juxscript 1.1.205 → 1.1.207

package/lib/storage/TabularDriver.js

@@ -486,12 +486,63 @@ export class TabularDriver {
             df = df.select(...selectCols);
         return df;
     }
+    /**
+     * Read raw cell values from first sheet of an Excel file.
+     * Returns rows with their actual sheet row indices.
+     * Used by both the preview UI and the parser to ensure consistency.
+     */
+    async readRawExcelRows(file, maxRows = 15) {
+        let XLSX;
+        try {
+            XLSX = await import('xlsx');
+        }
+        catch {
+            throw new Error('XLSX support requires the "xlsx" package.');
+        }
+        const buffer = await file.arrayBuffer();
+        const workbook = XLSX.read(buffer, {
+            type: 'array',
+            sheetRows: maxRows + 5,
+            dense: false
+        });
+        const sheetName = workbook.SheetNames[0];
+        const worksheet = workbook.Sheets[sheetName];
+        const ref = worksheet['!ref'];
+        if (!ref)
+            return [];
+        const range = XLSX.utils.decode_range(ref);
+        const startRow = range.s.r;
+        const endRow = Math.min(range.e.r, startRow + maxRows - 1);
+        const startCol = range.s.c;
+        const endCol = range.e.c;
+        console.log(`[readRawExcelRows] ref=${ref}, startRow=${startRow}, endRow=${endRow}, cols=${startCol}-${endCol}`);
+        const rows = [];
+        for (let r = startRow; r <= endRow; r++) {
+            const values = [];
+            for (let c = startCol; c <= endCol; c++) {
+                const addr = XLSX.utils.encode_cell({ r, c });
+                const cell = worksheet[addr];
+                if (!cell) {
+                    values.push(null);
+                    continue;
+                }
+                if (cell.w !== undefined) {
+                    values.push(cell.w);
+                    continue;
+                }
+                if (cell.v !== undefined) {
+                    values.push(cell.v);
+                    continue;
+                }
+                values.push(null);
+            }
+            rows.push({ sheetRow: r, values });
+        }
+        return rows;
+    }
     /**
      * ✅ FIXED: Stream Excel file with optional headerRow override
-     * headerRow is 0-based: 0 = first row, 1 = second row, etc.
-     *
-     * Uses direct cell access instead of sheet_to_json to avoid
-     * issues with blank row handling and sparse arrays.
+     * headerRow is the absolute sheet row index (same as sheetRow from readRawExcelRows).
      */
     async streamFileMultiSheet(file, options = {}) {
         const { maxSheetSize = 100000, sheetChunkSize = 10000, onProgress, headerRow = 0 } = options;
@@ -531,7 +582,6 @@ export class TabularDriver {
                 processedSheets++;
                 continue;
             }
-            // Direct cell reader - bypasses sheet_to_json completely
             const readCellValue = (r, c) => {
                 const addr = XLSX.utils.encode_cell({ r, c });
                 const cell = worksheet[addr];
@@ -550,46 +600,37 @@ export class TabularDriver {
                 }
                 return vals;
             };
-            // ✅ FIX: headerRow is the user's 0-based index counting from
-            // the first row of the file. Do NOT add startRow — the user
-            // sees row 0 as the first row regardless of where the sheet
-            // range begins. We simply use headerRow as the absolute sheet
-            // row index.
+            // headerRow is the absolute sheet row index — same value as
+            // sheetRow from readRawExcelRows(). No offset needed.
             const headerSheetRow = headerRow;
-            console.log(`[TabularDriver] Sheet "${sheetName}": range=${ref}, startRow=${startRow}, endRow=${endRow}, startCol=${startCol}, endCol=${endCol}`);
-            console.log(`[TabularDriver] headerRow=${headerRow}, headerSheetRow=${headerSheetRow} (absolute, no startRow offset)`);
+            console.log(`[TabularDriver] Sheet "${sheetName}": range=${ref}, startRow=${startRow}, endRow=${endRow}`);
+            console.log(`[TabularDriver] headerRow=${headerRow}, headerSheetRow=${headerSheetRow}`);
             if (headerSheetRow > endRow) {
-                console.warn(`[TabularDriver] headerRow ${headerRow} (sheet row ${headerSheetRow}) exceeds endRow ${endRow}`);
+                console.warn(`[TabularDriver] headerRow ${headerRow} exceeds endRow ${endRow}`);
                 processedSheets++;
                 continue;
             }
-            // Read header values directly from cells
             const headerValues = readRow(headerSheetRow);
             console.log(`[TabularDriver] Raw header values at sheet row ${headerSheetRow}:`, headerValues);
-            // Build headers array
             const headers = headerValues.map((h, i) => {
                 if (h === null || h === undefined || String(h).trim() === '') {
                     return `__EMPTY${i > 0 ? '_' + i : ''}`;
                 }
                 return String(h).trim();
             });
-            // Count valid headers
             const validHeaders = headers.filter(h => !h.startsWith('__EMPTY'));
             console.log(`[TabularDriver] Headers (${validHeaders.length} valid / ${headers.length} total):`, headers);
             if (validHeaders.length === 0) {
-                console.warn(`[TabularDriver] No valid headers found at row ${headerRow} (sheet row ${headerSheetRow}) in sheet "${sheetName}"`);
-                // Log surrounding rows for debugging
+                console.warn(`[TabularDriver] No valid headers found at row ${headerRow} in sheet "${sheetName}"`);
                 for (let debugR = Math.max(startRow, headerSheetRow - 2); debugR <= Math.min(endRow, headerSheetRow + 2); debugR++) {
                     console.log(`[TabularDriver]   row ${debugR}:`, readRow(debugR));
                 }
                 processedSheets++;
                 continue;
             }
-            // Build data rows: everything after the header row
             const rows = [];
             for (let r = headerSheetRow + 1; r <= endRow; r++) {
                 const rowData = readRow(r);
-                // Skip completely empty rows
                 const hasContent = rowData.some(cell => cell !== null && cell !== undefined && String(cell).trim() !== '');
                 if (!hasContent)
                     continue;
@@ -600,9 +641,6 @@ export class TabularDriver {
                 rows.push(row);
             }
             console.log(`[TabularDriver] Built ${rows.length} data rows for sheet "${sheetName}"`);
-            if (rows.length > 0) {
-                console.log(`[TabularDriver] First row:`, rows[0]);
-            }
             if (rows.length > 0) {
                 sheets[sheetName] = new DataFrame(rows);
             }

