# Changelog

## [1.1.11] - 2026-07-08

### 🐛 Critical Fix — XLSX with namespace-prefixed OOXML (`<x:workbook>`)
- **Symptom**: both parsers failed — `ExcelJS: Cannot read properties of undefined (reading 'sheets'). read-excel-file: Cannot read properties of undefined (reading 'trim')`.
- **Root cause (confirmed by dissecting the real file)**: the workbook's `xl/workbook.xml` uses namespace **prefixes** on every element — `<x:workbook><x:sheets><x:sheet .../>` with GUID-style `r:id`s — the signature of a non-Excel/programmatic generator. It's valid transitional OOXML, but ExcelJS and read-excel-file both match unprefixed `<sheet>` tags, find no sheets, and crash on the resulting `undefined`.
- **Fix**: replaced the `read-excel-file` fallback with **SheetJS (`xlsx`)**, which handles prefixed tags, Strict-OOXML, and non-Excel output. Verified against the real file (all 8 sheets, 200 data rows on the RoPA sheet) and against a synthetic prefixed-namespace fixture. SheetJS is lazy-loaded (a broken dep can't block node loading). Fallback chain is now ExcelJS → SheetJS.

### 🔧 Technical Details
- Removed dependency `read-excel-file`; added `xlsx@^0.18.5`.
- Renamed `parseXlsxWithReadExcelFile` → `parseXlsxWithSheetJS`.
- Added regression fixture `test/samples/prefixed-ns.xlsx` (synthetic, no real data) + test asserting ExcelJS fails while SheetJS parses it.

### ⚠️ Note
- `xlsx@0.18.5` is the latest on the npm registry; patched builds (0.20.x, addressing CVE-2023-30533 / CVE-2024-22363) are only on SheetJS's CDN. Consider switching to `https://cdn.sheetjs.com/xlsx-0.20.3/xlsx-0.20.3.tgz` if processing untrusted spreadsheets.

## [1.1.10] - 2026-07-08

### 🐛 Critical Fix — node failed to load on n8n community-node install
- **Symptom**: n8n UI update/install failed with `The specified package could not be loaded`. Server log: `Cannot find module '.../@nerme/.../node_modules/sanitize-html/node_modules/htmlparser2/lib/index.js'` while loading the node (`dist/FileToJsonNode.node.js:65 → require('sanitize-html')`).
- **Root cause**: `sanitize-html` and `cheerio` were **top-level imports**. In some installs (n8n's `~/.n8n/nodes` tree) `sanitize-html`'s nested `htmlparser2` resolves as ESM without `lib/index.js`, so the top-level `require` threw and the **entire node failed to load** — blocking even XLSX/CSV/PDF use.
- **Fix**: `sanitize-html` and `cheerio` are now **lazy-loaded inside `processHtml`**. A broken HTML dependency can no longer prevent the node from loading; it degrades to an HTML-only runtime error. Unblocks the 1.1.9 XLSX fallback fix that couldn't install.

### 📁 Files Modified
- `src/FileToJsonNode.node.ts`: move `cheerio`/`sanitize-html` to dynamic import in `processHtml`
- `package.json`: version bump to 1.1.10
- `CHANGELOG.md`: this entry

## [1.1.9] - 2026-07-08

### 🐛 Critical Bug Fix — XLSX parse crash on Strict/non-Excel workbooks
- **Symptom**: `XLSX processing error: Cannot read properties of undefined (reading 'sheets')` — whole execution aborts on certain `.xlsx` files.
- **Root cause**: The `xlsx` strategy relied solely on ExcelJS 4.4.0. On Strict-format (ISO/IEC 29500) or non-Excel-generated workbooks, ExcelJS's loader hits `model.sheets = workbook.sheets` (`exceljs/lib/xlsx/xlsx.js:323`) where `workbook` is `undefined`, throwing a raw `TypeError`.
- **Fix**: Added a `read-excel-file` fallback. ExcelJS stays the primary parser (rich cell model); if it throws (or returns zero worksheets), the file is re-parsed with `read-excel-file`, which tolerates Strict-OOXML and non-Excel generators. Output shape is identical (letter-keyed columns, `origRow`, `fileName`/`sheetName` metadata toggles).
- **Failure mode when both fail**: a clear `ProcessingError` suggesting a re-save as standard `.xlsx`, instead of a cryptic TypeError.
- **Also removed**: dead officeparser-then-throw preamble in the `xlsx` strategy (result was always discarded).

