@bendyline/squisq-formats 2.0.1 → 2.2.0

package/dist/chunk-OGS5VCGJ.js

@@ -0,0 +1,446 @@
+import {
+  extractPlainText
+} from "./chunk-SBUW7NHR.js";
+import {
+  createPackage
+} from "./chunk-4V3KCHAP.js";
+import {
+  escapeXml,
+  xmlDeclaration
+} from "./chunk-JU2RHXUB.js";
+import {
+  CONTENT_TYPE_XLSX_STYLES,
+  CONTENT_TYPE_XLSX_WORKBOOK,
+  CONTENT_TYPE_XLSX_WORKSHEET,
+  NS_R,
+  NS_SML,
+  REL_OFFICE_DOCUMENT,
+  REL_STYLES,
+  REL_WORKSHEET,
+  baseDirOf,
+  getPartRelationships,
+  getPartXml,
+  openPackage,
+  requireMainPartPath
+} from "./chunk-X2DEAXNK.js";
+
+// src/xlsx/index.ts
+import { markdownToDoc } from "@bendyline/squisq/doc";
+
+// src/xlsx/import.ts
+var XLSX_MAIN_PART = "xl/workbook.xml";
+function attrNS(el, ns, local, fallback) {
+  return el.getAttributeNS(ns, local) ?? el.getAttribute(fallback);
+}
+function resolveTarget(baseDir, target) {
+  if (target.startsWith("/")) return target.replace(/^\//, "");
+  const stack = baseDir ? baseDir.split("/") : [];
+  for (const seg of target.split("/")) {
+    if (seg === "..") stack.pop();
+    else if (seg !== ".") stack.push(seg);
+  }
+  return stack.join("/");
+}
+async function readWorkbook(pkg, mainPart) {
+  const wb = await getPartXml(pkg, mainPart);
+  if (!wb) {
+    throw new Error(`Invalid XLSX package: workbook part "${mainPart}" could not be parsed.`);
+  }
+  const rels = await getPartRelationships(pkg, mainPart);
+  const relById = new Map(rels.map((r) => [r.id, r.target]));
+  const out = [];
+  const sheetEls = wb.getElementsByTagNameNS(NS_SML, "sheet");
+  for (let i = 0; i < sheetEls.length; i++) {
+    const el = sheetEls[i];
+    const name = el.getAttribute("name") ?? `Sheet${i + 1}`;
+    const rid = attrNS(el, NS_R, "id", "r:id");
+    const target = rid ? relById.get(rid) : void 0;
+    if (target) out.push({ name, path: resolveTarget(baseDirOf(mainPart), target) });
+  }
+  const workbookPr = wb.getElementsByTagNameNS(NS_SML, "workbookPr")[0];
+  const date1904Value = workbookPr?.getAttribute("date1904");
+  return {
+    sheets: out,
+    date1904: date1904Value === "1" || date1904Value === "true"
+  };
+}
+function isPhonetic(el, root) {
+  for (let node = el.parentNode; node && node !== root; node = node.parentNode) {
+    if (node.nodeType !== 1) continue;
+    const parent = node;
+    if (parent.localName === "rPh" && parent.namespaceURI === NS_SML) return true;
+  }
+  return false;
+}
+function stringItemText(root) {
+  const tEls = root.getElementsByTagNameNS(NS_SML, "t");
+  let out = "";
+  for (let i = 0; i < tEls.length; i++) {
+    const t = tEls[i];
+    if (isPhonetic(t, root)) continue;
+    out += t.textContent ?? "";
+  }
+  return out;
+}
+async function readSharedStrings(pkg) {
+  const doc = await getPartXml(pkg, "xl/sharedStrings.xml");
+  if (!doc) return [];
+  const siEls = doc.getElementsByTagNameNS(NS_SML, "si");
+  const out = [];
+  for (let i = 0; i < siEls.length; i++) out.push(stringItemText(siEls[i]));
+  return out;
+}
+var BUILTIN_NUMBER_FORMATS = Object.freeze({
+  9: "0%",
+  10: "0.00%",
+  14: "mm-dd-yy",
+  15: "d-mmm-yy",
+  16: "d-mmm",
+  17: "mmm-yy",
+  18: "h:mm AM/PM",
+  19: "h:mm:ss AM/PM",
+  20: "h:mm",
+  21: "h:mm:ss",
+  22: "m/d/yy h:mm",
+  45: "mm:ss",
+  46: "[h]:mm:ss",
+  47: "mmss.0"
+});
+async function readCellStyles(pkg) {
