fmode-voc-projects/scripts/tools/lactic-analyze.js

// ==============================================================================
// 江中乳酸菌素片儿童版 · VOC 数据分析模块（参考 liver-analyze.js 架构）
// ==============================================================================
// 数据源：docs/乳酸菌/raw/_merged.json（已在 lactic-collect.js mergeAll 阶段预展平 items[]）
// 所以本模块不需要 flattenMerged（直接读取 items 即可）
// ==============================================================================

const fs = require('fs');
const path = require('path');

const ROOT = path.resolve(__dirname, '..', '..');
const RAW_DIR = path.join(ROOT, 'docs', '乳酸菌', 'raw');
const MERGED_PATH = path.join(RAW_DIR, '_merged.json');

// ------------------------------------------------------------
// 8 条核心假设定义（对应 docs/乳酸菌/2.VOC深度思路.md）
// ------------------------------------------------------------
const HYPOTHESES = {
    H1: { title: '药品益生菌痛点',         desc: '妈咪爱/亿活/金双岐 家长吐槽冷链、活菌失活、处方门槛、依赖焦虑' },
    H2: { title: '食品益生菌信任缺口',     desc: '合生元/万益蓝/拜奥/inne 家长"怕没效果"纠结' },
    H3: { title: '儿童肠道刚需',           desc: '便秘/腹泻/积食/挑食/消化不良 真实家长诉求' },
    H4: { title: '场景触发',               desc: '入园/换季/抗生素后/开学 集中发生的肠道问题' },
    H5: { title: '喂药依从性',             desc: '孩子抗拒吃药/口感差/独立包装/喂药像打仗' },
    H6: { title: '包装 × 颜值决策',         desc: '新生代妈妈专属感需求/货架辨识度/分龄/IP' },
    H7: { title: '营养吸收焦虑',           desc: '瓶瓶罐罐太多/补了不吸收/内源调理替代' },
    H8: { title: '糖分 × 添加剂焦虑',       desc: '新生代家长抵触糖/防腐剂/香精/色素' },
};

// 章节 → 假设映射
const CHAPTER_HYPOTHESIS_MAP = {
    challenge:   ['H1', 'H2', 'H3'],      // 诘问：现状三大痛点
    drug_voc:    ['H1'],                   // 药品益生菌 VOC
    food_voc:    ['H2', 'H3'],            // 食品益生菌 VOC
    kano:        ['H3', 'H5', 'H8'],      // KANO×JTBD
    scene:       ['H4', 'H3'],            // 场景四元素
    competitor:  ['H1', 'H2'],            // 三轴 × 竞品
    opportunity: ['H7', 'H5', 'H6'],      // 新机会
    blueprint:   ['H3', 'H7'],            // 4P 蓝图
};

// ------------------------------------------------------------
// 平台标签
// ------------------------------------------------------------
const PLATFORM_LABELS = {
    xhs:     { name: '小红书',   color: '#FF2442', short: '红' },
    douyin:  { name: '抖音',     color: '#1A1A1A', short: '抖' },
    taobao:  { name: '淘宝',     color: '#FF5000', short: '淘' },
    jd:      { name: '京东',     color: '#E1251B', short: '京' },
    tmall:   { name: '天猫',     color: '#FF0036', short: '猫' },
    unknown: { name: '其他',     color: '#888',    short: '—' },
};

// ------------------------------------------------------------
// 关键词 → 产品 / 竞品 归一化
// ------------------------------------------------------------
const KEYWORD_TO_PRODUCT = {
    '乳酸菌素片':         '江中乳酸菌素片',
    '乳酸菌素片儿童':     '江中乳酸菌素片儿童',
    '江中乳酸菌素片':     '江中乳酸菌素片',
    // 药品益生菌竞品
    '妈咪爱':             '妈咪爱',
    '妈咪爱 宝宝':        '妈咪爱',
    '亿活 儿童':          '亿活',
    '金双岐':             '金双岐',
    '宝乐安':             '宝乐安',
    // 食品益生菌竞品
    '合生元益生菌':       '合生元 儿童益生菌',
    '万益蓝益生菌':       '万益蓝 儿童益生菌',
    '拜奥益生菌':         '拜奥益生菌',
    'inne 噗噗宝':        'inne 噗噗宝',
    '小胖瓶益生菌':       '小胖瓶益生菌',
    '康萃乐益生菌':       '康萃乐益生菌',
    'lifespace 儿童':     'lifespace 儿童益生菌',
    '儿童益生菌推荐':     '儿童益生菌（品类）',
    '儿童益生菌':         '儿童益生菌（品类）',
    // 儿童肠道痛点话题
    '宝宝便秘':           '便秘话题',
    '宝宝积食':           '积食话题',
    '宝宝腹泻':           '腹泻话题',
    '儿童挑食':           '挑食话题',
    '抗生素 益生菌':      '抗生素后肠道话题',
    // 场景
    '宝宝入园生病':       '入园场景',
    '宝宝换季腹泻':       '换季场景',
    // 喂药依从 × 包装
    '喂药难':             '喂药难话题',
    '儿童咀嚼片':         '咀嚼片形态',
    '宝宝专用':           '儿童专属话题',
    // 营养吸收赛道
    '儿童营养软糖':       '儿童营养软糖（竞品）',
    '儿童DHA':            'DHA 补充话题',
    '瓶瓶罐罐':           '瓶瓶罐罐焦虑',
    '儿童补钙':           '补钙话题',
    // 糖分 / 安全焦虑
    '儿童无糖':           '无糖需求',
};

