big-rag-modified

import ExcelJS from "exceljs";
import * as XLSX from "xlsx";

/**
 * Parse XLSX files using ExcelJS — preserves sheet names and handles rich text
 * and formula results natively.
 */
export async function parseXLSX(filePath: string): Promise<string> {
  try {
    const workbook = new ExcelJS.Workbook();
    await workbook.xlsx.readFile(filePath);

    const sheetParts: string[] = [];

    workbook.eachSheet((sheet) => {
      const rows: string[] = [];

      sheet.eachRow({ includeEmpty: false }, (row) => {
        const cells: string[] = [];
        row.eachCell({ includeEmpty: true }, (cell) => {
          const val = cell.value;
          if (val === null || val === undefined) {
            cells.push("");
          } else if (typeof val === "object" && "richText" in val) {
            cells.push(val.richText.map((r) => r.text).join(""));
          } else if (typeof val === "object" && "result" in val) {
            cells.push(String((val as ExcelJS.CellFormulaValue).result ?? ""));
          } else if (typeof val === "object" && val instanceof Date) {
            cells.push(val.toISOString());
          } else {
            cells.push(String(val));
          }
        });
        rows.push(cells.join("\t"));
      });

      if (rows.length > 0) {
        sheetParts.push(`Sheet: ${sheet.name}\n${rows.join("\n")}`);
      }
    });

    return sheetParts.join("\n\n").trim();
  } catch (error) {
    console.error(`Error parsing XLSX file ${filePath}:`, error);
    return "";
  }
}

/**
 * Parse legacy Excel binary files (.xls) using SheetJS.
 *
 * Note: SheetJS (xlsx package) has known CVEs for prototype-pollution and
 * ReDoS when processing untrusted input. Formula parsing is disabled here
 * ({cellFormula: false}) to remove the ReDoS vector. This parser is
 * intentionally used only for LOCAL files the user has placed in their
 * own document directory — it should never be exposed to untrusted input.
 */
export function parseXLS(filePath: string): string {
  try {
    const workbook = XLSX.readFile(filePath, {
      cellFormula: false, // disables the ReDoS vector
      cellHTML: false,
    });

    const sheetParts: string[] = [];

    for (const sheetName of workbook.SheetNames) {
      const sheet = workbook.Sheets[sheetName];
      const rows = XLSX.utils.sheet_to_json<string[]>(sheet, {
        header: 1,
        defval: "",
        raw: false,
      });

      const nonEmpty = (rows as string[][]).filter((r) =>
        r.some((c) => c !== null && c !== undefined && String(c).trim() !== ""),
      );

      if (nonEmpty.length > 0) {
        const body = nonEmpty.map((r) => r.map(String).join("\t")).join("\n");
        sheetParts.push(`Sheet: ${sheetName}\n${body}`);
      }
    }

    return sheetParts.join("\n\n").trim();
  } catch (error) {
    console.error(`Error parsing XLS file ${filePath}:`, error);
    return "";
  }
}