pdf_utils.rb

# frozen_string_literal: true
require "date"
require "hexapdf"
require "pathname"
require "stringio"
require "tempfile"
require 'furi'

module PDFUtils
  MIME_TYPE = "application/pdf"
  APPROX_FILE_SIZE = 160_000
  LICENSE_TEMPLATE_NAME = "license-v3"
  W8_TEMPLATE_NAME = "fw8ben"
  W9_TEMPLATE_NAME = "fw9"

  TEMPLATE_CONFIG = {
    "fw8ben" => {
      title: "W-8BEN Tax Form",
      requires_address: true,
      required_profile_fields: %i[legal_name taxid birthday],
      signature_field: "part3_signature",
      fields: {
        "1_name": "topmostSubform[0].Page1[0].f_1[0]",
        "2_country": "topmostSubform[0].Page1[0].f_2[0]",
        "3_address": "topmostSubform[0].Page1[0].f_3[0]",
        "3_city": "topmostSubform[0].Page1[0].f_4[0]",
        "3_country": "topmostSubform[0].Page1[0].f_5[0]",
        "4_mailing_address": "topmostSubform[0].Page1[0].f_6[0]",
        "4_city": "topmostSubform[0].Page1[0].f_7[0]",
        "4_country": "topmostSubform[0].Page1[0].f_8[0]",
        "5_us_taxid": "topmostSubform[0].Page1[0].f_9[0]",
        "6a_foreign_taxid": "topmostSubform[0].Page1[0].f_10[0]",
        "6b_tin_required": "topmostSubform[0].Page1[0].c1_01[0]",
        "7_reference_numbres": "topmostSubform[0].Page1[0].f_11[0]",
        "8_birthday": "topmostSubform[0].Page1[0].f_12[0]",
        "9_resident": "topmostSubform[0].Page1[0].f_13[0]",
        "10_name": "topmostSubform[0].Page1[0].f_14[0]",
        "10_percentage": "topmostSubform[0].Page1[0].f_15[0]",
        "10_income": "topmostSubform[0].Page1[0].f_16[0]",
        "10_rate": "topmostSubform[0].Page1[0].f_17[0]",
        "10_amount": "topmostSubform[0].Page1[0].f_18[0]",
        "part3_confirmation": "topmostSubform[0].Page1[0].c1_02[0]",
        "part3_signature": "topmostSubform[0].Page1[0].f_20[0]",
        "part3_date": "topmostSubform[0].Page1[0].Date[0]",
        "part3_name": "topmostSubform[0].Page1[0].f_21[0]",
      }
    },

    "fw9" => {
      title: "W-9 Tax Form",
      requires_address: true,
      required_profile_fields: %i[legal_name taxid],
      signature_field: "part2_signature",
      fields: {
        "1_name": "topmostSubform[0].Page1[0].f1_01[0]",
        "2_business_name": "topmostSubform[0].Page1[0].f1_02[0]",
        "3a_invidual": "topmostSubform[0].Page1[0].Boxes3a-b_ReadOrder[0].c1_1[0]",
        "3a_c_corporation": "topmostSubform[0].Page1[0].Boxes3a-b_ReadOrder[0].c1_1[1]",
        "3a_s_corporation": "topmostSubform[0].Page1[0].Boxes3a-b_ReadOrder[0].c1_1[2]",
        "3a_partnership": "topmostSubform[0].Page1[0].Boxes3a-b_ReadOrder[0].c1_1[3]",
        "3a_trust": "topmostSubform[0].Page1[0].Boxes3a-b_ReadOrder[0].c1_1[4]",
        "3a_llc": "topmostSubform[0].Page1[0].Boxes3a-b_ReadOrder[0].c1_1[5]",
        "3a_tax_classification": "topmostSubform[0].Page1[0].Boxes3a-b_ReadOrder[0].f1_03[0]",
        "3a_other": "topmostSubform[0].Page1[0].Boxes3a-b_ReadOrder[0].c1_1[6]",
        "3a_corporation_type": "topmostSubform[0].Page1[0].Boxes3a-b_ReadOrder[0].f1_04[0]",
        "3b": "topmostSubform[0].Page1[0].Boxes3a-b_ReadOrder[0].c1_2[0]",
        "4_payee_code": "topmostSubform[0].Page1[0].f1_05[0]",
        "4_fatca": "topmostSubform[0].Page1[0].f1_06[0]",
        "5_address": "topmostSubform[0].Page1[0].Address_ReadOrder[0].f1_07[0]",
        "6_city_state_zip": "topmostSubform[0].Page1[0].Address_ReadOrder[0].f1_08[0]",
        "6_requester_name": "topmostSubform[0].Page1[0].f1_09[0]",
        "7_account_numbers": "topmostSubform[0].Page1[0].f1_10[0]",
        part1_ssn_1: "topmostSubform[0].Page1[0].f1_11[0]",
        part1_ssn_2: "topmostSubform[0].Page1[0].f1_12[0]",
        part1_ssn_3: "topmostSubform[0].Page1[0].f1_13[0]",
        part1_ein_1: "topmostSubform[0].Page1[0].f1_14[0]",
        part1_ein_2: "topmostSubform[0].Page1[0].f1_15[0]",
        part2_signature: "SignatureField",
        part2_date: "DateField",
      },
    },

