@ricsam/excel-to-formula-engine 0.1.1 → 0.2.0

package/dist/index.mjs

@@ -1538,6 +1538,238 @@ function coalesceCellsIntoAreas(sheetName, cells) {
   return areas;
 }
 
+// src/csv/parse.ts
+var CANDIDATE_DELIMITERS = [",", "\t", ";", "|"];
+function parseDelimitedText(text, delimiter) {
+  const rows = [];
+  let row = [];
+  let field = "";
+  let inQuotes = false;
+  let fieldWasQuoted = false;
+  let rowHasContent = false;
+  const endField = () => {
+    row.push(field);
+    field = "";
+    fieldWasQuoted = false;
+    rowHasContent = true;
+  };
+  const endRow = () => {
+    endField();
+    rows.push(row);
+    row = [];
+    rowHasContent = false;
+  };
+  for (let i = 0;i < text.length; i++) {
+    const char = text[i];
+    if (inQuotes) {
+      if (char === '"') {
+        if (text[i + 1] === '"') {
+          field += '"';
+          i++;
+        } else {
+          inQuotes = false;
+        }
+      } else {
+        field += char;
+      }
+      continue;
+    }
+    if (char === '"' && field.length === 0) {
+      inQuotes = true;
+      fieldWasQuoted = true;
+      continue;
+    }
+    if (char === delimiter) {
+      endField();
+      continue;
+    }
+    if (char === "\r") {
+      if (text[i + 1] === `
+`)
+        i++;
+      endRow();
+      continue;
+    }
+    if (char === `
+`) {
+      endRow();
+      continue;
+    }
+    field += char;
+  }
+  if (field.length > 0 || fieldWasQuoted || rowHasContent) {
+    endRow();
+  }
+  return rows;
+}
+function detectDelimiter(text, candidates = CANDIDATE_DELIMITERS) {
+  const lines = takeSampleLines(text, 20);
+  if (lines.length === 0)
+    return ",";
+  let best;
+  for (const delimiter of candidates) {
+    const counts = lines.map((line) => countOutsideQuotes(line, delimiter));
+    const present = counts.filter((count) => count > 0);
+    if (present.length === 0)
+      continue;
+    const first = counts[0];
+    const consistent = counts.filter((count) => count === first).length;
+    const score = consistent / counts.length + Math.min(first, 50) / 1000;
+    if (!best || score > best.score) {
+      best = { delimiter, score };
+    }
+  }
+  return best?.delimiter ?? ",";
+}
+function takeSampleLines(text, limit) {
+  const lines = [];
+  let current = "";
+  let inQuotes = false;
+  for (let i = 0;i < text.length && lines.length < limit; i++) {
+    const char = text[i];
+    if (char === '"') {
+      if (inQuotes && text[i + 1] === '"') {
+        current += '""';
+        i++;
+        continue;
+      }
+      inQuotes = !inQuotes;
+      current += char;
+      continue;
+    }
+    if (!inQuotes && (char === `
+` || char === "\r")) {
+      if (char === "\r" && text[i + 1] === `
+`)
+        i++;
+      if (current.trim().length > 0)
+        lines.push(current);
+      current = "";
+      continue;
+    }
+    current += char;
+  }
+  if (current.trim().length > 0 && lines.length < limit)
+    lines.push(current);
+  return lines;
+}
+function countOutsideQuotes(line, delimiter) {
+  let count = 0;
+  let inQuotes = false;
+  for (let i = 0;i < line.length; i++) {
+    const char = line[i];
+    if (char === '"') {
+      if (inQuotes && line[i + 1] === '"') {
+        i++;
+        continue;
+      }
+      inQuotes = !inQuotes;
+      continue;
+    }
+    if (!inQuotes && char === delimiter)
+      count++;
+  }
+  return count;
+}
+function stripBom(text) {
+  return text.charCodeAt(0) === 65279 ? text.slice(1) : text;
+}
+
+// src/csv/convert.ts
+var MAX_CELLS = 2000000;
+function convertDelimitedText(text, options = {}) {
+  const {
+    sheetName = "Sheet1",
+    fileName,
+    parseNumbers = true,
+    parseBooleans = true,
+    trimFields = false
+  } = options;
+  const diagnostics = [];
+  const body = stripBom(text);
+  const delimiter = options.delimiter ?? detectDelimiter(body);
+  const rows = parseDelimitedText(body, delimiter);
+  const widest = rows.reduce((max, row) => Math.max(max, row.length), 0);
+  if (rows.length * widest > MAX_CELLS) {
+    throw new Error(`The file has about ${rows.length * widest} cells, more than the ` + `${MAX_CELLS} this importer will load at once.`);
+  }
+  const content = new Map;
+  for (let rowIndex = 0;rowIndex < rows.length; rowIndex++) {
+    const row = rows[rowIndex];
+    for (let colIndex = 0;colIndex < row.length; colIndex++) {
+      const raw = trimFields ? row[colIndex].trim() : row[colIndex];
