param(
  [Parameter(Position=0)][string]$InputPath,
  [Parameter(Position=1)][string]$OutputPath,
  [switch]$Execute
)
# kb_sanitize_hashtable_v1.0.ps1
# Minimal, PS5.1-safe sanitizer for text files (eg. hashtable-like).
# - Normalizes common Unicode punctuation to ASCII
# - Trims trailing spaces
# - Ensures CRLF line endings
# If OutputPath is empty, will write back to InputPath only when -Execute is provided.

$ErrorActionPreference = "Stop"

function Normalize-Ascii([string]$s){
  if ([string]::IsNullOrEmpty($s)) { return "" }
  $t = $s

  # Replace curly quotes and dashes
  $t = $t -replace "[\u2018\u2019\u201A\u201B]", "'"    # single quotes
  $t = $t -replace "[\u201C\u201D\u201E\u201F]", '"'    # double quotes
  $t = $t -replace "[\u2013\u2014\u2212]", "-"          # en/em/minus
  $t = $t -replace "\u00A0", " "                        # nbsp -> space
  $t = $t -replace "\u2026", "..."                      # ellipsis
  $t = $t -replace "[\u2000-\u200F\u202A-\u202E]", ""   # bidi/zwsp remove

  # Strip non-ASCII control except CR/LF/TAB
  $chars = New-Object System.Text.StringBuilder
  foreach($ch in $t.ToCharArray()){
    $code = [int][char]$ch
    if ( ($code -eq 9) -or ($code -eq 10) -or ($code -eq 13) -or ($code -ge 32 -and $code -le 126) ){
      [void]$chars.Append($ch)
    } else {
      # drop or replace with '?'
      [void]$chars.Append('?')
    }
  }
  return $chars.ToString()
}

if ([string]::IsNullOrWhiteSpace($InputPath)){
  Write-Host "[INFO] No input path provided. Nothing to do."
  exit 0
}

if (-not (Test-Path -LiteralPath $InputPath)){
  Write-Error ("Input not found: {0}" -f $InputPath)
  exit 1
}

$raw = Get-Content -LiteralPath $InputPath -Raw
# Normalize line endings to CRLF and trimmed lines
$norm = Normalize-Ascii $raw
$norm = ($norm -split "`r?`n") | ForEach-Object { $_.TrimEnd() } | ForEach-Object { $_ + "`r" } | Out-String
# Out-String adds a trailing newline; enforce CRLF
$norm = $norm -replace "`r`n`r`n$", "`r`n"

$target = $OutputPath
if ([string]::IsNullOrWhiteSpace($target)) { $target = $InputPath }

if (-not $Execute){
  Write-Host ("[PREVIEW] Sanitized length = {0} bytes (no write)" -f ([Text.Encoding]::UTF8.GetByteCount($norm)))
  exit 0
}

# Atomic write: write to .tmp then move
$dir = Split-Path -LiteralPath $target -Parent
if (-not (Test-Path -LiteralPath $dir)){ New-Item -ItemType Directory -Path $dir | Out-Null }
$tmp = "$target.tmp"
# Write UTF-8 with BOM
$bytes = [Text.Encoding]::UTF8.GetPreamble() + [Text.Encoding]::UTF8.GetBytes($norm)
[IO.File]::WriteAllBytes($tmp, $bytes)
if (Test-Path -LiteralPath $target){ Remove-Item -LiteralPath $target -Force }
Move-Item -LiteralPath $tmp -Destination $target -Force

Write-Host ("[OK] Written: {0}" -f $target)
exit 0