Rendering of large PDF-Files: flushing of memory possible?

[Robert Neumayr]

Hello!

I have a project that renders pretty big PDF files. In one instance the PDF file is more than 600 pages with 4 images on almost every page. This eventually leads to a NoMemoryError by ruby cause the PDF is held entirely in memory... Is there any way to flush the memory to the file during the creation process? I am tinking something along the lines of "create 100 pages - flush - next 100 - flush - ..."

Thanks for taking the time and for this awesome library!

[William Ross]

Dear Robert,

I would suggest writing out a series of smaller files and concatenating them with pdftk:

https://github.com/webkrak/pdftk

We do something similar to combine generated with uploaded pdf files and it seems reliable. For example:

  def collate_correspondence

    FileUtils.mkdir_p(Rails.root.join("tmp/applications/#{id}"))

    merged_path = Rails.root.join("tmp/applications/#{id}/correspondence_#{id}.pdf")

    source_pathnames = (reviews + references).map(&:local_tempfile_path).map {|f| "'#{f}'"}

    if merge_pdfs(source_pathnames, merged_path)

      self.correspondence_file = File.open(merged_path)

      self.save

    end

  end

  def merge_pdfs(source_paths, destination_path)

    Pdftk.new.merge(source_paths, destination_path)

  end

This is all quite slow because the files are pulled from S3, but the concatenation operation seems quick.

yours,

Will

--
You received this message because you are subscribed to the Google Groups "Prawn" group.
To unsubscribe from this group and stop receiving emails from it, send an email to prawn-ruby+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.