package/lib/storage/TabularDriver.ts

@@ -606,12 +606,59 @@ export class TabularDriver {
         return df;
     }
 
+    /**
+     * Read raw cell values from first sheet of an Excel file.
+     * Returns rows with their actual sheet row indices.
+     * Used by both the preview UI and the parser to ensure consistency.
+     */
+    async readRawExcelRows(file: File, maxRows: number = 15): Promise<{ sheetRow: number; values: any[] }[]> {
+        let XLSX: any;
+        try {
+            XLSX = await import('xlsx');
+        } catch {
+            throw new Error('XLSX support requires the "xlsx" package.');
+        }
+
+        const buffer = await file.arrayBuffer();
+        const workbook = XLSX.read(buffer, {
+            type: 'array',
+            sheetRows: maxRows + 5,
+            dense: false
+        });
+
+        const sheetName = workbook.SheetNames[0];
+        const worksheet = workbook.Sheets[sheetName];
+        const ref = worksheet['!ref'];
+        if (!ref) return [];
+
+        const range = XLSX.utils.decode_range(ref);
+        const startRow = range.s.r;
+        const endRow = Math.min(range.e.r, startRow + maxRows - 1);
+        const startCol = range.s.c;
+        const endCol = range.e.c;
+
+        console.log(`[readRawExcelRows] ref=${ref}, startRow=${startRow}, endRow=${endRow}, cols=${startCol}-${endCol}`);
+
+        const rows: { sheetRow: number; values: any[] }[] = [];
+        for (let r = startRow; r <= endRow; r++) {
+            const values: any[] = [];
+            for (let c = startCol; c <= endCol; c++) {
+                const addr = XLSX.utils.encode_cell({ r, c });
+                const cell = worksheet[addr];
+                if (!cell) { values.push(null); continue; }
+                if (cell.w !== undefined) { values.push(cell.w); continue; }
+                if (cell.v !== undefined) { values.push(cell.v); continue; }
+                values.push(null);
+            }
+            rows.push({ sheetRow: r, values });
+        }
+
+        return rows;
+    }
+
     /**
      * ✅ FIXED: Stream Excel file with optional headerRow override
-     * headerRow is 0-based: 0 = first row, 1 = second row, etc.
-     * 
-     * Uses direct cell access instead of sheet_to_json to avoid
-     * issues with blank row handling and sparse arrays.
+     * headerRow is the absolute sheet row index (same as sheetRow from readRawExcelRows).
      */
     async streamFileMultiSheet(file: File, options: ParseOptions = {}): Promise<Record<string, DataFrame>> {
         const { maxSheetSize = 100000, sheetChunkSize = 10000, onProgress, headerRow = 0 } = options;
@@ -659,7 +706,6 @@ export class TabularDriver {
                 continue;
             }
 
