smart-fin-chart/scripts/normalize_ppmt_financials.mjs

import fs from 'node:fs';
import path from 'node:path';
import { fileURLToPath } from 'node:url';

const ROOT = path.resolve(path.dirname(fileURLToPath(import.meta.url)), '..');
const RAW_DIR = path.join(ROOT, 'data/ppmt/raw');
const OUTPUT_DIR = path.join(ROOT, 'data/ppmt');

const FILES = {
  cashflow: '泡泡玛特_现金流量表_合并报表_20260828_134646.csv',
  income: '泡泡玛特_利润表_合并报表_20260828_134510.csv',
  balance: '泡泡玛特_资产负债表_合并报表_20260828_134503.csv',
  indicators: '泡泡玛特_财务指标_合并报表_20260828_134712.csv',
  revenueComposition: '泡泡玛特_营收构成_20260828_134749.csv',
};

function clean(cell = '') {
  return cell.trim().replace(/^\uFEFF/, '').replace(/^="(.*)"$/, '$1');
}

function toNumber(cell) {
  const value = clean(cell).replace(/^=/, '');
  if (!value) return null;
  if (value.endsWith('%')) {
    const percentage = Number(value.slice(0, -1));
    return Number.isFinite(percentage) ? percentage / 100 : null;
  }
  const number = Number(value);
  return Number.isFinite(number) ? number : null;
}

function readStatement(filename) {
  const lines = fs.readFileSync(path.join(RAW_DIR, filename), 'utf8')
    .trim()
    .split(/\r?\n/)
    .map((line) => line.split(','));
  const dates = lines[1].slice(3).map(clean);
  const values = new Map(lines.slice(2).map((cells) => [clean(cells[0]), cells.slice(3)]));
  return { dates, values };
}

function row(statement, name) {
  const values = statement.values.get(name);
  if (!values) throw new Error(`原始报表缺少行：${name}`);
  return values.map(toNumber);
}

function periodFor(date) {
  const [year, month] = date.split('-');
  if (month === '06') return `${year}H1`;
  if (month === '12') return `${year}FY`;
  throw new Error(`不支持的报表日期：${date}`);
}

function periodForRevenue(date) {
  const [year, quarter] = date.split('-Q');
  if (quarter === '2') return `${year}H1`;
  if (quarter === '4') return `${year}FY`;
  throw new Error(`不支持的营收构成期间：${date}`);
}

function readRevenueComposition(filename) {
  const lines = fs.readFileSync(path.join(RAW_DIR, filename), 'utf8')
    .trim()
    .split(/\r?\n/)
    .map((line) => line.split(','));
  const dates = lines[0].slice(2).map(clean);
  const channels = new Map();
  let section = '';
  let isTotal = false;

  lines.slice(1).forEach((cells) => {
    const first = clean(cells[0]);
    const second = clean(cells[1]);
    if (first.startsWith('按服務類型分-')) {
      section = first.replace('按服務類型分-', '');
      isTotal = false;
      return;
    }
    if (first === '合計') {
      section = '合計';
      isTotal = true;
      return;
    }
    if (second !== '收入' && second !== '收入同比') return;
    const values = cells.slice(2).map(toNumber);
    if (isTotal) channels.set(second === '收入' ? 'totalRevenue' : 'revenueYoy', values);
    else if (section) channels.set(`${section}:${second}`, values);
  });
  return { dates, channels };
}

function writeCSV(filename, headers, rows) {
  const text = [headers.join(','), ...rows.map((values) => values.join(','))].join('\n').concat('\n');
  fs.writeFileSync(path.join(OUTPUT_DIR, filename), text);
}

const cashflow = readStatement(FILES.cashflow);
const income = readStatement(FILES.income);
const balance = readStatement(FILES.balance);
const indicators = readStatement(FILES.indicators);
const revenueComposition = readRevenueComposition(FILES.revenueComposition);

const operatingCashFlow = row(cashflow, '经营活动产生的现金流量净额');
const netProfit = row(income, '三、净利润');

const cashConversionRows = cashflow.dates.map((date, index) => ({
  date,
  operatingCashFlow: operatingCashFlow[index],
  netProfit: netProfit[index],
})).filter(({ date, operatingCashFlow: cfo, netProfit: profit }) => (
  date >= '2020-01-01' && date.endsWith('-12-31') && cfo !== null && profit !== null
)).map(({ date, operatingCashFlow: cfo, netProfit: profit }) => [
  periodFor(date),
  (cfo / 100000000).toFixed(2),
  (profit / 100000000).toFixed(2),
  (cfo / profit * 100).toFixed(1),
]).reverse();

writeCSV('ppmt_cash_conversion.csv', [
  'period', 'operating_cash_flow', 'net_profit', 'cash_conversion',
], cashConversionRows);

const cash = row(balance, '货币资金');
const liabilities = row(balance, '二、负债合计');
const currentAssets = row(balance, '流动资产合计');
const currentLiabilities = row(balance, '流动负债合计');

const liquidityRows = balance.dates.map((date, index) => ({
  date,
  cash: cash[index],
  liabilities: liabilities[index],
  currentAssets: currentAssets[index],
  currentLiabilities: currentLiabilities[index],
})).filter((item) => (
  item.date >= '2020-01-01'
  && item.cash !== null
  && item.liabilities !== null
  && item.currentAssets !== null
  && item.currentLiabilities !== null
)).map((item) => [
  periodFor(item.date),
  (item.cash / 100000000).toFixed(2),
  (item.liabilities / 100000000).toFixed(2),
  ((item.currentAssets - item.currentLiabilities) / 100000000).toFixed(2),
]).reverse();

