Haskell library that extracts text from PDF's

erik flister via haskell-pdftotext

2 years ago
2 years ago


The pdftotext package provides functions for extraction of plain text from PDF documents. It uses C++ library Poppler, which is required to be installed in the system. Output of Haskell pdftotext library is identical to output of Poppler's tool pdftotext.


import qualified Data.Text.IO as T
import Pdftotext

main :: IO ()
main = do
  Just pdf <- openFile "path/to/file.pdf"
  T.putStrLn $ pdftotext Physical pdf


pdftotext comes with executable program pdftotext.hs which can print text extracted from PDF and basic information from the document.

$> pdftotext.hs info test/simple.pdf
File      : test/simple.pdf
Pages     : 4
  Title   : Simple document for testing
  Author  : G. Eyaeb
  Subject : Testing
  Creator : pdflatex
  Producer: LaTeX with hyperref
  Keywords: haskell,pdf
$> pdftotext.hs text --pages 1,4 test/simple.pdf
Simple document for testing

                  deserve neither
liberty nor safety.

See help for more information:

$> pdftotext.hs --help
$> pdftotext.hs text --help
$> pdftotext.hs info --help


The library uses poppler via FFI, therefore internally all functions are of type IO. However, their non-IO variants (using unsafePerformIO) should be safe to use. Module Pdftotext.Internal exposes all IO-typed functions.


Project is hosted at https://sr.ht/~geyaeb/haskell-pdftotext/ . The homepage provides links to Mercurial repository, mailing list and ticket tracker.

Patches, suggestions, questions and general discussions can be send to the mailing list. Detailed information about sending patches by email can be found at https://man.sr.ht/hg.sr.ht/email.md.