sankey/miniapp/utils/sankey.js

/**
 * 统一清洗文本，避免空格导致节点重复。
 */
function normalizeText(value) {
  return String(value || '').trim();
}

/**
 * 尝试懒加载 xlsx 解析库。
 * 说明：
 * - 小程序若未完成 npm 构建，此处会拿不到模块
 * - 使用缓存避免每次解析都重复 require
 */
let cachedXlsxModule;
function getXlsxModule() {
  if (cachedXlsxModule !== undefined) {
    return cachedXlsxModule;
  }
  try {
    cachedXlsxModule = require('xlsx');
  } catch (error) {
    // 兜底：兼容开发者工具未正确挂载 npm 解析时，直接从 miniprogram_npm 目录加载。
    try {
      cachedXlsxModule = require('../miniprogram_npm/xlsx');
    } catch (fallbackError) {
      cachedXlsxModule = null;
    }
  }
  return cachedXlsxModule;
}

/**
 * 将二维数组统一整理为 headers + rows 结构。
 * 约定第一行为表头，后续行为数据行。
 */
function toRawTable(rows) {
  if (!Array.isArray(rows) || rows.length === 0) {
    return { headers: [], rows: [] };
  }

  const firstRow = Array.isArray(rows[0]) ? rows[0] : [];
  const maxColumns = rows.reduce((max, row) => {
    const length = Array.isArray(row) ? row.length : 0;
    return Math.max(max, length);
  }, firstRow.length);

  const headers = Array.from({ length: maxColumns }, (_, index) => {
    const header = normalizeText(firstRow[index] || '');
    return header || `列${index + 1}`;
  });

  const dataRows = rows.slice(1).map((row) => {
    const safeRow = Array.isArray(row) ? row : [];
    return Array.from({ length: maxColumns }, (_, index) => normalizeText(safeRow[index] || ''));
  });

  return { headers, rows: dataRows };
}

/**
 * 解析数字，支持千分位。
 */
function parseNumericValue(text) {
  const normalized = normalizeText(text).replace(/,/g, '');
  if (!normalized) {
    return null;
  }
  const parsed = Number(normalized);
  if (Number.isNaN(parsed)) {
    return null;
  }
  return parsed;
}

/**
 * 将单元格值格式化为日志可读文本。
 */
function formatCellValueForWarning(value) {
  const text = String(value || '');
  return text.length > 0 ? text : '(空)';
}

/**
 * 组装“列位置 + 列名 + 原始值”的调试文本。
 */
function buildColumnDebugText(row, headers, columns) {
  if (!Array.isArray(columns) || columns.length === 0) {
    return '未选择列';
  }
  return columns
    .map((columnIndex) => {
      const headerName = headers[columnIndex] || `列${columnIndex + 1}`;
      const rawValue = row[columnIndex] || '';
      return `第 ${columnIndex + 1} 列（${headerName}）="${formatCellValueForWarning(rawValue)}"`;
    })
    .join('，');
}

/**
 * 简单 CSV 解析（支持双引号与双引号转义）。
 */
function parseCsvText(csvText) {
  const text = String(csvText || '').replace(/^\uFEFF/, '');
  const rows = [];
  let row = [];
  let cell = '';
  let inQuotes = false;

  for (let i = 0; i < text.length; i += 1) {
    const ch = text[i];
    const next = text[i + 1];
    if (ch === '"') {
      if (inQuotes && next === '"') {
        cell += '"';
        i += 1;
      } else {
        inQuotes = !inQuotes;
      }
      continue;
    }

    if (!inQuotes && ch === ',') {
      row.push(cell);
      cell = '';
      continue;
    }

    if (!inQuotes && (ch === '\n' || ch === '\r')) {
      if (ch === '\r' && next === '\n') {
        i += 1;
      }
      row.push(cell);
      cell = '';
      rows.push(row);
      row = [];
      continue;
    }

    cell += ch;
  }

  row.push(cell);
  rows.push(row);

  const normalizedRows = rows
    .map((items) => items.map((item) => normalizeText(item)))
    .filter((items) => items.some((item) => item.length > 0));

  return toRawTable(normalizedRows);
}

/**
 * 解析 xls/xlsx 二进制内容。
 */
function parseXlsxBuffer(buffer) {
  const xlsx = getXlsxModule();
  if (!xlsx) {
    throw new Error('当前环境未启用 xlsx 解析，请先在开发者工具执行“构建 npm”');
  }

  const workbook = xlsx.read(buffer, { type: 'array' });
  const firstSheetName = workbook.SheetNames[0];
  if (!firstSheetName) {
    throw new Error('Excel 文件中没有工作表');
  }

  const sheet = workbook.Sheets[firstSheetName];
  const rows = xlsx.utils.sheet_to_json(sheet, {
    header: 1,
    raw: false,
    defval: ''
  });
  return toRawTable(rows);
}