### 🔧 Technical Details
- New module functions: `parseXlsxWithExcelJS`, `parseXlsxWithReadExcelFile`, `buildSheetsFromRows` (exported for testing).
- New dependency: `read-excel-file@^6.0.3` (same parser upstream `mazixs` adopted in v1.2.2).

### 📁 Files Modified
- `src/FileToJsonNode.node.ts`: primary→fallback XLSX parsing + helper extraction + test exports
- `test/integration/xlsx-fallback.test.ts`: new — shape-parity + fallback-trigger tests (6 cases)
- `package.json`: add `read-excel-file`, version bump to 1.1.9
- `CHANGELOG.md`: this entry

## [1.1.8] - 2025-01-04

### 🐛 Critical Bug Fix (Second Attempt)
- **Fixed n8n Array Format Issue**: Resolved persistent "outputData.entries is not a function" error 
- **Root Cause**: n8n requires return format to be array of arrays: `[separateItems]` not `separateItems`
- **Solution**: Changed `return separateItems;` to `return [separateItems];` to match working grouped output format
- **Verification**: Now matches the exact format used in working grouped mode `[[{json: {...}}]]`

### 🔧 Technical Details
- **Fixed Return Format**: `return [separateItems];` (array containing array of items)
- **Consistent Structure**: Both modes now return `[array_of_items]` format
- **n8n Compatibility**: Proper array nesting for n8n workflow execution engine

### 📁 Files Modified
- `src/FileToJsonNode.node.ts`: Fixed array nesting for separate items output
- `package.json`: Version bump to 1.1.8
- `CHANGELOG.md`: This critical fix entry

---

## [1.1.7] - 2025-01-04

### 🐛 Critical Bug Fix
- **Fixed n8n Compatibility Issue**: Resolved "outputData.entries is not a function" error when using "Output Sheets as Separate Items" feature
- **Root Cause**: n8n requires each workflow item to be wrapped in `{ json: ... }` structure
- **Solution**: Updated return format to properly wrap each sheet item for n8n workflow processing
- **Impact**: "Output Sheets as Separate Items" feature now works correctly in n8n workflows

### 🔧 Technical Details
- **Fixed Return Format**: Each item now properly wrapped as `{ json: separateItem }` instead of raw `separateItem`
- **n8n Compatibility**: Ensures proper iteration over output items in n8n workflow execution
- **No Breaking Changes**: Fix only affects the separate items mode, standard grouped output unchanged

### 📁 Files Modified
- `src/FileToJsonNode.node.ts`: Fixed return format for separate items output
- `package.json`: Version bump to 1.1.7
- `CHANGELOG.md`: This critical fix entry

---

## [1.1.6] - 2025-01-04

### 🚀 Major New Feature: Output Sheets as Separate Items

#### ✨ New Functionality
- **Individual Sheet Processing**: Added "Output Sheets as Separate Items" toggle that outputs each sheet as a separate n8n workflow item instead of grouped by file
- **Perfect for n8n Workflows**: Enables sheet-level processing, parallel execution, and individual sheet transformations
- **Text File Filtering**: PDF, DOCX, and other text-based files are automatically ignored when separate items mode is enabled
- **Clean, Flat Structure**: Removed nested JSON wrappers for streamlined data access

#### 🔧 Interface and Naming Improvements
- **Toggle Renamed**: "Include Spreadsheet Name" → "Include File Name" 
- **Parameter Renamed**: `includeSpreadsheetName` → `includeFileName`
- **Key Unified**: `spreadsheetName` → `fileName` (they were identical values, now consistently named)
- **Simplified Structure**: Removed redundant `"json"` wrapper from output
- **Data Organization**: Changed `"data"` to `"rows"` array for clearer semantics

