homebrew-core/Formula/ocrmypdf.rb

class Ocrmypdf < Formula
  include Language::Python::Virtualenv

  desc "Adds an OCR text layer to scanned PDF files"
  homepage "https://github.com/jbarlow83/OCRmyPDF"
  url "https://files.pythonhosted.org/packages/e5/95/7a9a874ca183a0f61d30fc44d51d9ea42737514173aafc727e6eb325af4c/ocrmypdf-11.0.2.tar.gz"
  sha256 "b9200f07a8b65472392d9aed04986e41c1b307011e51af6a727489ace426f662"
  license "MPL-2.0"

  livecheck do
    url :stable
  end

  bottle do
    cellar :any
    sha256 "243684db373fa773ce5877d05328e9828a719df247e55139260442b40e3f6f9c" => :catalina
    sha256 "ad6a60b71b77b549991264e8b2894ccb8c3bd1fbe81a7a3fefb6798709b28c23" => :mojave
    sha256 "9270b3a2c473b145a8c86ef2ef85ec70c2cbd58918aa175cf4255130431b17bb" => :high_sierra
  end

  depends_on "pkg-config" => :build
  depends_on "freetype"
  depends_on "ghostscript"
  depends_on "jbig2enc"
  depends_on "jpeg"
  depends_on "leptonica"
  depends_on "libpng"
  depends_on "pngquant"
  depends_on "pybind11"
  depends_on "python@3.8"
  depends_on "qpdf"
  depends_on "tesseract"
  depends_on "unpaper"

  uses_from_macos "libffi"
  uses_from_macos "libxml2"
  uses_from_macos "libxslt"
  uses_from_macos "zlib"

  resource "cffi" do
    url "https://files.pythonhosted.org/packages/f7/09/88bbe20b76ca76be052c366fe77aa5e3cd6e5f932766e5597fecdd95b2a8/cffi-1.14.2.tar.gz"
    sha256 "ae8f34d50af2c2154035984b8b5fc5d9ed63f32fe615646ab435b05b132ca91b"
  end

  resource "chardet" do
    url "https://files.pythonhosted.org/packages/fc/bb/a5768c230f9ddb03acc9ef3f0d4a3cf93462473795d18e9535498c8f929d/chardet-3.0.4.tar.gz"
    sha256 "84ab92ed1c4d4f16916e05906b6b75a6c0fb5db821cc65e70cbd64a3e2a5eaae"
  end

  resource "coloredlogs" do
    url "https://files.pythonhosted.org/packages/84/1b/1ecdd371fa68839cfbda15cc671d0f6c92d2c42688df995a9bf6e36f3511/coloredlogs-14.0.tar.gz"
    sha256 "a1fab193d2053aa6c0a97608c4342d031f1f93a3d1218432c59322441d31a505"
  end

  resource "cryptography" do
    url "https://files.pythonhosted.org/packages/12/be/c9cc7d7ab71dbcc9e4e517ead0cdd48e8c9a48d7b8bdddb738e90d08279a/cryptography-3.1.tar.gz"
    sha256 "26409a473cc6278e4c90f782cd5968ebad04d3911ed1c402fc86908c17633e08"
  end

  resource "humanfriendly" do
    url "https://files.pythonhosted.org/packages/6c/19/8e3b4c6fa7cca4788817db398c05274d98ecc6a35e0eaad2846fde90c863/humanfriendly-8.2.tar.gz"
    sha256 "bf52ec91244819c780341a3438d5d7b09f431d3f113a475147ac9b7b167a3d12"
  end

  resource "img2pdf" do
    url "https://files.pythonhosted.org/packages/80/ed/5167992abaf268f5a5867e974d9d36a8fa4802800898ec711f4e1942b4f5/img2pdf-0.4.0.tar.gz"
    sha256 "eaee690ab8403dd1a9cb4db10afee41dd3e6c7ed63bdace02a0121f9feadb0c9"
  end

  resource "lxml" do
    url "https://files.pythonhosted.org/packages/2c/4d/3ec1ea8512a7fbf57f02dee3035e2cce2d63d0e9c0ab8e4e376e01452597/lxml-4.5.2.tar.gz"
    sha256 "cdc13a1682b2a6241080745b1953719e7fe0850b40a5c71ca574f090a1391df6"
  end

  resource "pdfminer.six" do
    url "https://files.pythonhosted.org/packages/34/4a/bc38040c4a970870ccbef975bae2b6b1bccaa2d308e61111203cccc58e89/pdfminer.six-20200726.tar.gz"
    sha256 "8513b27d780ac83e8474cef245c43d1ebf65ca9550dc07ac0525b4d47eb2c310"
  end

  resource "pikepdf" do
    url "https://files.pythonhosted.org/packages/4e/62/47ee84138503c58926947de4a2e08a35c45ffc51e3387fe43e8f9a727aaf/pikepdf-1.19.1.tar.gz"
    sha256 "fffe9cb5e4f026216c0f6f9170c30a1f9093b061275d9562b39fd74d5156a2ae"
  end