+  const doc = await getPartXml(pkg, "xl/styles.xml");
+  if (!doc) return [];
+  const custom = /* @__PURE__ */ new Map();
+  const numFmtEls = doc.getElementsByTagNameNS(NS_SML, "numFmt");
+  for (let i = 0; i < numFmtEls.length; i++) {
+    const el = numFmtEls[i];
+    const id = Number.parseInt(el.getAttribute("numFmtId") ?? "", 10);
+    const code = el.getAttribute("formatCode");
+    if (Number.isFinite(id) && code) custom.set(id, code);
+  }
+  const cellXfs = doc.getElementsByTagNameNS(NS_SML, "cellXfs")[0];
+  if (!cellXfs) return [];
+  const xfEls = cellXfs.getElementsByTagNameNS(NS_SML, "xf");
+  const out = [];
+  for (let i = 0; i < xfEls.length; i++) {
+    const numFmtId = Number.parseInt(xfEls[i].getAttribute("numFmtId") ?? "0", 10);
+    out.push({
+      numFmtId,
+      formatCode: custom.get(numFmtId) ?? BUILTIN_NUMBER_FORMATS[numFmtId] ?? "General"
+    });
+  }
+  return out;
+}
+function colIndex(ref) {
+  const m = /^([A-Za-z]+)/.exec(ref);
+  if (!m) return 0;
+  let n = 0;
+  for (const ch of m[1].toUpperCase()) n = n * 26 + (ch.charCodeAt(0) - 64);
+  return n - 1;
+}
+function normalizeFormatCode(formatCode) {
+  return formatCode.split(";", 1)[0].replace(/"(?:[^"]|"")*"/g, "").replace(/\\./g, "").replace(/[_*]./g, "").replace(/\[(?!h+\]|m+\]|s+\])[^\]]*\]/gi, "").toLowerCase();
+}
+function numberFormatKind(formatCode) {
+  const normalized = normalizeFormatCode(formatCode);
+  if (normalized === "general") return "general";
+  if (normalized.includes("%")) return "percent";
+  if (/^0+$/.test(normalized)) return "zero-pad";
+  const hasDate = /[yd]/.test(normalized);
+  const hasTime = /[hs]|\[[hms]+\]/.test(normalized);
+  if (hasDate && hasTime) return "datetime";
+  if (hasDate) return "date";
+  if (hasTime) return "time";
+  return "general";
+}
+function twoDigits(value) {
+  return String(value).padStart(2, "0");
+}
+function excelDateText(serial, date1904) {
+  const wholeDays = Math.floor(serial);
+  if (!date1904 && wholeDays === 60) return "1900-02-29";
+  const epoch = date1904 ? Date.UTC(1904, 0, 1) : Date.UTC(1899, 11, 31);
+  const adjustedDays = !date1904 && wholeDays > 60 ? wholeDays - 1 : wholeDays;
+  const date = new Date(epoch + adjustedDays * 864e5);
+  if (!Number.isFinite(date.getTime())) return null;
+  return `${date.getUTCFullYear()}-${twoDigits(date.getUTCMonth() + 1)}-${twoDigits(date.getUTCDate())}`;
+}
+function excelTimeText(serial, includeSeconds, elapsedHours) {
+  const totalSeconds = Math.round((serial - Math.floor(serial)) * 86400);
+  const hours = elapsedHours ? Math.floor(serial * 24) : Math.floor(totalSeconds / 3600) % 24;
+  const minutes = Math.floor(totalSeconds / 60) % 60;
+  const seconds = totalSeconds % 60;
+  return `${twoDigits(hours)}:${twoDigits(minutes)}${includeSeconds ? `:${twoDigits(seconds)}` : ""}`;
+}
+function formattedNumberText(raw, style, date1904) {
+  if (!style) return raw;
+  const value = Number(raw);
+  if (!Number.isFinite(value)) return raw;
+  const normalized = normalizeFormatCode(style.formatCode);
+  const kind = numberFormatKind(style.formatCode);
+  if (kind === "date") return excelDateText(value, date1904) ?? raw;
+  if (kind === "time") {
+    return excelTimeText(value, /s/.test(normalized), /\[[h]+\]/.test(normalized));
+  }
+  if (kind === "datetime") {
