warframe-bilder/wireframe-system/scripts/wireframe/extract-documents.ps1

param(
  [Parameter(Mandatory = $false)]
  [string]$InputPath = ".",

  [Parameter(Mandatory = $false)]
  [string]$OutputDir = "workspace/artifacts/wireframe-gen/extracted"
)

$ErrorActionPreference = "Stop"

function New-Slug {
  param([string]$Value)
  $slug = [System.IO.Path]::GetFileNameWithoutExtension($Value).ToLowerInvariant()
  $slug = $slug -replace "[^a-z0-9]+", "-"
  $slug = $slug.Trim("-")
  if ([string]::IsNullOrWhiteSpace($slug)) { return "source" }
  return $slug
}

function Get-TextFileContent {
  param([string]$Path)
  return Get-Content -LiteralPath $Path -Raw -Encoding UTF8
}

function Get-DocxText {
  param([string]$Path)
  Add-Type -AssemblyName System.IO.Compression.FileSystem
  $zip = [System.IO.Compression.ZipFile]::OpenRead((Resolve-Path -LiteralPath $Path))
  try {
    $parts = $zip.Entries | Where-Object {
      $_.FullName -eq "word/document.xml" -or
      $_.FullName -like "word/header*.xml" -or
      $_.FullName -like "word/footer*.xml"
    }
    $texts = New-Object System.Collections.Generic.List[string]
    foreach ($part in $parts) {
      $reader = New-Object System.IO.StreamReader($part.Open())
      try {
        $xml = $reader.ReadToEnd()
        $xml = $xml -replace "</w:p>", "`n"
        $matches = [regex]::Matches($xml, "<w:t[^>]*>(.*?)</w:t>")
        foreach ($match in $matches) {
          $texts.Add([System.Net.WebUtility]::HtmlDecode($match.Groups[1].Value))
        }
        $texts.Add("`n")
      }
      finally {
        $reader.Dispose()
      }
    }
    return (($texts -join "") -replace "`r", "" -replace "`n{3,}", "`n`n").Trim()
  }
  finally {
    $zip.Dispose()
  }
}

function Get-PdfText {
  param([string]$Path)
  $tool = Get-Command pdftotext -ErrorAction SilentlyContinue
  if ($null -eq $tool) {
    return @{
      Text = ""
      Note = "PDF extraction requires pdftotext in PATH. No OCR fallback is used in v1."
      Status = "needs_external_extractor"
    }
  }

  $tempFile = Join-Path ([System.IO.Path]::GetTempPath()) ("wireframe-pdf-" + [guid]::NewGuid().ToString() + ".txt")
  & $tool.Source "-layout" $Path $tempFile | Out-Null
  $text = Get-Content -LiteralPath $tempFile -Raw -Encoding UTF8
  Remove-Item -LiteralPath $tempFile -Force
  return @{
    Text = $text
    Note = "Extracted with pdftotext."
    Status = "ok"
  }
}

function Get-SourceFiles {
  param([string]$Path)
  $resolved = Resolve-Path -LiteralPath $Path
  $item = Get-Item -LiteralPath $resolved
  $extensions = @(".md", ".markdown", ".txt", ".pdf", ".docx")
  if ($item.PSIsContainer) {
    return Get-ChildItem -LiteralPath $item.FullName -Recurse -File |
      Where-Object { $extensions -contains $_.Extension.ToLowerInvariant() } |
      Sort-Object FullName
  }
  if ($extensions -contains $item.Extension.ToLowerInvariant()) {
    return @($item)
  }
  throw "Unsupported input file type: $($item.Extension)"
}

New-Item -ItemType Directory -Force -Path $OutputDir | Out-Null
$files = Get-SourceFiles -Path $InputPath
$sources = New-Object System.Collections.Generic.List[object]
$index = 1

foreach ($file in $files) {
  $sourceId = "SRC-{0:D3}" -f $index
  $extension = $file.Extension.ToLowerInvariant()
  $status = "ok"
  $notes = ""
  $text = ""

  try {
    switch ($extension) {
      ".md" { $text = Get-TextFileContent -Path $file.FullName }
      ".markdown" { $text = Get-TextFileContent -Path $file.FullName }
      ".txt" { $text = Get-TextFileContent -Path $file.FullName }
      ".docx" { $text = Get-DocxText -Path $file.FullName }
      ".pdf" {
        $pdf = Get-PdfText -Path $file.FullName
        $text = $pdf.Text
        $notes = $pdf.Note
        $status = $pdf.Status
      }
    }
  }
  catch {
    $status = "error"
    $notes = $_.Exception.Message
  }

  $slug = New-Slug -Value $file.Name
  $outName = "$sourceId-$slug.txt"
  $outPath = Join-Path $OutputDir $outName
  Set-Content -LiteralPath $outPath -Value $text -Encoding UTF8

  $sources.Add([ordered]@{
    source_id = $sourceId
    path = $file.FullName
    type = $extension.TrimStart(".")
    status = $status
    extracted_text_path = $outPath
    character_count = $text.Length
    notes = $notes
  })
  $index += 1
}

$inventory = [ordered]@{
  generated_at = (Get-Date).ToUniversalTime().ToString("o")
  input_path = (Resolve-Path -LiteralPath $InputPath).Path
  output_dir = (Resolve-Path -LiteralPath $OutputDir).Path
  sources = $sources
}

$inventoryPath = Join-Path $OutputDir "source_inventory.json"
$inventory | ConvertTo-Json -Depth 10 | Set-Content -LiteralPath $inventoryPath -Encoding UTF8
Write-Output "Extracted $($sources.Count) source(s) to $OutputDir"
Write-Output $inventoryPath