+      if (raw.length === 0)
+        continue;
+      const reference = `${columnIndexToLabel(colIndex)}${rowIndex + 1}`;
+      content.set(reference, toCellValue(raw, {
+        parseNumbers,
+        parseBooleans,
+        sheetName,
+        reference,
+        diagnostics
+      }));
+    }
+  }
+  const data = {
+    sheets: [{ name: sheetName, content }],
+    workbookMetadata: {
+      fileName,
+      sheetOrder: [sheetName],
+      diagnostics: diagnostics.length > 0 ? diagnostics : undefined
+    }
+  };
+  return { data, diagnostics };
+}
+function toCellValue(raw, context) {
+  if (raw.startsWith("=") && raw.length > 1) {
+    const translated = translateFormula(raw, {
+      sheetName: context.sheetName,
+      cellReference: context.reference
+    });
+    context.diagnostics.push(...translated.diagnostics);
+    return translated.formula;
+  }
+  if (context.parseBooleans) {
+    const upper = raw.toUpperCase();
+    if (upper === "TRUE")
+      return true;
+    if (upper === "FALSE")
+      return false;
+  }
+  if (context.parseNumbers) {
+    const number = toNumber(raw);
+    if (number !== undefined)
+      return number;
+  }
+  return raw;
+}
+function toNumber(raw) {
+  const text = raw.trim();
+  if (text.length === 0)
+    return;
+  if (!/^[+-]?(\d+(\.\d*)?|\.\d+)([eE][+-]?\d+)?$/.test(text)) {
+    return;
+  }
+  const digits = text.replace(/^[+-]/, "");
+  if (/^0\d/.test(digits))
+    return;
+  const value = Number(text);
+  if (!Number.isFinite(value))
+    return;
+  if (!text.includes(".") && !/[eE]/.test(text) && !Number.isSafeInteger(value)) {
+    return;
+  }
+  return value;
+}
+async function readTextFrom(source) {
+  if (typeof source === "string")
+    return source;
+  if (source instanceof Uint8Array || source instanceof ArrayBuffer) {
+    return new TextDecoder("utf-8").decode(source);
+  }
+  return source.text();
+}
+
 // index.ts
 async function excelToFormulaEngine(source, options) {
   const { data } = await convertWorkbook(source, withFileName(source, options));
@@ -1553,10 +1785,55 @@ function withFileName(source, options) {
   const name = typeof File !== "undefined" && source instanceof File ? source.name : undefined;
   return name ? { ...options, fileName: name } : { ...options };
 }
+async function csvToFormulaEngine(source, options) {
+  const { data } = await csvToFormulaEngineDetailed(source, options);
+  return data;
+}
+async function csvToFormulaEngineDetailed(source, options) {
+  const text = await readTextFrom(source);
+  const resolved = withCsvFileName(source, options);
+  return convertDelimitedText(text, resolved);
+}
+async function spreadsheetToFormulaEngine(source, options) {
+  const { data } = await spreadsheetToFormulaEngineDetailed(source, options);
+  return data;
+}
+async function spreadsheetToFormulaEngineDetailed(source, options) {
+  const bytes = await toBytes2(source);
+  if (looksLikeZip(bytes)) {
+    return convertWorkbook(bytes, withFileName(source, options));
+  }
+  return csvToFormulaEngineDetailed(bytes, withCsvFileName(source, options));
+}
+function looksLikeZip(bytes) {
+  return bytes.length >= 4 && bytes[0] === 80 && bytes[1] === 75 && bytes[2] === 3 && bytes[3] === 4;
+}
+async function toBytes2(source) {
+  if (source instanceof Uint8Array)
+    return source;
+  if (source instanceof ArrayBuffer)
+    return new Uint8Array(source);
+  return new Uint8Array(await source.arrayBuffer());
+}
+function withCsvFileName(source, options) {
+  const name = typeof File !== "undefined" && source instanceof File ? source.name : undefined;
+  const fileName = options?.fileName ?? name;
+  if (!fileName)
+    return { ...options };
+  return {
+    ...options,
+    fileName,
+    sheetName: options?.sheetName ?? (fileName.replace(/\.(csv|tsv|tab|txt)$/i, "").trim() || "Sheet1")
+  };
+}
 export {
   columnIndexToLabel,
   columnLabelToIndex,
+  convertDelimitedText,
   convertWorkbook,
+  csvToFormulaEngine,
+  csvToFormulaEngineDetailed,
+  detectDelimiter,
   excelSerialToIso,
   excelToFormulaEngine,
   excelToFormulaEngineDetailed,
@@ -1564,8 +1841,12 @@ export {
   offsetFormulaReferences,
   parseA1,
   parseA1Range,
+  parseDelimitedText,
+  spreadsheetToFormulaEngine,
+  spreadsheetToFormulaEngineDetailed,
+  stripBom,
   toA1,
   translateFormula
 };
 
-//# debugId=0B0A199FAC38549264756E2164756E21
+//# debugId=D5F9B2A423DC266864756E2164756E21