+    const date = excelDateText(value, date1904);
+    return date ? `${date} ${excelTimeText(value, /s/.test(normalized), false)}` : raw;
+  }
+  if (kind === "percent") {
+    const decimals = /0\.(0+)%/.exec(normalized)?.[1]?.length ?? 0;
+    return `${(value * 100).toFixed(decimals)}%`;
+  }
+  if (kind === "zero-pad" && Number.isInteger(value)) {
+    const width = normalized.length;
+    const sign = value < 0 ? "-" : "";
+    return `${sign}${String(Math.abs(value)).padStart(width, "0")}`;
+  }
+  return raw;
+}
+function cellText(cell, shared, styles, date1904) {
+  const t = cell.getAttribute("t");
+  if (t === "inlineStr") {
+    const is = cell.getElementsByTagNameNS(NS_SML, "is")[0];
+    return is ? stringItemText(is) : "";
+  }
+  const vEls = cell.getElementsByTagNameNS(NS_SML, "v");
+  const v = vEls.length ? vEls[0].textContent ?? "" : "";
+  if (t === "s") return shared[Number.parseInt(v, 10)] ?? "";
+  if (t === "b") return v === "1" ? "TRUE" : "FALSE";
+  const styleIndex = Number.parseInt(cell.getAttribute("s") ?? "0", 10);
+  return formattedNumberText(v, styles[styleIndex], date1904);
+}
+async function sheetToGrid(pkg, path, shared, styles, date1904) {
+  const doc = await getPartXml(pkg, path);
+  if (!doc) return [];
+  const rowEls = doc.getElementsByTagNameNS(NS_SML, "row");
+  const byIndex = /* @__PURE__ */ new Map();
+  let maxContentIdx = -1;
+  let fallbackIdx = 0;
+  for (let r = 0; r < rowEls.length; r++) {
+    const rowEl = rowEls[r];
+    const cells = rowEl.getElementsByTagNameNS(NS_SML, "c");
+    const rowArr = [];
+    for (let c = 0; c < cells.length; c++) {
+      const cell = cells[c];
+      const ref = cell.getAttribute("r");
+      const idx = ref ? colIndex(ref) : rowArr.length;
+      while (rowArr.length < idx) rowArr.push("");
+      rowArr[idx] = cellText(cell, shared, styles, date1904);
+    }
+    const rowRef = Number.parseInt(rowEl.getAttribute("r") ?? "", 10);
+    const rowIdx = Number.isFinite(rowRef) && rowRef > 0 ? rowRef - 1 : fallbackIdx;
+    fallbackIdx = rowIdx + 1;
+    byIndex.set(rowIdx, rowArr);
+    if (rowArr.some((v) => v !== "")) maxContentIdx = Math.max(maxContentIdx, rowIdx);
+  }
+  const grid = [];
+  for (let i = 0; i <= maxContentIdx; i++) grid.push(byIndex.get(i) ?? []);
+  return grid;
+}
+function gridToTable(grid) {
+  const maxCols = grid.reduce((m, r) => Math.max(m, r.length), 1);
+  const rows = grid.map((cells, rowIdx) => {
+    const children = [];
+    for (let c = 0; c < maxCols; c++) {
+      const value = cells[c] ?? "";
+      children.push({
+        type: "tableCell",
+        ...rowIdx === 0 ? { isHeader: true } : {},
+        children: value ? [{ type: "text", value }] : []
+      });
+    }
+    return { type: "tableRow", children };
+  });
+  return { type: "table", children: rows };
+}
+async function xlsxToMarkdownDoc(data, options = {}) {
+  const pkg = await openPackage(data, options);
+  const mainPart = requireMainPartPath(pkg, XLSX_MAIN_PART, "XLSX");
+  const [{ sheets, date1904 }, shared, styles] = await Promise.all([
+    readWorkbook(pkg, mainPart),
+    readSharedStrings(pkg),
+    readCellStyles(pkg)
+  ]);
+  let selected = sheets;
+  if (options.sheet !== void 0) {
+    const picked = typeof options.sheet === "number" ? sheets[options.sheet] : sheets.find((s) => s.name === options.sheet);