// ------------------------------------------------------------
// 启发式再判（兜底 · lactic-collect.js 已做主要打标）
// ------------------------------------------------------------
const NEG_KEYS = ['智商税', '没用', '没效', '没有效果', '骗', '差评', '退货', '拉肚子', '副作用', '难喝', '难吃', '难喂', '抗拒', '吐出', '失望', '浪费钱', '坑'];
const POS_KEYS = ['有效', '真香', '好用', '推荐', '回购', '复购', '神器', '爱吃', '管用', '对症', '好转', '有改善', '主动要'];
const CONF_KEYS = ['纠结', '矛盾', '不知道.*好', '又.*又', '害怕', '担心', '想停', '是不是'];

function inferSentiment(item) {
    if (item.sentiment && ['positive', 'negative', 'neutral', 'conflicted'].includes(item.sentiment)) return item.sentiment;
    const c = String(item.content || '').toLowerCase();
    for (const k of NEG_KEYS) if (c.includes(k.toLowerCase())) return 'negative';
    for (const k of POS_KEYS) if (c.includes(k.toLowerCase())) return 'positive';
    for (const k of CONF_KEYS) if (new RegExp(k).test(c)) return 'conflicted';
    return 'neutral';
}

// ------------------------------------------------------------
// 数据加载
// ------------------------------------------------------------
const SKELETON = {
    meta: {
        collectedAt: '待采集',
        platforms: {},
        hypotheses: {},
        products: {},
        stage: 'skeleton',
        sourceNote: '暂无采集数据',
    },
    items: [],
    raw: null,
};

function loadMerged() {
    if (fs.existsSync(MERGED_PATH)) {
        try {
            const raw = JSON.parse(fs.readFileSync(MERGED_PATH, 'utf8'));
            const items = (raw.items || []).map((it) => ({
                ...it,
                product: KEYWORD_TO_PRODUCT[it.keyword] || it.product || it.keyword,
                sentiment: inferSentiment(it),
            }));
            return {
                meta: {
                    sourceTier: raw.meta?.sourceTier || 'real-collected',
                    stage: raw.meta?.stage || 'batch-real',
                    collectedAt: raw.meta?.collectedAt || new Date().toISOString().slice(0, 10),
                    product: '江中乳酸菌素片儿童版',
                    stats: raw.meta || {},
                    sourceNote: 'docs/乳酸菌/raw/_merged.json · 真实多平台采集',
                },
                items,
                raw,
            };
        } catch (err) {
            console.warn(`⚠ _merged.json 解析失败：${err.message}`);
        }
    }
    return SKELETON;
}

// ------------------------------------------------------------
// 元数据汇总（给 cover / agenda 用）
// ------------------------------------------------------------
function getMeta(data) {
    const items = (data && data.items) || [];
    const platforms = {};
    const products = {};
    const hypotheses = {};
    const sources = {};
    const sentiments = {};
    const keywords = new Set();
    const tags = new Set();

    for (const it of items) {
        const pf = it.platform || 'unknown';
        platforms[pf] = (platforms[pf] || 0) + 1;
        const prod = it.product || 'unknown';
        products[prod] = (products[prod] || 0) + 1;
        const hs = Array.isArray(it.hypothesis) ? it.hypothesis : (it.hypothesis ? [it.hypothesis] : []);
        for (const h of hs) hypotheses[h] = (hypotheses[h] || 0) + 1;
        const src = it.source || 'unknown';
        sources[src] = (sources[src] || 0) + 1;
        sentiments[it.sentiment || 'unknown'] = (sentiments[it.sentiment || 'unknown'] || 0) + 1;
        if (it.keyword) keywords.add(it.keyword);
        if (Array.isArray(it.tags)) it.tags.forEach((t) => tags.add(t));
    }

    return {
        comments: items.length,
        keywords: keywords.size,
        tagsTotal: tags.size,
        platforms,
        products,
        productsCount: Object.keys(products).length,
        hypotheses,
        sources,
        sentiments,
        stage: data?.meta?.stage || 'unknown',
        sourceTier: data?.meta?.sourceTier || 'unknown',
        collectedAt: data?.meta?.collectedAt || '待采集',
        sourceNote: data?.meta?.sourceNote || '',
    };
}