    "license-v3" => {
      title: "License Agreement",
      requires_address: false,
      required_profile_fields: %i[legal_name],
      signature_field: "signature",
      fields: {
        name: "name",
        date: "date",
        signature: "signature",
      }
    },
  }.with_indifferent_access.freeze

  TEST_SIGNATURE_PATH = Rails.root.join("test/fixtures/files/signature.png")
  TEST_FILE_PATH = Rails.root.join("tmp/test-document.pdf")
  # Even in development, we use production links
  # To ensure external iframe PDF preview works locally.
  FAKE_DATA = "%PDF-1.4\nfake\n%%EOF\n"

  # 1x1 full transparency PNG
  BLANK_IMAGE = StringIO.new(
    "\x89PNG\r\n\x1A\n" \
    "\x00\x00\x00\rIHDR" \
    "\x00\x00\x00\x01\x00\x00\x00\x01" \
    "\x08\x06\x00\x00\x00\x1F\x15\xC4\x89" \
    "\x00\x00\x00\x0AIDAT" \
    "\x78\x9C\x63\x00\x01\x00\x00\x05\x00\x01" \
    "\x0D\x0A\x2D\xB4" \
    "\x00\x00\x00\x00IEND\xAE\x42\x60\x82",
    "rb"
  )

  module_function

  def preview_checksum(template_name)
    Digest::SHA256.file(preview_path(template_name)).hexdigest
  end

  def test_signature
    File.open(Rails.root.join("test/fixtures/files/signature.png"))
  end

  def fake_data_io
    StringIO.new(FAKE_DATA)
  end

  def write_test_license
    generate_license(
      name: "Bradley Kam",
      date: Date.today,
      signature: test_signature,
    ) do |pdf_io|
      File.binwrite(TEST_FILE_PATH, pdf_io.read)
    end
  end

  def write_test_w8
    generate_io(
      W8_TEMPLATE_NAME,
      "1_name": "Bradley Kam",
      "2_country": "Singapore",
      "3_address": "111 Mockingbird Lane",
      "3_city": "Singapore",
      "3_country": "Singapore",
      "6a_foreign_taxid": "S12345678",
      "8_birthday": "02/03/2026",
      "part3_confirmation": true,
      "part3_signature": "Bradley Kam",
      "part3_date": "02/03/2026",
      "part3_name": "Bradley Kam",
    ) do |pdf_io|
      File.binwrite(TEST_FILE_PATH, pdf_io.read)
    end
  end

  def write_test_w9
    generate_io(
      W9_TEMPLATE_NAME,
      "1_name": "Bradley Kam",
      "3a_invidual": true,
      "5_address": "111 Mockingbird Lane",
      "6_city_state_zip": "San Francisco, CA 94105",
      part1_ssn_1: "123",
      part1_ssn_2: "45",
      part1_ssn_3: "6789",
      part2_signature: test_signature,
      part2_date: "02/03/2026",
    ) do |pdf_io|
      File.binwrite(TEST_FILE_PATH, pdf_io.read)
    end
  end

  def generate_license(**fields, &block)
    generate_io("license-v3", **fields, &block)
  end

  def generate_document(template_name, **fields)
    doc = open_template(template_name)
    # ensure_font_in_acroform_dr(doc, :BrushScriptStd)
    form = doc.acro_form
    fields_map = TEMPLATE_CONFIG.fetch(template_name).fetch(:fields)
    raise "Form not found" unless form
    fields.each do |name, value|
      name = fields_map[name] if fields_map[name]
      next if value.nil?
      value = formatted_date(value) if value.is_a?(Date)
      if value.respond_to?(:to_io)
        replace_field(doc, form, name.to_s, value)
      else
        set_field_value(form, name.to_s, value)
      end
    end

