feat: 增加不同地区各渠道收入占比图

2026-09-08 16:53:44 +08:00
parent 44c6437b64
commit bab02e1efd
25 changed files with 888 additions and 29 deletions
@@ -0,0 +1,95 @@
+import fs from 'node:fs/promises';
+import path from 'node:path';
+
+const sourcePath = path.resolve('data/brk/rawData/brk_holdings_1988_1997_export.csv');
+const outputPath = path.resolve('data/brk/brk_holdings.csv');
+
+const COMPANY_NAMES = [
+  ['american express|美国运通', '美国运通（American Express）'],
+  ['coca-cola|可口可乐', '可口可乐（Coca-Cola）'],
+  ['disney|迪士尼', '华特迪士尼（Walt Disney）'],
+  ['freddie mac|房地美|联邦住', '房地美（Freddie Mac）'],
+  ['gillette|吉列', '吉列（Gillette）'],
+  ['travelers|旅行家', '旅行家集团（Travelers Group）'],
+  ['washington post|华盛顿邮报', '华盛顿邮报（Washington Post）'],
+  ['wells fargo|富国银行', '富国银行（Wells Fargo）'],
+  ['capital cities|大都会', '大都会/ABC（Capital Cities/ABC）'],
+  ['geico|盖可', '盖可保险（GEICO）'],
+  ['gannett|甘尼特', '甘尼特（Gannett）'],
+  ['pnc', 'PNC 银行（PNC Bank）'],
+  ['general dynamics|通用动力', '通用动力（General Dynamics）'],
+  ['guinness|健力士', '健力士（Guinness）'],
+  ["mcdonald|麦当劳", '麦当劳（McDonald\'s）'],
+].map(([pattern, name]) => ({ pattern: new RegExp(pattern, 'i'), name }));
+
+function parseCSV(text) {
+  const rows = [];
+  let row = [];
+  let value = '';
+  let quoted = false;
+
+  for (let i = 0; i < text.length; i += 1) {
+    const char = text[i];
+    if (char === '"') {
+      if (quoted && text[i + 1] === '"') {
+        value += '"';
+        i += 1;
+      } else {
+        quoted = !quoted;
+      }
+    } else if (char === ',' && !quoted) {
+      row.push(value.trim());
+      value = '';
+    } else if ((char === '\n' || char === '\r') && !quoted) {
+      if (char === '\r' && text[i + 1] === '\n') i += 1;
+      row.push(value.trim());
+      rows.push(row);
+      row = [];
+      value = '';
+    } else {
+      value += char;
+    }
+  }
+  if (value || row.length) rows.push([...row, value.trim()]);
+  return rows;
+}
+
+function canonicalName(value) {
+  return COMPANY_NAMES.find(({ pattern }) => pattern.test(value))?.name || value.trim();
+}
+
+function parseShares(value) {
+  const digits = value.replace(/[^\d]/g, '');
+  return digits ? Number(digits) : NaN;
+}
+
+const rows = parseCSV(await fs.readFile(sourcePath, 'utf8'));
+const holdings = [];
+
+for (let index = 0; index < rows.length; index += 1) {
+  const dateMatch = rows[index][0]?.match(/^(\d{4})\/12\/31$/);
+  if (!dateMatch) continue;
+
+  const year = Number(dateMatch[1]);
+  const header = rows[index + 1] || [];
+  const sharesFirst = /股(份)?数量|股数|持股数量/.test(header[0] || '');
+  const companyIndex = sharesFirst ? 1 : 0;
+  const sharesIndex = sharesFirst ? 0 : 1;
+
+  for (let rowIndex = index + 2; rowIndex < rows.length; rowIndex += 1) {
+    const record = rows[rowIndex];
+    if (!record.some(Boolean)) break;
+    if (/^\d{4}\/12\/31$/.test(record[0] || '')) break;
+    const company = record[companyIndex];
+    const shares = parseShares(record[sharesIndex] || '');
+    if (!company || !Number.isFinite(shares)) continue;
+    holdings.push({ year, company: canonicalName(company), shares });
+  }
+}
+
+holdings.sort((a, b) => a.year - b.year || a.company.localeCompare(b.company, 'zh-CN'));
+const lines = ['year,company,shares', ...holdings.map((row) => `${row.year},${row.company},${row.shares}`)];
+await fs.mkdir(path.dirname(outputPath), { recursive: true });
+await fs.writeFile(outputPath, `${lines.join('\n')}\n`, 'utf8');
+
+console.log(`Normalized ${holdings.length} holdings across ${new Set(holdings.map((row) => row.year)).size} years.`);
@@ -714,10 +714,24 @@ const CHANNEL_NAME_TO_KEY = {
  线上渠道: 'online',
  批发及其他: 'wholesale',
 };
+const ONLINE_SUB_NAME_TO_KEY = {
+  泡泡玛特抽盒机: 'roboshop',
+  天猫旗舰店: 'tmall',
+  抖音平台: 'douyin',
+  自研APP及官方网站: 'own_app',
+  Shopee: 'shopee',
+  TikTok: 'tiktok',
+  亚马逊: 'amazon',
+  其他线上渠道: 'online_other',
+};