+    selected = picked ? [picked] : [];
+  }
+  const children = [];
+  const single = selected.length === 1 && options.sheet !== void 0;
+  for (const sheet of selected) {
+    const grid = await sheetToGrid(pkg, sheet.path, shared, styles, date1904);
+    if (!single) {
+      children.push({ type: "heading", depth: 1, children: [{ type: "text", value: sheet.name }] });
+    }
+    if (grid.length > 0) children.push(gridToTable(grid));
+  }
+  return { type: "document", children };
+}
+
+// src/xlsx/export.ts
+import { docToMarkdown } from "@bendyline/squisq/doc";
+var NUMERIC_RE = /^-?\d+(\.\d+)?$/;
+function isSafeNumericCell(value) {
+  if (!NUMERIC_RE.test(value)) return false;
+  const unsigned = value.startsWith("-") ? value.slice(1) : value;
+  const [integer] = unsigned.split(".");
+  if (integer.length > 1 && integer.startsWith("0")) return false;
+  const significantDigits = unsigned.replace(".", "").replace(/^0+/, "");
+  if (significantDigits.length > 15) return false;
+  return Number.isFinite(Number(value));
+}
+function cleanSheetName(value) {
+  return value.replace(/[[\]:*?/\\]/g, "").trim().slice(0, 31).trim().replace(/^'+|'+$/g, "");
+}
+function sanitizeSheetName(raw, used, fallback) {
+  const safeFallback = cleanSheetName(fallback) || "Sheet";
+  const name = cleanSheetName(raw) || safeFallback;
+  const key = name.toLocaleLowerCase("en-US");
+  if (!used.has(key)) {
+    used.add(key);
+    return name;
+  }
+  for (let n = 2; ; n++) {
+    const suffix = String(n);
+    const base = name.slice(0, 31 - suffix.length);
+    const candidate = `${base}${suffix}`;
+    const candidateKey = candidate.toLocaleLowerCase("en-US");
+    if (!used.has(candidateKey)) {
+      used.add(candidateKey);
+      return candidate;
+    }
+  }
+}
+function tableToGrid(table) {
+  return table.children.map(
+    (row) => row.children.map((cell) => extractPlainText(cell.children))
+  );
+}
+function collectSheets(doc, prefix) {
+  const used = /* @__PURE__ */ new Set();
+  const sheets = [];
+  let pendingHeading = null;
+  const visit = (nodes) => {
+    for (const node of nodes) {
+      if (node.type === "heading") {
+        pendingHeading = extractPlainText(node.children);
+      } else if (node.type === "table") {
+        const fallback = `${prefix}${sheets.length + 1}`;
+        const name = sanitizeSheetName(pendingHeading ?? fallback, used, fallback);
+        sheets.push({ name, grid: tableToGrid(node) });
+        pendingHeading = null;
+      }
+    }
+  };
+  visit(doc.children);
+  return sheets;
+}
+function columnLetter(index) {
+  let n = index + 1;
+  let letters = "";
+  while (n > 0) {
+    const rem = (n - 1) % 26;
+    letters = String.fromCharCode(65 + rem) + letters;
+    n = Math.floor((n - 1) / 26);
+  }
+  return letters;
+}
+function cellXml(value, ref, inferNumericCells) {
+  if (inferNumericCells && isSafeNumericCell(value)) {
+    return `<c r="${ref}"><v>${escapeXml(value)}</v></c>`;
+  }
+  return `<c r="${ref}" t="inlineStr"><is><t xml:space="preserve">${escapeXml(value)}</t></is></c>`;
+}
+function worksheetXml(grid, inferNumericCells) {
+  const rows = [];
+  for (let r = 0; r < grid.length; r++) {
+    const cells = grid[r];
+    const cellsXml = cells.map((value, c) => cellXml(value, `${columnLetter(c)}${r + 1}`, inferNumericCells)).join("");
