Converting doc to txt without WIN32::OLE

[mrguy123]

Hi Monks
I have about 30 word documents that I need to parse, and would like to transform them to text. I know the best solution is to use Win32::OLE but since I am working in Unix this is a bit of a problem. Thre are solutions but for now I am looking for other directions.
What I really need is a super simple Perl module that gets text from word documents and dumps it in a string. I tried using Text::Extract::Word but it only worked on one doc out of 5.
So, any ideas?

There’s no point in being grown up if you can’t act a little childish sometimes

P.S: I need this solution primarily for doc and not docx documents

UPDATE: Installed the latest version of Text::Extract::Word (it appears that I was working with the legacy code for some reason). It now works! Thanks for all your help
Guy

[Corion]

If you don't want to use Win32::OLE, I would look at OpenOffice / LibreOffice for converting the doc file to something else. Searching for MSWord also shows some modules, but there is nothing like Spreadsheet::ParseExcel for Word.

Simply scanning the file for strings will likely not work well if the strings are stored as UTF-16 or wide characters.

[mrguy123]

Hi Corion, thanks for the tip
I will try some of the modules you suggested, maybe I will get lucky :)
Open Office is also a good solution, but I need to see if there is a batch option for conversion before I go there

[ww]

You don't really need a "batch option" with OO/LO; just write your script, allowing it to accept (@ARGV or a list in a file) multiple docs which names/fullpaths it then feeds to OO.

UPDATE: PS. You can write offsite links here as "[http://www.foo/bar.htm]" and, for PM nodes, as "[id://977242]." In the latter case, that's especially preferred, so you don't inadvertently logout a reader following your node on say, PM.org by a link to PM.com.

See also Markup in the Monastery for less expensive (keystroke-wise) ways a marking paras. </update>

[Khen1950fx]

I did some work on Text::Extract::Word last September. See: Re: Problem in Text::Extract::Word.

I put two super simple scripts together for you that put the files into a string. For the test.docs, I used the test.docs from Text::Extract::Word/t directory:

#single file
#!/usr/bin/perl -l

use strict qw/refs/;
use warnings FATAL => 'all';
use Text::Extract::Word;

binmode STDOUT, ':encoding(UTF-8)';

my $file = '/root/Desktop/xls/test1.doc';
my $extractor = Text::Extract::Word->new($file);
my $string = $extractor->get_text;
print "$string";

close STDOUT;

#!/usr/bin/perl 

BEGIN {
    $| = 1;
}

use autodie;
use strict qw/refs/;
use Text::Extract::Word;
use warnings FATAL => 'all';

binmode STDOUT, ':encoding(UTF-8)';

my(@data) = qw(
    test1.doc
    test2.doc
    test3.doc
    test4.doc
    test5.doc
    test6.doc
);

foreach my $data(@data) {
    my $file = Text::Extract::Word->new($data);
    my $str  = $file->get_text;
    print "$str===>File done<===\n\n";
    sleep 2;
}

close STDOUT;

[mrguy123]

Hi, thanks for the scripts
Problem is that when I run them I get:

Can't locate object method "new" via package "Text::Extract::Word" at 
+monks.pl line 11.

I guess that there was something wrong with the installation or that the module isn't stable- I will try to fix it. The legacy interface works but not very well (some docs just aren't parsed)

# legacy interface
use Text::Extract::Word qw(get_all_text);
my $text = get_all_text("test1.doc");

[mrguy123]

OK, so it seems I am working with the legacy code for some reason. I am now installing the new version...hope it works better