filter((it) => re.test(it.content || ''));
- }
- // ==========================================================
- // 6. 核心 API · getEvidence
- // ==========================================================
- /**
- * 从 items 池里取出证据卡
- * @param {Array<Item>} items
- * @param {Object} opts
- * @param {string} [opts.keyword]
- * @param {string} [opts.product]
- * @param {string} [opts.platform]
- * @param {RegExp|string} [opts.contentMatch]
- * @param {Array<string>} [opts.hypotheses] - 必含的 H 标签(AND 关系)
- * @param {string} [opts.sentiment] - positive / negative / neutral
- * @param {number} [opts.minChars=10]
- * @param {number} [opts.minLikes=0]
- * @param {number} [opts.top=5]
- * @param {number} [opts.seed=0] - 乱序种子(避免章节重复)
- * @returns {Array<Item>}
- */
- function getEvidence(items, opts = {}) {
- let pool = items;
- if (opts.keyword) pool = filterByKeyword(pool, opts.keyword);
- if (opts.product) pool = filterByProduct(pool, opts.product);
- if (opts.platform) pool = filterByPlatform(pool, opts.platform);
- if (opts.sentiment) pool = filterBySentiment(pool, opts.sentiment);
- if (opts.contentMatch) pool = filterByContentMatch(pool, opts.contentMatch);
- if (opts.hypotheses?.length) {
- pool = pool.filter((it) => opts.hypotheses.every((h) => (it.hypotheses || []).includes(h)));
- }
- const minChars = opts.minChars ?? 10;
- pool = pool.filter((it) => isSubstantive(it.content, minChars));
- if (opts.minLikes) pool = pool.filter((it) => (it.likes || 0) >= opts.minLikes);
- // 去重:前 40 字相同视为同内容
- const seen = new Set();
- pool = pool.filter((it) => {
- const key = String(it.content || '').slice(0, 40);
- if (seen.has(key)) return false;
- seen.add(key);
- return true;
- });
- // 按 likes 降序
- pool.sort((a, b) => (b.likes || 0) - (a.likes || 0));
- const top = opts.top || 5;
- // 取前 top*2 再按 seed 乱序选 top(避免每章都抽到同几条)
- const top2 = pool.slice(0, top * 2);
- if (opts.seed) return shuffleWithSeed(top2, opts.seed).slice(0, top);
- return top2.slice(0, top);
- }
- // ==========================================================
- // 7. 工具
- // ==========================================================
- function shuffleWithSeed(arr, seed) {
- const result = arr.slice();
- let rnd = seed || 1;
- for (let i = result.length - 1; i > 0; i--) {
- rnd = (rnd * 9301 + 49297) % 233280;
- const j = Math.floor((rnd / 233280) * (i + 1));
- [result[i], result[j]] = [result[j], result[i]];
- }
- return result;
- }
- function fmtLikes(n) {
- if (!n) return '0';
- if (n >= 10000) return (n / 10000).toFixed(1) + 'w';
- if (n >= 1000) return (n / 1000).toFixed(1) + 'k';
- return String(n);
- }
- // ==========================================================
- // 8. 聚合统计
- // ==========================================================
- function statsByPlatform(items) {
- const map = {};
- for (const it of items) map[it.platform] = (map[it.platform] || 0) + 1;
- return map;
- }
- function statsByHypothesis(items) {
- const map = {};
- for (const it of items) {
- for (const h of it.hypotheses || []) map[h] = (map[h] || 0) + 1;
- }
- return map;
- }
- function statsByProduct(items) {
- const map = {};
- for (const it of items) map[it.product] = (map[it.product] || 0) + 1;
- return map;
- }
- function statsBySentiment(items) {
- const map = { positive: 0, negative: 0, neutral: 0 };
- for (const it of items) map[it.sentiment || 'neutral']++;
- return map;
- }
- // ==========================================================
- // 9. 导出
- // ==========================================================
- module.exports = {
- loadMerged,
- getEvidence,
- isSubstantive,
- filterByPlatform,
- filterByHypothesis,
- filterByProduct,
- filterByKeyword,
- filterBySentiment,
- filterByContentMatch,
- shuffleWithSeed,
- fmtLikes,
- statsByPlatform,
- statsByHypothesis,
- statsByProduct,
- statsBySentiment,
- HYPOTHESES,
- PLATFORMS,
- };
|