wicked_pdf generation in a resque queue?

[mileszs]

wicked_pdf generation in a resque queue?

[kandadaboggu]

I addressed the issue by monkey patching the pdf_from_string method as follows:

class WickedPdf
  def pdf_from_string(string, options={})
    command = "#{@exe_path} #{parse_options(options)} -q - - " # -q for no errors on stdout
    p "*"*15 + command + "*"*15 unless defined?(Rails) and Rails.env != 'development'
    pdf, err = Open3.popen3(command) do |stdin, stdout, stderr|
      stdin.binmode
      stdout.binmode
      stderr.binmode
      stdin.write(string)
      stdin.close
      if block_given?
        success=false
        while(line=stdout.read(1024))
          success = true
          yield line
        end        
        [ ( success ? "SUCCESS" : ""), stderr.read]
      else
        [stdout.read, stderr.read]
      end      
    end
    raise "PDF could not be generated!" if pdf and pdf.rstrip.length == 0
    pdf
  rescue Exception => e
    raise "Failed to execute:\n#{command}\nError: #{e}"
  end
end

This allows me to process the output stream 1024 bytes at a time.

File.open(out_file, 'wb') do |file| 
  WickedPdf.new.pdf_from_string("HELLO") {|line| file << line}
end

This might be an good enhancement to add to the gem.

[unixmonkey]

This is totally doable. In fact, I used to do the same thing (generate pdfs from the model via delayed_job). It's a bit tricky, though. Let me know if something like this works for you:

class Pdfmaker
  def self.generate_pdf(org,items)
     # create instance vars that are needed in the view
     @org = org
     @items = items

     # setup paths
     outfile_path = Rails.root.join('public','uploads','pdfs')
     view_path   = Rails.root.join('app','views','pdfs')

     # parse erb templates
     body = File.read(view_path.join('show.pdf.erb')
     body_render = Erb.new(body).render(binding)
     footer = File.read(view_path.join('_footer.pdf.erb')
     footer_render = Erb.new(body).render(binding)

     # pre-save footer to file for inclusion
     tf = WickedPdfTempfile.new('wicked_pdf_footer_pdf.html')
     tf.write(footer_render)
     tf.flush

     # run through wicked_pdf
     pdf = WickedPdf.new.pdf_from_string(body_render,
       :footer   => { :html => { :url => "file://#{tf.path}" } }, # this line includes the footer we just made
       :margin => { :bottom => 15 } # other options work fine here
     )

     # save final pdf to filesystem
     File.open(outfile_path.join("#{org.id}.pdf"), 'w') {|f| f << pdf }
  end
end

[dgdosen]

I'm not that well versed on calling the rendering engine for rails - what do you declare Erb as?

[unixmonkey]

Erb is already declared. It is the .erb rendering library included in Ruby.
Erb.new("hello <%= 'world' %>!").render(binding) #=> "Hello world!"

[dgdosen]

Trying to get this to work...
Is the code above for Rails 3.1?

When I try to declare an object of type Erb, I get an uninitialized constant error on Erb. From looking at the source code for rails, I can declare an object as ERB, but that doesn't have a render method. For that I'd need something like an instance of the ActionView...

Any guidance? The only thing I'm blocking on is actually rendering the content from the worker queue

[unixmonkey]

Whoops. I use Haml, so my bits were different and I swapped in ERB from memory.

What you really need is ERB.new(body).result(binding)

[dgdosen]

Almost there! -

I can generate PDF's with data, and attach them to an email, but only if I'm not using the wicked_pdf_stylesheet_link_tag in my template. If I try to use it, the template throws an error. undefined method 'wicked_pdf_stylesheet_link_tag' To be clear, I don't get this issue when generating a PDF for consumption in the browser.

Do you know what I'd need to get me some css when generating the doc from the scope of the worker queue?

[unixmonkey]

ActionView Helpers like wicked_pdf_stylesheet_link_tag or even link_to and number_to_currency aren't available in that context. You could instantiate actionview and call render on that as shown here:
#34

Or, you could just avoid using helpers and write out the html you need by hand like so:
<link type='text/css' rel='stylesheet' src='file://<%= Rails.root.join('public','stylesheets','pdf.css').to_s %>' />

[dgdosen]

FYI - I couldn't get this to work by manually loading the css file, but I was able to get everything to work by using the method where you create an instance of the ActionView (per #34). When you take that approach, the css file(s) load (using wicked_pdf_stylesheet_link_tag) and you have access to all the view helpers.

[kandadaboggu]

I am trying to generate the PDF and save it to a file. I am considering using the WickedPdf.new.pdf_from_string technique. I am concerned that this will simply bloat the memory when the PDF generated is large. Is it possible to write the PDF results using some sort of PDF result enumerator

E.g:

File.open("out.pdf", 'wb')  do |file| 
  WickedPdf.new.pdf_from_string(html_string) do | pdf|
    file << pdf
  end
end

[danblaker]

I added a wiki page https://github.com/mileszs/wicked_pdf/wiki/Background-PDF-creation-via-delayed_job-gem to show a simple example of using delayed_job to generate PDFs in the background.