/**
 * 按文件名后缀自动分流解析器。
 */
function parseTableByFileName(fileName, payload) {
  const lowerName = String(fileName || '').toLowerCase();
  if (lowerName.endsWith('.csv')) {
    return parseCsvText(String(payload || ''));
  }
  if (lowerName.endsWith('.xlsx') || lowerName.endsWith('.xls')) {
    return parseXlsxBuffer(payload);
  }
  throw new Error('仅支持 .csv / .xlsx / .xls 文件');
}

/**
 * 构建 source 名称。
 */
function buildSourceName(row, config) {
  if (!Array.isArray(config.sourceDescriptionColumns) || config.sourceDescriptionColumns.length === 0) {
    return normalizeText(row[config.sourceDataColumn] || '');
  }
  const parts = config.sourceDescriptionColumns
    .map((column) => normalizeText(row[column] || ''))
    .filter((item) => item.length > 0);
  return parts.join(config.delimiter || '-');
}

/**
 * 构建 target 名称，支持向下补全。
 */
function buildTargetName(row, config, lastNonEmptyTargetValueByColumn) {
  const parts = (config.targetDescriptionColumns || [])
    .map((column) => {
      const raw = normalizeText(row[column] || '');
      if (raw.length > 0) {
        lastNonEmptyTargetValueByColumn[column] = raw;
        return raw;
      }
      return lastNonEmptyTargetValueByColumn[column] || '';
    })
    .filter((item) => item.length > 0);
  return parts.join(config.delimiter || '-');
}

/**
 * 与 Web 端保持一致的聚合规则。
 */
function buildSankeyData(table, config) {
  if (config.sourceDataColumn === null || config.sourceDataColumn === undefined) {
    throw new Error('必须选择源数据列');
  }
  if (!Array.isArray(config.targetDescriptionColumns) || config.targetDescriptionColumns.length === 0) {
    throw new Error('必须至少选择一个目标描述列');
  }

  const sourceDataColumnIndex = config.sourceDataColumn;
  const sourceDataColumnName = table.headers[sourceDataColumnIndex] || `列${sourceDataColumnIndex + 1}`;
  const linkValueMap = {};
  const warnings = [];
  let droppedRows = 0;
  const lastNonEmptyTargetValueByColumn = {};

  (table.rows || []).forEach((row, rowIndex) => {
    const excelRow = rowIndex + 2;
    const sourceCellRaw = row[sourceDataColumnIndex] || '';
    const sourceValue = parseNumericValue(sourceCellRaw);

    if (sourceValue === null) {
      warnings.push(
        `第 ${excelRow} 行, 第 ${sourceDataColumnIndex + 1} 列（${sourceDataColumnName}）: 源数据不是有效数字，原始值="${formatCellValueForWarning(sourceCellRaw)}"，已跳过`
      );
      droppedRows += 1;
      return;
    }

    const sourceName = buildSourceName(row, config);
    if (!sourceName) {
      warnings.push(
        `第 ${excelRow} 行: 源描述为空，字段=${buildColumnDebugText(
          row,
          table.headers || [],
          config.sourceDescriptionColumns || []
        )}，已跳过`
      );
      droppedRows += 1;
      return;
    }

    const targetName = buildTargetName(row, config, lastNonEmptyTargetValueByColumn);
    if (!targetName) {
      warnings.push(
        `第 ${excelRow} 行: 目标描述为空，字段=${buildColumnDebugText(
          row,
          table.headers || [],
          config.targetDescriptionColumns || []
        )}，且无可继承的上方值，已跳过`
      );
      droppedRows += 1;
      return;
    }

    const key = `${sourceName}@@${targetName}`;
    linkValueMap[key] = (linkValueMap[key] || 0) + sourceValue;
  });

  const links = [];
  const sourceSet = {};
  const targetSet = {};
  Object.keys(linkValueMap).forEach((key) => {
    const pair = key.split('@@');
    const source = pair[0];
    const target = pair[1];
    if (!source || !target) {
      return;
    }
    sourceSet[source] = true;
    targetSet[target] = true;
    links.push({
      source,
      target,
      value: linkValueMap[key]
    });
  });

  const nodes = [];
  Object.keys(sourceSet).forEach((name) => {
    nodes.push({ name, kind: 'source' });
  });
  Object.keys(targetSet).forEach((name) => {
    if (!sourceSet[name]) {
      nodes.push({ name, kind: 'target' });
    }
  });

  return {
    nodes,
    links,
    meta: {
      droppedRows,
      warnings
    }
  };
}

module.exports = {
  parseCsvText,
  parseXlsxBuffer,
  parseTableByFileName,
  buildSankeyData
};