 const channelRevenueByPeriod = new Map();
 readCsvRows(RAWDATA_DIR, '地区渠道收入.csv').forEach(({ region, channel, period, revenue }) => {
  const key = CHANNEL_NAME_TO_KEY[channel];
-  if (!key) throw new Error(`地区渠道收入底表存在未知渠道：${channel}`);
+  if (!key) {
+    // 线上细分行只进分地区图；未知渠道仍视为底表错误。
+    if (!ONLINE_SUB_NAME_TO_KEY[channel]) throw new Error(`地区渠道收入底表存在未知渠道：${channel}`);
+    return;
+  }
  if (!period || !Number.isFinite(revenue)) throw new Error(`地区渠道收入底表行不完整：${region} ${channel} ${period} ${revenue}`);
  if (!channelRevenueByPeriod.has(period)) {
    channelRevenueByPeriod.set(period, { retail: 0, online: 0, robot_store: 0, wholesale: 0 });
@@ -745,3 +759,66 @@ writeRuntimeViews('ppmt_channel_share', ['period', 'retail', 'online', 'robot_st
 });

 console.log('已生成渠道收入占比的全部、年度、半年 CSV。');
+
+// 分地区渠道收入占比：线上渠道按披露细分拆开，各地区细分不同。
+// 底表中「线上渠道」合计行仅作核对值；细分空缺期已按 0 录入并核对
+// （美洲区 2026H1 起 TikTok 未再产生收入、亚马逊 2026 年起开售，与线上合计勾稽一致）。
+const REGION_KEY_BY_NAME = { 中国: 'china', 亚太: 'apac', 美洲: 'americas', 欧洲及其他: 'europe' };
+const REGION_SERIES = {
+  china: ['retail', 'robot_store', 'roboshop', 'tmall', 'douyin', 'online_other', 'wholesale'],
+  apac: ['retail', 'robot_store', 'own_app', 'shopee', 'online_other', 'wholesale'],
+  americas: ['retail', 'robot_store', 'own_app', 'tiktok', 'amazon', 'online_other', 'wholesale'],
+  europe: ['retail', 'robot_store', 'own_app', 'online_other', 'wholesale'],
+};
+const TOP_AND_SUB_NAME_TO_KEY = { ...CHANNEL_NAME_TO_KEY, ...ONLINE_SUB_NAME_TO_KEY };
+const regionAmounts = new Map();
+const regionOnlineDisclosed = new Map();
+readCsvRows(RAWDATA_DIR, '地区渠道收入.csv').forEach(({ region, channel, period, revenue }) => {
+  const regionKey = REGION_KEY_BY_NAME[region];
+  if (!regionKey) throw new Error(`地区渠道收入底表存在未知地区：${region}`);
+  if (channel === '线上渠道') {
+    regionOnlineDisclosed.set(`${regionKey}:${period}`, revenue);
+    return;
+  }
+  const key = TOP_AND_SUB_NAME_TO_KEY[channel];
+  if (!key) throw new Error(`地区渠道收入底表存在未知渠道：${channel}`);
+  if (!period || !Number.isFinite(revenue)) throw new Error(`地区渠道收入底表行不完整：${region} ${channel} ${period} ${revenue}`);
+  const bucketKey = `${regionKey}:${period}`;
+  if (!regionAmounts.has(bucketKey)) regionAmounts.set(bucketKey, {});
+  const bucket = regionAmounts.get(bucketKey);
+  bucket[key] = (bucket[key] ?? 0) + revenue;
+});
+regionOnlineDisclosed.forEach((disclosed, bucketKey) => {
+  const bucket = regionAmounts.get(bucketKey);
+  if (!bucket) throw new Error(`地区渠道收入底表缺少线上合计对应的期间：${bucketKey}`);
+  const subSum = Object.values(ONLINE_SUB_NAME_TO_KEY).reduce((sum, key) => sum + (bucket[key] ?? 0), 0);
+  if (Math.abs(subSum - disclosed) > 0.05) {
+    throw new Error(`${bucketKey} 线上细分合计 ${subSum.toFixed(2)} ≠ 披露线上渠道 ${disclosed.toFixed(2)}`);
+  }
+});
+
+Object.entries(REGION_SERIES).forEach(([regionKey, seriesKeys]) => {
+  const amountViews = buildViews((year, type) => {
+    const bucket = regionAmounts.get(`${regionKey}:${type === 'H1' ? `${year}H1` : `${year}FY`}`);
+    if (!bucket) return null;
+    const values = Object.fromEntries(seriesKeys.map((key) => [key, bucket[key] ?? 0]));
+    return isComplete(values) ? values : null;
+  });
+  const shareViews = Object.fromEntries(Object.entries(amountViews).map(([view, rows]) => [
+    view,
+    rows.map(({ period, ...amounts }) => {
+      const total = Object.values(amounts).reduce((sum, value) => sum + value, 0);
+      return {
+        period,
+        ...Object.fromEntries(Object.entries(amounts).map(([key, value]) => [key, value / total * 100])),
+        total_revenue: total,
+      };
+    }),
+  ]));
+  writeRuntimeViews(`ppmt_channel_share_${regionKey}`, ['period', ...seriesKeys, 'total_revenue'], shareViews, {
+    ...Object.fromEntries(seriesKeys.map((key) => [key, 1])),
+    total_revenue: 2,
+  });
+});
+
+console.log('已生成四地区渠道收入占比（线上细分）的全部、年度、半年 CSV。');