How to convert PDF to Excel or CSV in Rails 4

Question

I have searched a lot. I have no choice unless asking this here. Do you guys know an online convertor which has API or Gem/s that can convert PDF to Excel or CSV file?

Answer 1

Check Tabula-Extractor project and also check how it is used in projects like NYPD Moving Summonses Parser and CompStat criminal complaints parser.

Answer 2

Ok, After lots of research I couldn't find an API or even a proper software that does it. Here how I did it.

I first extract the Table out of the PDF into the Table with this API pdftables. It is cheap.

Then I convert the HTML table to CSV.

(This is not ideal but it works)

Here is the code:

require 'httmultiparty'
class PageTextReceiver
  include HTTMultiParty
  base_uri 'http://localhost:3000'

  def run
    response = PageTextReceiver.post('https://pdftables.com/api?key=myapikey', :query => { f: File.new("/path/to/pdf/uploaded_pdf.pdf", "r") })

    File.open('/path/to/save/as/html/response.html', 'w') do |f|
      f.puts response
    end
  end

  def convert
    f = File.open("/path/to/saved/html/response.html")
    doc = Nokogiri::HTML(f)
    csv = CSV.open("path/to/csv/t.csv", 'w',{:col_sep => ",", :quote_char => '\'', :force_quotes => true})
    doc.xpath('//table/tr').each do |row|
      tarray = []
      row.xpath('td').each do |cell|
        tarray << cell.text
      end
      csv << tarray
    end
    csv.close
  end
end

Now Run it like this:

#> page = PageTextReceiver.new
#> page.run
#> page.convert

It is not refactored. Just proof of concept. You need to consider performance.

I might use the gem Sidekiq to run it in background and move the result to the main thread.

Answer 3

Ryan Bates covers csv exports in his rails casts > http://railscasts.com/episodes/362-exporting-csv-and-excel this might give you some pointers.

Edit: as you now mention you need the raw data from an uploaded PDF, you could use JavaScript to read the PDF file and the populate the data into Ryan Bates' export method. Reading PDF's was covered excellently in the following question:

extract text from pdf in Javascript

I would imagine the flow would be something like:

PDF new action
    user uploads PDF 

PDF show action
    PDF is displayed
    JavaScript reads PDF
    JavaScript populates Ryan's raw data
    Raw data is exported with PDF data included