Haskell bzlib-conduit/zlib-conduit example

Question

Let's assume we create the file a.txt.gz as follows:

$ echo "foobar" > a.txt
$ gzip a.txt

I intend to use zlib-conduit in order to emulate zcat in Haskell.

I'm looking for a simple example that can also be applied to bzlib-conduit.

Note: This question was answered immediately in a Q&A-Style. Therefore it intentionally does not show any research effort.

Aspergillosis

Answer 1

If you intend to work with conduits, I highly recommend to read the excellent Conduit overview by Michael Snoyman and the FP Complete tutorial on this topic first.

I've turned on my -vv flag to make this suitable for Haskell beginners (like myself).

You need three things:

• A file source
• A zlib decompression filter
• A stdout sink

Let's start off with this simple file copy example:

import Data.Conduit (runResourceT, ($$))
import qualified Data.Conduit.Binary as CB
import Data.Conduit.Zlib

main = do
    runResourceT $ CB.sourceFile "input.txt" $$ CB.sinkFile "output.txt"

What do we need to modify here?

• The input filename is not a.txt.gz
• zlib decompressor is missing
• We want to output to stdout, not to output.txt

Indeed the decompress documentation contains an example of how to decompress.

Note that you can't use decompress for gzip-generated files. decompress decompresses .Z files generated by the old compress UNIX program.

After modifying the above example we get:

import Data.Conduit (runResourceT, ($$), ($=))
import qualified Data.Conduit.Binary as CB
import Data.Conduit.Zlib
import System.IO

main = do
    runResourceT $ CB.sourceFile "a.txt.gz" $= ungzip $$ CB.sinkHandle stdout

The difference when using bzlib-conduit is minimal:

import Data.Conduit (runResourceT, ($$), ($=))
import qualified Data.Conduit.Binary as CB
import Data.Conduit.BZlib
import System.IO

main = do
    runResourceT $ CB.sourceFile "a.txt.bz2" $= bunzip2 $$ CB.sinkHandle stdout

Aspergillosis