A Perl or Gawk script that returns a Keyword, the word before, and the word after?

I need a simple script to run in Windows that searches large xml files for a keyword, then returns the word before it, the keyword, and the word after.

For example: "How can I extract keywords in context" I want: "extract keywords in"

I'm a novice with enough knowledge to return each line with the Keyword, and the lines before and after, but I'm stumped on getting the individual words I 开发者_StackOverflow中文版need out.

Anyone have any clever ideas?

Here's one way:

#!/usr/bin/perl

use 5.12.0;
my $keyword = 'keywords';

while (<DATA>)
{
    say for /\b(\S+\s+\b\Q$keyword\E[[:punct:]]*\s+\S+)\b/g;
}

__END__
How can I extract keywords in context, even if there are many keywords to
extract? So many keywords, no idea how to deal with them.

grep -o is enough:

grep -Po '(\S+\s)?keywords(\s\S+)?' << END
How can I extract keywords in context
How can I extract keywords
keywords in context
END

returns

extract keywords in
extract keywords
keywords in