    form.flatten
    # form.create_appearances
    doc
  end


  def ensure_font_in_acroform_dr(doc, font_key)
    form = doc.acro_form
    form[:DR] ||= {}
    form[:DR][:Font] ||= {}

    return if form[:DR][:Font].key?(font_key)

    any_font = nil
    form.each_field do |f|
      f.each_widget do |w|
        ap = w[:AP]
        next unless ap && ap[:N]
        n = ap[:N].is_a?(HexaPDF::Dictionary) ? ap[:N].values.first : ap[:N]
        res = n[:Resources]
        fonts = res && res[:Font]
        next unless fonts && fonts[font_key]
        any_font = fonts[font_key]
        break
      end
      break if any_font
    end

    raise "Font #{font_key} not found in widget resources" unless any_font

    form[:DR][:Font][font_key] = any_font
  end

  def generate_io(template_name, **fields, &block)
    raise ArgumentError, "generate_io requires a block" unless block
    signature = fields[:signature]
    if signature.respond_to?(:content_type) && signature.content_type != "image/png"
      raise ArgumentError, "Signature content type is invalid."
    end

    Tempfile.create(["license", ".pdf"]) do |tempfile|
      tempfile.binmode
      doc = generate_document(template_name, **fields)
      doc.write(tempfile, optimize: true)
      tempfile.rewind
      return yield tempfile
    end
  end

  def generate_previews
    TEMPLATE_CONFIG.each_key do |template_name|
      doc = open_template(template_name)
      form = doc.acro_form
      fields = TEMPLATE_CONFIG.fetch(template_name).fetch(:fields)
      fields = fields.values if fields.is_a?(Hash)
      fields.each do |name|
        replace_field(doc, form, name, BLANK_IMAGE)
      end

      doc.catalog[:AcroForm].delete(:Fields)
      doc.catalog.delete(:AcroForm)
      preview_path = preview_path(template_name)
      FileUtils.mkdir_p(preview_path.dirname)
      doc.write(preview_path.to_s, optimize: true).first
    end
  end

  def preview_path(template_name)
    Rails.root.join("public/pdf", "#{template_name}.pdf")
  end

  def public_url(template_name)
    Furi.build(**ENV_URL_OPTIONS[:production], path: "/pdf/#{template_name}.pdf")
  end

  def preview_url(template_name)
    Furi.update(
      "https://mozilla.github.io/pdf.js/web/viewer.html",
      query: {file: public_url(template_name)},
      anchor: "zoom=auto"
    )
  end

  def open_template(template_name)
    HexaPDF::Document.open(Rails.root.join("app/pdf", "#{template_name}.pdf").to_s)
  end

  def formatted_date(value)
    value.strftime("%B %e, %Y")
  end

  def set_field_value(form, name, value)
    field = find_field(form, name)
    if value == true && field.is_a?(HexaPDF::Type::AcroForm::ButtonField)
      allowed_values = field.allowed_values
      value = allowed_values.first if allowed_values.present?
    end
    require 'debug'
    debugger if name == "topmostSubform[0].Page1[0].f_20[0]"
    field.field_value = value
    field.create_appearances(force: true)
  end

  def find_field(form, name)
    form.each_field.find { it.full_field_name.to_s == name } ||
      raise("field not found #{name}")
  end

  def replace_field(doc, form, field, image_io)
    image_io = image_io.to_io if image_io.is_a?(ActionDispatch::Http::UploadedFile)
    input = find_field(form, field)
    widget = input.form_field.each_widget.first
    box = widget[:Rect]
    page = doc.pages.find do |p|
      p[:Annots]&.any? do |element|
        element.form_field&.full_field_name == field
      end
    end || raise("#{field} widget page not found")

    canvas = page.canvas(type: :overlay)
    canvas.image(image_io, at: [box[0], box[1]], width: box[2] - box[0])
    unless input.is_a?(HexaPDF::Type::AcroForm::ButtonField)
      input.field_value = ""
    end
    page[:Annots].each do |element|
      if element.form_field&.full_field_name == field
        page[:Annots].delete(element)
      end
    end
    form.delete_field(field)
  end
end