// ------------------------------------------------------------
// 筛选接口（与 liver-analyze 一致）
// ------------------------------------------------------------
function filterByHypothesis(items, h) {
    return items.filter((it) => {
        const hs = Array.isArray(it.hypothesis) ? it.hypothesis : (it.hypothesis ? [it.hypothesis] : []);
        return hs.includes(h);
    });
}
function filterByProduct(items, product) {
    return items.filter((it) => (it.product || '').includes(product) || (it.keyword || '').includes(product));
}
function filterByKeyword(items, kw) {
    return items.filter((it) => it.keyword === kw);
}
function filterByPlatform(items, platform) {
    return items.filter((it) => it.platform === platform);
}
function filterBySentiment(items, sentiment) {
    return items.filter((it) => it.sentiment === sentiment);
}
function filterByTag(items, tag) {
    return items.filter((it) => Array.isArray(it.tags) && it.tags.some((t) => t.includes(tag)));
}
function filterByContent(items, re) {
    const rx = re instanceof RegExp ? re : new RegExp(String(re), 'i');
    return items.filter((it) => rx.test(String(it.content || '')));
}
function filterByMinLikes(items, min = 1) {
    return items.filter((it) => (it.likes || 0) >= min);
}

// ------------------------------------------------------------
// 排序 / 抽样
// ------------------------------------------------------------
function topByLikes(items, n = 10) {
    return items.slice().sort((a, b) => (b.likes || 0) - (a.likes || 0)).slice(0, n);
}
function sample(items, n = 6, seed = 1) {
    const arr = items.slice();
    const result = [];
    let s = seed;
    while (result.length < n && arr.length) {
        s = (s * 9301 + 49297) % 233280;
        const idx = Math.floor((s / 233280) * arr.length);
        result.push(arr.splice(idx, 1)[0]);
    }
    return result;
}
function groupByTag(items) {
    const map = new Map();
    for (const it of items) {
        if (!Array.isArray(it.tags)) continue;
        for (const t of it.tags) {
            if (!map.has(t)) map.set(t, { tag: t, count: 0, items: [] });
            const g = map.get(t);
            g.count++;
            g.items.push(it);
        }
    }
    return Array.from(map.values()).sort((a, b) => b.count - a.count);
}

// ------------------------------------------------------------
// 主查询接口：getEvidence
// ------------------------------------------------------------
function isSubstantive(content, minChars) {
    const s = String(content || '').trim();
    if (s.length < minChars) return false;
    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));
}

function getEvidence(items, opts = {}) {
    const {
        hypothesis, product, keyword, platform, sentiment, tag,
        minLikes = 0, minChars = 10, contentMatch, requireContentHit = false,
        dedupByContent = true, dedupByNickname = false,
        top = 6, seed = 7, sortBy = 'likes',
    } = opts;
    let filtered = items.slice();
    if (hypothesis) filtered = filterByHypothesis(filtered, hypothesis);
    if (product)    filtered = filterByProduct(filtered, product);
    if (keyword)    filtered = filterByKeyword(filtered, keyword);
    if (platform)   filtered = filterByPlatform(filtered, platform);
    if (sentiment)  filtered = filterBySentiment(filtered, sentiment);
    if (tag)        filtered = filterByTag(filtered, tag);
    if (minLikes)   filtered = filterByMinLikes(filtered, minLikes);

    filtered = filtered.filter((it) => isSubstantive(it.content, minChars));
    if (contentMatch) filtered = filterByContent(filtered, contentMatch);

    if (requireContentHit) {
        filtered = filtered.filter((it) => {
            const c = String(it.content || '').toLowerCase();
            const kw = String(it.keyword || '').toLowerCase();
            if (!kw) return true;
            const fragments = kw.split(/\s+|儿童|宝宝|益生菌/).filter((x) => x.length >= 2);
            if (fragments.length === 0) return c.includes(kw);
            return fragments.some((f) => c.includes(f));
        });
    }

    if (sortBy === 'likes') {
        filtered = filtered.sort((a, b) => (b.likes || 0) - (a.likes || 0));
    }

    if (dedupByContent) {
        const seen = new Set();
        filtered = filtered.filter((it) => {
            const key = dedupByNickname
                ? `${(it.content || '').slice(0, 40)}|${it.nickname || ''}`
                : (it.content || '').slice(0, 40);
            if (seen.has(key)) return false;
            seen.add(key);
            return true;
        });
    }

    const pool = filtered.slice(0, Math.max(top * 2, top + 3));
    return sample(pool, Math.min(top, pool.length), seed);
}

// ------------------------------------------------------------
// 源标识辅助
// ------------------------------------------------------------
function isSeed(item) { return (item?.source || '').includes('pattern') || (item?.source || '').includes('seed'); }
function isReal(item) { return (item?.source || '') === 'real-collected'; }

function getGlobalSourceLabel(meta) {
    if (!meta) return '未加载';
    const tier = meta.sourceTier;
    if (tier === 'real-collected') return '真实采集';
    if (tier === 'pattern-curated') return '公开模式归纳 · 种子样本';
    return '骨架占位';
}

module.exports = {
    HYPOTHESES,
    CHAPTER_HYPOTHESIS_MAP,
    PLATFORM_LABELS,
    loadMerged,
    getMeta,
    // filters
    filterByHypothesis, filterByProduct, filterByKeyword, filterByPlatform,
    filterBySentiment, filterByTag, filterByContent, filterByMinLikes,
    // sorting/sampling
    topByLikes, sample, groupByTag,
    // high-level
    getEvidence,
    isSeed, isReal, getGlobalSourceLabel,
    SKELETON,
};