writeCSV('ppmt_liquidity_working_capital.csv', [
  'period', 'cash', 'total_liabilities', 'net_working_capital',
], liquidityRows);

const investingCashFlow = row(cashflow, '投资活动产生的现金流量净额');
const financingCashFlow = row(cashflow, '筹资活动产生的现金流量净额');
const netCashIncrease = row(cashflow, '现金及现金等价物的净增加额');
const cashflowCompositionRows = cashflow.dates.map((date, index) => ({
  date,
  operating: operatingCashFlow[index],
  investing: investingCashFlow[index],
  financing: financingCashFlow[index],
  netIncrease: netCashIncrease[index],
})).filter((item) => (
  item.date >= '2020-01-01' && item.date.endsWith('-12-31')
  && Object.values(item).every((value) => value !== null)
)).map((item) => [
  periodFor(item.date),
  (item.operating / 100000000).toFixed(2),
  (item.investing / 100000000).toFixed(2),
  (item.financing / 100000000).toFixed(2),
  (item.netIncrease / 100000000).toFixed(2),
]).reverse();

writeCSV('ppmt_cashflow_composition.csv', [
  'period', 'operating_cash_flow', 'investing_cash_flow', 'financing_cash_flow', 'net_cash_increase',
], cashflowCompositionRows);

const inventoryDays = row(indicators, '存货周转天数');
const receivableDays = row(indicators, '应收票据和应收账款周转天数');
const payableDays = row(indicators, '应付票据和应付账款周转天数');
const workingCapitalDaysRows = indicators.dates.map((date, index) => ({
  date,
  inventory: inventoryDays[index],
  receivable: receivableDays[index],
  payable: payableDays[index],
})).filter((item) => (
  item.date >= '2020-01-01'
  && item.inventory !== null && item.receivable !== null && item.payable !== null
)).map((item) => [
  periodFor(item.date),
  item.inventory.toFixed(1),
  item.receivable.toFixed(1),
  item.payable.toFixed(1),
  (item.inventory + item.receivable - item.payable).toFixed(1),
]).reverse();

writeCSV('ppmt_working_capital_days.csv', [
  'period', 'inventory_days', 'receivable_days', 'payable_days', 'cash_conversion_cycle',
], workingCapitalDaysRows);

const revenue = row(income, '一、营业总收入');
const salesExpense = row(income, '减：销售费用');
const adminExpense = row(income, '管理费用');
const financialExpense = row(income, '财务费用');
const fourExpenseRate = row(income, '四项费用率');
const expenseRows = income.dates.map((date, index) => ({
  date,
  revenue: revenue[index],
  sales: salesExpense[index],
  admin: adminExpense[index],
  financial: financialExpense[index],
  rate: fourExpenseRate[index],
})).filter((item) => (
  item.date >= '2020-01-01' && item.date.endsWith('-12-31')
  && Object.values(item).every((value) => value !== null)
)).map((item) => [
  periodFor(item.date),
  (item.sales / 100000000).toFixed(2),
  (item.admin / 100000000).toFixed(2),
  (item.financial / 100000000).toFixed(2),
  (item.rate * 100).toFixed(1),
]).reverse();

writeCSV('ppmt_expense_structure.csv', [
  'period', 'sales_expense', 'admin_expense', 'financial_expense', 'four_expense_rate',
], expenseRows);

const channelKeys = [
  ['零售店銷售收益', 'retail'],
  ['線上銷售收益', 'online'],
  ['機器人商店銷售收益', 'robot_store'],
  ['批發收益', 'wholesale'],
];
const channelRows = revenueComposition.dates.map((date, index) => {
  const values = Object.fromEntries(channelKeys.map(([name, key]) => [
    key,
    revenueComposition.channels.get(`${name}:收入`)?.[index] ?? null,
  ]));
  return {
    date,
    ...values,
    totalRevenue: revenueComposition.channels.get('totalRevenue')?.[index] ?? null,
    revenueYoy: revenueComposition.channels.get('revenueYoy')?.[index] ?? null,
  };
}).filter((item) => (
  item.date >= '2020-Q2'
  && Object.values(item).every((value) => value !== null)
)).map((item) => [
  periodForRevenue(item.date),
  (item.retail / 100000000).toFixed(2),
  (item.online / 100000000).toFixed(2),
  (item.robot_store / 100000000).toFixed(2),
  (item.wholesale / 100000000).toFixed(2),
  (item.totalRevenue / 100000000).toFixed(2),
  (item.revenueYoy * 100).toFixed(1),
]).reverse();

writeCSV('ppmt_channel_revenue.csv', [
  'period', 'retail', 'online', 'robot_store', 'wholesale', 'total_revenue', 'revenue_yoy',
], channelRows);

console.log(`已生成现金转换 ${cashConversionRows.length}、流动性 ${liquidityRows.length}、现金流 ${cashflowCompositionRows.length}、营运资本 ${workingCapitalDaysRows.length}、费用 ${expenseRows.length}、渠道 ${channelRows.length} 行数据。`);