It very much depends on what kind of pdfs you are trying to read, I use something like this to scrape pdf bank statements:

{pdf_as_text,_} = System.cmd("pdftotext", ~w[-raw bank_statement.pdf -], cd: "/Users/myuser/hello_phoenix/pdfs/")

pdf_as_text
|> String.split("\n") 
|> Stream.map(fn line ->
  regexline = Regex.run(~r/(^\d{1,2}\.\d{1,2}\.\d{4})\s(.+?(?=\s\d{1,2}\.\d{1,2}))\s(\d{1,2}\.\d{1,2})\s(.+?(?=\s))\s(.*)/, line)
  case regexline do
  	[_,full_date, text, date, amount, total] ->
  	  {full_date, text, date, amount, total}
  	_ -> 
  	    nil
  end
end) |> Enum.reject( fn(x) -> x == nil end) |> Enum.to_list |> IO.inspect

basically gets raw text from pdf and then runs a regex on each line, also explore the -layout option for pdftotext, depending on your use case…