Class: Libera::Parser

Inherits:
Object
  • Object
show all
Defined in:
lib/libera.rb

Instance Method Summary collapse

Instance Method Details

#generate_tei(page_list) ⇒ Object 



71
72
73
74
75
76
77
78
79
80
81
82
 
# File 'lib/libera.rb', line 71

def generate_tei(page_list)
  tei_path = "#{Libera.configuration.working_dir}/tei.xml"
  tei_xml = Libera::Tei.new
  
  page_list.each do |k, v|
    tei_xml.add_page_break(k)
    tei_xml.add_anon_block(v)
  end
  
  # write out xml to file
  IO.write(tei_path, tei_xml.to_xml)
end

#mk_working_dir ⇒ Object 



84
85
86
87
 
# File 'lib/libera.rb', line 84

def mk_working_dir
  # Check if working dir exists - If not, make it
  FileUtils.mkdir_p("#{Libera.configuration.working_dir}") unless File.exists? "#{Libera.configuration.working_dir}"
end

#parse_image(image_path, i) ⇒ Object 



65
66
67
68
69
 
# File 'lib/libera.rb', line 65

def parse_image(image_path, i)
  file_path = "#{Libera.configuration.working_dir}/pdf-page-#{i}"
  `tesseract #{image_path} #{file_path} >> /dev/null 2>&1`
  return File.read(file_path + ".txt")
end