Remove obsolete PDF test files from MiniPdf.Scripts/pdf_output directory

Author: shps951023

AGENTS.md

@@ -1,2 +1,6 @@
 - English
-- .NET security policy
+- .NET security policy
+
+## Issue Summary Workflow
+When user says "summary on issue #N", use `gh issue comment N --body-file -` 
+to post a summary of completed work to the GitHub issue.

Run-Benchmark.ps1

@@ -0,0 +1,72 @@
+<#
+.SYNOPSIS
+    One-click benchmark: generate Excel → convert to PDF (MiniPdf + LibreOffice) → compare → report.
+
+.DESCRIPTION
+    This script orchestrates the full MiniPdf self-evolution pipeline on Windows.
+    It installs Python dependencies, runs all steps, and opens the final report.
+
+.EXAMPLE
+    .\Run-Benchmark.ps1
+    .\Run-Benchmark.ps1 -CompareOnly
+    .\Run-Benchmark.ps1 -SkipReference
+#>
+
+param(
+    [switch]$CompareOnly,
+    [switch]$SkipGenerate,
+    [switch]$SkipMiniPdf,
+    [switch]$SkipReference,
+    [switch]$SkipInstall
+)
+
+$ErrorActionPreference = "Continue"
+$ScriptRoot = $PSScriptRoot
+$BenchmarkDir = Join-Path $ScriptRoot "tests" "MiniPdf.Benchmark"
+
+Write-Host "`n============================================================" -ForegroundColor Cyan
+Write-Host "  MiniPdf Self-Evolution Benchmark Pipeline" -ForegroundColor Cyan
+Write-Host "============================================================`n" -ForegroundColor Cyan
+
+# Step 0: Install Python dependencies
+if (-not $SkipInstall) {
+    Write-Host "[Step 0] Installing Python dependencies..." -ForegroundColor Yellow
+    pip install openpyxl pymupdf --quiet 2>$null
+    if ($LASTEXITCODE -ne 0) {
+        Write-Host "  WARNING: pip install had issues. Continuing anyway..." -ForegroundColor DarkYellow
+    } else {
+        Write-Host "  OK" -ForegroundColor Green
+    }
+}
+
+# Build args for Python pipeline
+$pyArgs = @()
+if ($CompareOnly) { $pyArgs += "--compare-only" }
+if ($SkipGenerate) { $pyArgs += "--skip-generate" }
+if ($SkipMiniPdf) { $pyArgs += "--skip-minipdf" }
+if ($SkipReference) { $pyArgs += "--skip-reference" }
+
+# Run the benchmark pipeline
+Write-Host "`n[Running] python run_benchmark.py $($pyArgs -join ' ')`n" -ForegroundColor Yellow
+Push-Location $BenchmarkDir
+try {
+    python run_benchmark.py @pyArgs
+} finally {
+    Pop-Location
+}
+
+# Open the report if it exists
+$reportPath = Join-Path $BenchmarkDir "reports" "comparison_report.md"
+if (Test-Path $reportPath) {
+    Write-Host "`n[Done] Report: $reportPath" -ForegroundColor Green
+    Write-Host "Opening report..." -ForegroundColor Cyan
+    # Open in VS Code if available, otherwise notepad
+    $code = Get-Command code -ErrorAction SilentlyContinue
+    if ($code) {
+        code $reportPath
+    } else {
+        Start-Process notepad.exe -ArgumentList $reportPath
+    }
+} else {
+    Write-Host "`nNo report generated. Check the output above for errors." -ForegroundColor Red
+}

src/MiniPdf/ExcelReader.cs

@@ -223,8 +223,27 @@ private static List<List<ExcelCell>> ReadSheet(ZipArchiveEntry entry, List<strin
         var doc = XDocument.Load(stream);
         var ns = doc.Root?.GetDefaultNamespace() ?? XNamespace.None;
 
+        var lastRowNumber = 0;
+
         foreach (var row in doc.Descendants(ns + "row"))
         {
+            // Parse the row number to detect gaps (sparse rows)
+            var rowNumAttr = row.Attribute("r")?.Value;
+            if (int.TryParse(rowNumAttr, out var rowNumber))
+            {
+                // Insert empty rows for any gaps
+                while (lastRowNumber + 1 < rowNumber)
+                {
+                    rows.Add(new List<ExcelCell>());
+                    lastRowNumber++;
+                }
+                lastRowNumber = rowNumber;
+            }
+            else
+            {
+                lastRowNumber++;
+            }
+
             var cells = new List<ExcelCell>();
             var lastColIndex = 0;