#### 📊 New Output Structure

**Individual Sheet Items Mode (when toggle enabled):**
```json
[
  {
    "fileName": "example.xlsx",
    "sheetName": "Sheet1",
    "fileType": "xlsx",
    "fileSize": 23456,
    "processedAt": "2025-01-04T12:00:00.000Z",
    "rows": [
      {"A": "Value1", "B": "Value2"},
      {"A": "Value3", "B": "Value4"}
    ]
  },
  {
    "fileName": "example.xlsx",
    "sheetName": "Sheet2", 
    "fileType": "xlsx",
    "fileSize": 23456,
    "processedAt": "2025-01-04T12:00:00.000Z",
    "rows": [
      {"A": "Data1", "B": "Data2"}
    ]
  }
]
```

**Standard Grouped Mode (default, unchanged):**
```json
[{
  "json": {
    "files": [
      {
        "sheets": {
          "Sheet1": {
            "fileName": "example.xlsx",
            "sheetName": "Sheet1",
            "data": [...]
          }
        }
      }
    ]
  }
}]
```

#### 🎯 Metadata Handling
- **Always Included**: `fileType`, `fileSize`, `processedAt`, `rows`
- **Conditionally Included**: `fileName` and `sheetName` based on their respective toggle settings
- **Flat Structure**: All metadata at the same level as the rows data
- **No Redundancy**: Eliminated duplicate information between fileName and spreadsheetName

#### 🔄 Backward Compatibility
- **Default Behavior**: New toggle defaults to `false`, maintaining existing grouped output
- **Existing Workflows**: No breaking changes for current implementations
- **Progressive Enhancement**: Users can opt-in to new functionality as needed

#### 💼 Use Cases
- **Parallel Processing**: Process different sheets simultaneously in n8n workflows
- **Sheet-Level Filtering**: Apply different logic to different sheets
- **Individual Transformations**: Transform each sheet with specific business rules
- **Conditional Processing**: Skip or process sheets based on their content or name

#### 📁 Files Modified
- `src/FileToJsonNode.node.ts`: Complete overhaul of output logic and parameter handling
- `package.json`: Version bump to 1.1.6
- `README.md`: Updated documentation with new features and examples
- `CHANGELOG.md`: This comprehensive changelog entry

#### 🎉 Impact
This release transforms the node from a file-centric processor to a flexible sheet-centric processor, opening up new possibilities for granular data processing in n8n workflows while maintaining full backward compatibility.

---

## [1.1.5] - 2025-01-04

### ✨ New Features
- **Include Spreadsheet Name Toggle**: Added user-configurable option to control whether to include the original filename in each sheet object (default: enabled)
- **Include Sheet Name Toggle**: Added user-configurable option to control whether to include the sheet name in each sheet object (default: enabled)
- **Enhanced n8n Interface**: Added new toggle controls in the node configuration panel
- **Flexible Output Structure**: Sheet objects now support dynamic metadata inclusion based on user preferences

### 🔧 Technical Implementation
- Extended `ProcessingOptions` interface with `includeSpreadsheetName` and `includeSheetName` parameters
- Updated all sheet processing strategies (xlsx, csv, processExcel, streamCsvStrategy) to respect toggle settings
- Implemented conditional object spreading for dynamic metadata inclusion
- Added proper TypeScript type support for flexible sheet structures

### 📊 Output Structure Examples

**Full metadata (default behavior):**
```json
{
  "sheets": {
    "Sheet1": {
      "spreadsheetName": "example.xlsx",
      "sheetName": "Sheet1", 
      "data": [{"A": "Value1", "B": "Value2"}]
    }
  }
}
```

**Minimal output (both toggles disabled):**
```json
{
  "sheets": {
    "Sheet1": {
      "data": [{"A": "Value1", "B": "Value2"}]
    }
  }
}
```

### 🎯 Impact
- **Backward Compatibility**: All new toggles default to enabled, maintaining existing workflow behavior
- **User Control**: Users can now customize output structure based on their needs
- **Data Efficiency**: Option to reduce output size by excluding unnecessary metadata
- **Applies to**: Both XLSX and CSV file processing

### 📁 Files Modified
- `src/FileToJsonNode.node.ts`: Added toggle options and processing logic
- `README.md`: Updated documentation with new features and examples
- `package.json`: Version bump to 1.1.5