-            // Direct cell reader - bypasses sheet_to_json completely
             const readCellValue = (r: number, c: number): any => {
                 const addr = XLSX.utils.encode_cell({ r, c });
                 const cell = worksheet[addr];
@@ -677,27 +723,22 @@ export class TabularDriver {
                 return vals;
             };
 
-            // ✅ FIX: headerRow is the user's 0-based index counting from
-            // the first row of the file. Do NOT add startRow — the user
-            // sees row 0 as the first row regardless of where the sheet
-            // range begins. We simply use headerRow as the absolute sheet
-            // row index.
+            // headerRow is the absolute sheet row index — same value as
+            // sheetRow from readRawExcelRows(). No offset needed.
             const headerSheetRow = headerRow;
 
-            console.log(`[TabularDriver] Sheet "${sheetName}": range=${ref}, startRow=${startRow}, endRow=${endRow}, startCol=${startCol}, endCol=${endCol}`);
-            console.log(`[TabularDriver] headerRow=${headerRow}, headerSheetRow=${headerSheetRow} (absolute, no startRow offset)`);
+            console.log(`[TabularDriver] Sheet "${sheetName}": range=${ref}, startRow=${startRow}, endRow=${endRow}`);
+            console.log(`[TabularDriver] headerRow=${headerRow}, headerSheetRow=${headerSheetRow}`);
 
             if (headerSheetRow > endRow) {
-                console.warn(`[TabularDriver] headerRow ${headerRow} (sheet row ${headerSheetRow}) exceeds endRow ${endRow}`);
+                console.warn(`[TabularDriver] headerRow ${headerRow} exceeds endRow ${endRow}`);
                 processedSheets++;
                 continue;
             }
 
-            // Read header values directly from cells
             const headerValues = readRow(headerSheetRow);
             console.log(`[TabularDriver] Raw header values at sheet row ${headerSheetRow}:`, headerValues);
 
-            // Build headers array
             const headers: string[] = headerValues.map((h: any, i: number) => {
                 if (h === null || h === undefined || String(h).trim() === '') {
                     return `__EMPTY${i > 0 ? '_' + i : ''}`;
@@ -705,13 +746,11 @@ export class TabularDriver {
                 return String(h).trim();
             });
 
-            // Count valid headers
             const validHeaders = headers.filter(h => !h.startsWith('__EMPTY'));
             console.log(`[TabularDriver] Headers (${validHeaders.length} valid / ${headers.length} total):`, headers);
 
             if (validHeaders.length === 0) {
-                console.warn(`[TabularDriver] No valid headers found at row ${headerRow} (sheet row ${headerSheetRow}) in sheet "${sheetName}"`);
-                // Log surrounding rows for debugging
+                console.warn(`[TabularDriver] No valid headers found at row ${headerRow} in sheet "${sheetName}"`);
                 for (let debugR = Math.max(startRow, headerSheetRow - 2); debugR <= Math.min(endRow, headerSheetRow + 2); debugR++) {
                     console.log(`[TabularDriver]   row ${debugR}:`, readRow(debugR));
                 }
@@ -719,12 +758,10 @@ export class TabularDriver {
                 continue;
             }
 
-            // Build data rows: everything after the header row
             const rows: Record<string, any>[] = [];
             for (let r = headerSheetRow + 1; r <= endRow; r++) {
                 const rowData = readRow(r);
 
-                // Skip completely empty rows
                 const hasContent = rowData.some(cell =>
                     cell !== null && cell !== undefined && String(cell).trim() !== ''
                 );
@@ -738,9 +775,6 @@ export class TabularDriver {
             }
 
             console.log(`[TabularDriver] Built ${rows.length} data rows for sheet "${sheetName}"`);
-            if (rows.length > 0) {
-                console.log(`[TabularDriver] First row:`, rows[0]);
-            }
 
             if (rows.length > 0) {
                 sheets[sheetName] = new DataFrame(rows);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "juxscript",
-  "version": "1.1.205",
+  "version": "1.1.207",
   "type": "module",
   "description": "A JavaScript UX authorship platform",
   "main": "index.js",