CPAN has several modules that can guess a files type, like File::MimeInfo and File::MimeInfo::Magic. However most of those modules can't give you much more information that the filetype. Therefor I prefer using File::Extractor, which can collect detailed metadata for lots of filetypes like the resolution for image files, artist, title, etc from audio files and much more. The only downside is that it requires an external c library: libextractor.

Also note that I might be quite a bit biased as I'm the author of File::Extractor.

Thanks, I need information more than MIME type, so File::Type and File::MimeInfo seem not OK. I installed your module through cpan shell, but it doesnot pass the compilation.

cpan> install ExtUtils::PkgConfig
...
Can't locate ExtUtils/PkgConfig.pm in @INC ....
...
cpan> install ExtUtils::PkgConfig
...snip ok info...
cpan> install File::Extractor    
Running install for module File::Extractor
Running make for F/FL/FLORA/File-Extractor-0.03.tar.gz
  Is already unwrapped into directory /root/.cpan/build/File-Extractor
+-0.03
  Makefile.PL returned status 65280
Running make test
  Make had some problems, maybe interrupted? Won't test
Running make install
  Make had some problems, maybe interrupted? Won't install
[download]

I installed libextractor(http://gnunet.org/libextractor/download/Extractor-0.5.tar.gz) by following the instruction(i.e. running "python setup.py install", everything looks ok), but I still get the same error information. ? ? any idea? many thanks...

lihao

CPANs error message is caused the cpan shell caching some state. Restarting the shell and invoking trying to install the module again should fix that.

Also it looks like you installed the python bindings for libextractor but not libextractor itself. This is most likely what you want.

I haven't used File::Type myself, but it looks like what you want. You can also use backticks to call file, if you want.

Update: Actually, I have used File::Type, but it was so long ago, I forgot. It worked fine at the time.

Depending on your task you (is it getting image size by chance?) you may find a module like Image::Size or Image::Info does the trick. If you want to see how its done here it is in Perl. Essentially just read the header and see what it matches, then unpack the size data.

#!/usr/bin/perl -w
use strict;

sub image_size {
  return unless $_[0];
    my ($width, $height, $sig);
    if ( $_[0] =~ m/^GIF8..(....)/s ) {
        $sig = 'GIF'; 
        ($width, $height) = unpack( "SS", $1 );
    }
    elsif ( $_[0] =~ m/^^\xFF\xD8.{4}JFIF/s  ) {
        $sig = 'JPEG';
        ($height,$width) = unpack( "nn", $1 ) if $_[0] =~ /\xFF\xC0...
+(....)/s;
    }
    elsif ( $_[0] =~ /^\x89PNG\x0d\x0a\x1a\x0a/  ) {
        $sig = 'PNG';
        ($width, $height) = unpack( "NN", $1 ) if $_[0] =~ /IHDR(.{8})
+/s;
    }
    elsif ( $_[0] =~ /BM.{16}(.{8})/s ) {
        $sig = 'BMP';
        ($width, $height) = unpack( "LL", $1);
    }
  return $width, $height, $sig;
}

for my $img ( qw( c:/sample.bmp c:/sample.jpg c:/sample.gif c:/sample.
+png) ) {
    my $data = get_file($img);
    my @res = image_size($data);
    print "$img $res[2] width $res[0] height $res[1]\n";
}

sub get_file {
    open my $fh $_[0] or die $!;
    binmode $fh;
    local $/;
    <$fh>
}

__END__
c:/sample.bmp BMP width 512 height 384
c:/sample.jpg JPEG width 100 height 149
c:/sample.gif GIF width 110 height 58
c:/sample.png PNG width 256 height 192
[download]

Hi, tachyon-II:

You are right, I have used GD::Image to grab size-infomation of about 99% image files. only those with type "PC bitmap data, Windows 3.x format" not available through GD::Image. but I can grab this missing information from running Linux 'file' command anyway. Just want to know if there is a Perl module which can handle this, instead of use code like below(untested):

  my $imginfo = qx[ file $myimg ];
  my (width, height)= ($imginfo =~ m{(\d+) x (\d+) x \d+});
[download]

thanks.

lihao

Does the code above not work? Is Image::Size not a module that does this?