---

## [1.1.4] - 2025-01-04

### ✨ New Features  
- **Automatic Spreadsheet Name Inclusion**: Each sheet object now automatically includes the source spreadsheet filename
- **Automatic Sheet Name Inclusion**: Each sheet object now includes the sheet name for better data lineage
- **Enhanced Metadata Structure**: Improved data tracking capabilities

### 🔧 Technical Changes
- Modified xlsx and csv processing strategies to include filename metadata
- Updated TypeScript interfaces to support new sheet structure
- Enhanced processExcel function with metadata handling

### 📊 Breaking Change
- **Sheet Structure**: Changed from `sheets.SheetName: [...]` to `sheets.SheetName: {spreadsheetName, sheetName, data: [...]}`
- **Migration**: Update downstream processes expecting the old array format

---

## [1.0.11.1] - 2025-01-27

### 🔧 Bug Fixes
- **GitHub Actions**: Fixed permissions error in release workflow
  - Added `checks: write` permission to release.yml
  - Resolves workflow error when calling ci.yml from release.yml
- **Code Quality**: Fixed all TypeScript linter errors (18 issues)
  - Replaced `any` types with proper interfaces for YML processing
  - Added type-safe interfaces: YmlCurrency, YmlCategory, YmlOffer, YmlShop, YmlCatalog
  - Fixed `require()` import in integration tests
  - Added eslint-disable comments for test files where needed
- **Build**: All tests passing (60/60) and linter clean

### 📝 Technical Details
- Enhanced type safety for YML processing functions
- Improved code maintainability and IDE support
- No functional changes - purely technical improvements

---

## [1.0.11] - 2025-01-27

### ✨ New Features
- **YML Support**: Added comprehensive support for YML file format
  - Specialized processing for Yandex Market catalog files (yml_catalog format)
  - Structured JSON output with sections: shop_info, currencies, categories, offers, statistics
  - Automatic detection of Yandex Market YML structure
  - Fallback to standard XML processing for regular YML files
  - Support for product parameters, images, delivery options
  - Performance optimized for typical catalog sizes with warnings for large datasets

### 📄 Technical Implementation
- Added `processYandexMarketYml` function for specialized YML processing
- Enhanced file type detection and processing strategies
- Comprehensive test coverage with integration and unit tests
- Updated documentation with YML examples and usage guidelines

### 📁 Files Added/Modified
- `src/FileToJsonNode.node.ts`: Added YML processing strategy
- `test/samples/sample_yandex_market.yml`: Sample YML test file
- `test/integration/yml-integration.test.ts`: Integration tests
- `test/unit/yml-processor.test.ts`: Unit tests
- `docs/yml_support.md`: Comprehensive YML documentation
- `README.md`: Updated with YML support information

### 🎯 Impact
- Users can now process Yandex Market catalog files with structured output
- Enhanced data extraction capabilities for e-commerce integrations
- Backward compatible - no breaking changes

---

## [1.0.10] - 2025-06-20

### 🐛 Bug Fixes
- **Critical**: Fixed support for ODT, ODP, ODS, and JSON file formats
  - Added missing format extensions to supported formats list
  - Resolves "Unsupported file type" error for these formats
  - Format processing strategies were already implemented but not accessible

### 📋 Technical Details
- Added `odt`, `odp`, `ods`, `json` to the `supported` array in FileToJsonNode
- All format handlers were previously implemented in the `strategies` object
- This was a configuration oversight that prevented format recognition

### 🔧 Files Changed
- `src/FileToJsonNode.node.ts`: Updated supported formats array

### 🎯 Impact
- Users can now successfully process OpenDocument formats (ODT, ODP, ODS)
- JSON files are now properly recognized and processed
- No breaking changes - purely additive fix

---

## [1.0.9] - 2025-06-20

### 🐛 Bug Fixes
- **CI/CD**: Fixed Jest parameter compatibility issues
  - Updated `--testPathPattern` to `--testPathPatterns` in all CI commands
  - Resolves Jest 30+ compatibility problems
  - All CI tests now pass successfully

### 🔧 Files Changed
- `.github/workflows/ci.yml`: Updated Jest command parameters

---

## Previous Versions

For changes in versions 1.0.8 and earlier, please see the [GitHub releases page](https://github.com/mazixs/n8n-node-converter-documents/releases).