+    rows.push(`<row r="${r + 1}">${cellsXml}</row>`);
+  }
+  return `${xmlDeclaration()}
+<worksheet xmlns="${NS_SML}" xmlns:r="${NS_R}"><sheetData>${rows.join("")}</sheetData></worksheet>`;
+}
+function workbookXml(sheets) {
+  const sheetEls = sheets.map(
+    (sheet, i) => `<sheet name="${escapeXml(sheet.name)}" sheetId="${i + 1}" r:id="rId${i + 1}"/>`
+  ).join("");
+  return `${xmlDeclaration()}
+<workbook xmlns="${NS_SML}" xmlns:r="${NS_R}"><sheets>${sheetEls}</sheets></workbook>`;
+}
+function stylesXml() {
+  return `${xmlDeclaration()}
+<styleSheet xmlns="${NS_SML}"><fonts count="1"><font><sz val="11"/><name val="Calibri"/></font></fonts><fills count="1"><fill><patternFill patternType="none"/></fill></fills><borders count="1"><border/></borders><cellStyleXfs count="1"><xf numFmtId="0" fontId="0" fillId="0" borderId="0"/></cellStyleXfs><cellXfs count="1"><xf numFmtId="0" fontId="0" fillId="0" borderId="0" xfId="0"/></cellXfs></styleSheet>`;
+}
+async function markdownDocToXlsx(doc, options = {}) {
+  options.signal?.throwIfAborted();
+  const prefix = cleanSheetName(options.sheetNamePrefix ?? "Sheet") || "Sheet";
+  const inferNumericCells = options.inferNumericCells ?? false;
+  let sheets = collectSheets(doc, prefix);
+  const maxCells = options.maxCells ?? 1e5;
+  if (!Number.isSafeInteger(maxCells) || maxCells < 0) {
+    throw new RangeError("maxCells must be a non-negative safe integer");
+  }
+  const cellCount = sheets.reduce(
+    (total, sheet) => total + sheet.grid.reduce((sum, row) => sum + row.length, 0),
+    0
+  );
+  if (cellCount > maxCells) {
+    throw new RangeError(`XLSX export exceeds the ${maxCells}-cell safety limit`);
+  }
+  if (sheets.length === 0) {
+    sheets = [{ name: sanitizeSheetName(`${prefix}1`, /* @__PURE__ */ new Set(), "Sheet1"), grid: [] }];
+  }
+  const pkg = createPackage();
+  sheets.forEach((sheet, i) => {
+    if ((i & 31) === 0) options.signal?.throwIfAborted();
+    const sheetPath = `xl/worksheets/sheet${i + 1}.xml`;
+    pkg.addPart(
+      sheetPath,
+      worksheetXml(sheet.grid, inferNumericCells),
+      CONTENT_TYPE_XLSX_WORKSHEET
+    );
+    pkg.addRelationship("xl/workbook.xml", {
+      id: `rId${i + 1}`,
+      type: REL_WORKSHEET,
+      target: `worksheets/sheet${i + 1}.xml`
+    });
+  });
+  pkg.addPart("xl/styles.xml", stylesXml(), CONTENT_TYPE_XLSX_STYLES);
+  pkg.addRelationship("xl/workbook.xml", {
+    id: `rId${sheets.length + 1}`,
+    type: REL_STYLES,
+    target: "styles.xml"
+  });
+  pkg.addPart("xl/workbook.xml", workbookXml(sheets), CONTENT_TYPE_XLSX_WORKBOOK);
+  pkg.addRelationship("", {
+    id: "rId1",
+    type: REL_OFFICE_DOCUMENT,
+    target: "xl/workbook.xml"
+  });
+  if (options.title || options.author) {
+    pkg.setCoreProperties({
+      title: options.title,
+      creator: options.author,
+      created: (/* @__PURE__ */ new Date()).toISOString(),
+      modified: (/* @__PURE__ */ new Date()).toISOString()
+    });
+  }
+  return pkg.toArrayBuffer();
+}
+async function docToXlsx(doc, options) {
+  return markdownDocToXlsx(docToMarkdown(doc), options);
+}
+
+// src/xlsx/index.ts
+async function xlsxToDoc(data, options) {
+  return markdownToDoc(await xlsxToMarkdownDoc(data, options));
+}
+
+export {
+  xlsxToMarkdownDoc,
+  markdownDocToXlsx,
+  docToXlsx,
+  xlsxToDoc
+};