  resource "Pillow" do
    url "https://files.pythonhosted.org/packages/3e/02/b09732ca4b14405ff159c470a612979acfc6e8645dc32f83ea0129709f7a/Pillow-7.2.0.tar.gz"
    sha256 "97f9e7953a77d5a70f49b9a48da7776dc51e9b738151b22dacf101641594a626"
  end

  resource "pluggy" do
    url "https://files.pythonhosted.org/packages/f8/04/7a8542bed4b16a65c2714bf76cf5a0b026157da7f75e87cc88774aa10b14/pluggy-0.13.1.tar.gz"
    sha256 "15b2acde666561e1298d71b523007ed7364de07029219b604cf808bfa1c765b0"
  end

  resource "pycparser" do
    url "https://files.pythonhosted.org/packages/0f/86/e19659527668d70be91d0369aeaa055b4eb396b0f387a4f92293a20035bd/pycparser-2.20.tar.gz"
    sha256 "2d475327684562c3a96cc71adf7dc8c4f0565175cf86b6d7a404ff4c771f15f0"
  end

  resource "reportlab" do
    url "https://files.pythonhosted.org/packages/d5/d9/a6fc317cad5d8a182f22d8b8712532c1a558221a77bd6bd8283ca8be96cd/reportlab-3.5.49.tar.gz"
    sha256 "2ccf5165aa64e51abf240cd3f0062b860bb19346bd2c268fb00c33c09a53f8a8"
  end

  resource "six" do
    url "https://files.pythonhosted.org/packages/6b/34/415834bfdafca3c5f451532e8a8d9ba89a21c9743a0c59fbd0205c7f9426/six-1.15.0.tar.gz"
    sha256 "30639c035cdb23534cd4aa2dd52c3bf48f06e5f4a941509c8bafd8ce11080259"
  end

  resource "sortedcontainers" do
    url "https://files.pythonhosted.org/packages/3b/fb/48f6fa11e4953c530b09fa0f2976df5234b0eaabcd158625c3e73535aeb8/sortedcontainers-2.2.2.tar.gz"
    sha256 "4e73a757831fc3ca4de2859c422564239a31d8213d09a2a666e375807034d2ba"
  end

  resource "tqdm" do
    url "https://files.pythonhosted.org/packages/7c/a2/4cc95d7766a5d17ea2541d88da357d5905f75b6dbdfd17dfffd6c37647ae/tqdm-4.48.2.tar.gz"
    sha256 "564d632ea2b9cb52979f7956e093e831c28d441c11751682f84c86fc46e4fd21"
  end

  def install
    venv = virtualenv_create(libexec, Formula["python@3.8"].bin/"python3")

    resource("Pillow").stage do
      inreplace "setup.py" do |s|
        sdkprefix = MacOS.sdk_path_if_needed ? MacOS.sdk_path : ""
        s.gsub! "openjpeg.h", "probably_not_a_header_called_this_eh.h"
        s.gsub! "xcb.h", "probably_not_a_header_called_this_eh.h"
        s.gsub! "ZLIB_ROOT = None",
                "ZLIB_ROOT = ('#{sdkprefix}/usr/lib', '#{sdkprefix}/usr/include')"
        s.gsub! "JPEG_ROOT = None",
                "JPEG_ROOT = ('#{Formula["jpeg"].opt_prefix}/lib', '#{Formula["jpeg"].opt_prefix}/include')"
        s.gsub! "FREETYPE_ROOT = None",
                "FREETYPE_ROOT = ('#{Formula["freetype"].opt_prefix}/lib', " \
                                 "'#{Formula["freetype"].opt_prefix}/include')"
      end

      # avoid triggering "helpful" distutils code that doesn't recognize Xcode 7 .tbd stubs
      unless MacOS::CLT.installed?
        ENV.append "CFLAGS", "-I#{MacOS.sdk_path}/System/Library/Frameworks/Tk.framework/Versions/8.5/Headers"
      end
      venv.pip_install Pathname.pwd
    end

    # Fix "ld: file not found: /usr/lib/system/libsystem_darwin.dylib" for lxml
    ENV["SDKROOT"] = MacOS.sdk_path if MacOS.version == :sierra

    res = resources.map(&:name).to_set - ["Pillow"]
    res.each do |r|
      venv.pip_install resource(r)
    end

    venv.pip_install_and_link buildpath
    bash_completion.install "misc/completion/ocrmypdf.bash" => "ocrmypdf"
    fish_completion.install "misc/completion/ocrmypdf.fish"
  end

  test do
    system "#{bin}/ocrmypdf", "-f", "-q", "--deskew",
                              test_fixtures("test.pdf"), "ocr.pdf"
    assert_predicate testpath/"ocr.pdf", :exist?
  end
end