I am working on a module that I hope to release shortly, which constructs
various regexps for solving (loosely) crossword-style problems. As per
the subject line, my working name for it is Regexp::Word; anyone got
suggestions for a better name?
Pod attached.
Hugo
---
=head1 NAME
Regexp::Word - regular expressions for various crossword-style word patterns
=head1 SYNOPSIS
use Regexp::Word;
use re qw/ eval /;
if ("sett" =~ anagram("test")) {
print "sett is an anagram of test\n";
}
if ("true" =~ subset("turbulent")) {
print "true is a subset of turbulent\n";
}
if ("true" =~ subset_anagram("suture")) {
print "true is a subset of an anagram of suture\n";
}
# etc
=head1 DESCRIPTION
Given an input pattern of a recognisable form, these functions will
return an anchored regular expression that will match words that
satisfy one or other variant of the pattern, as described for the
individual functions below.
Recognisable patterns look like normal regular expressions, but are
constrained to consist only of normal characters, character classes
and quantifiers, so that for example 'ab[c-m]n?o*p+q{3,5}[^r-z]' is
accepted. Parentheses and anchors are not recognised.
Functions that have 'anagram' in the name will usually have embedded
eval groups in the returned regular expression, so you will usually
need to C<use re 'eval';> before applying these.
The results returned are simple strings, so that you can apply
modifiers to them when you use them:
# test case-insensitive anagram
my $re = anagram("test");
print "ok" if "SETT" =~ /$re/i;
NOTE that a single regular expression may not be the best way of
solving some of these problems in the general case: for example,
if you want to check whether two words are anagrams of each other
it may be more efficient to sort the characters and compare the
resulting ordered strings:
my($word1, $word2) = ("sett", "test");
my $order1 = join '', sort split //, $word1;
my $order2 = join '', sort split //, $word2;
print "ok" if $order1 eq $order2;
.. but non-regexp solutions to these problems would belong in a
different module.
=head2 EXPORT
All the functions described below are exported by default.
=head2 FUNCTIONS
=over 4
=item anagram ( pattern )
Returns a regular expression that matches any word that is an anagram of
the components of the input pattern. May require C<use re 'eval'>.
=for testing
use re 'eval';
my $re = anagram('ab[cd]{1,2}[^ef].');
print "xycacb" =~ $re ? "ok 1" : "not ok 1";
print "eecacb" !~ $re ? "ok 2" : "not ok 2";
print "cdbaaa" =~ $re ? "ok 3" : "not ok 3";
=item subset ( pattern )
Returns a regular expression that matches any word that is a subset of
the input pattern.
=for testing
my $re = subset('ab[cd]{1,2}[^ef].');
print "acccc" =~ $re ? "ok 1" : "not ok 1";
print "accccc" !~ $re ? "ok 2" : "not ok 2";
print "zz" =~ $re ? "ok 3" : "not ok 3";
=item subset_anagram ( pattern )
Returns a regular expression that matches any word that is a subset of
any anagram of the input pattern. May require C<use re 'eval'>.
=for testing
use re 'eval';
my $re = subset_anagram('ab[cd]{1,2}[^ef].');
print "cccca" =~ $re ? "ok 1" : "not ok 1";
print "ccccca" !~ $re ? "ok 2" : "not ok 2";
print "zzb" =~ $re ? "ok 3" : "not ok 3";
=item superset ( pattern )
Returns a regular expression that matches any word that is a superset
of the input pattern.
=for testing
my $re = superset('ab[cd]{1,2}[^ef].');
print "accccc" !~ $re ? "ok 1" : "not ok 1";
print "abcccc" =~ $re ? "ok 2" : "not ok 2";
print "zzzzabcdz" =~ $re ? "ok 3" : "not ok 3";
=item superset_anagram ( pattern )
Returns a regular expression that matches any word that is a superset
of any anagram of the input pattern. May require C<use re 'eval'>.
=for testing
my $re = superset_anagram('ab[cd]{1,2}[^ef].');
print "ccccca" !~ $re ? "ok 1" : "not ok 1";
print "ccccba" =~ $re ? "ok 2" : "not ok 2";
print "zdcba" =~ $re ? "ok 3" : "not ok 3";
=item insert ( count, pattern )
Returns a regular expression that matches any word that inserts
exactly I<count> characters (not necessarily consecutive) into the
input pattern.
=for testing
my $re = insert(2, 'ab[cd]{1,2}[^ef].');
print "abcdezz" =~ $re ? "ok 1" : "not ok 1";
print "abcddezz" =~ $re ? "ok 2" : "not ok 2";
print "abcdddezz" !~ $re ? "ok 3" : "not ok 3";
print "abcezz" !~ $re ? "ok 4" : "not ok 4";
=item insert_anagram ( count, pattern )
Returns a regular expression that matches any word that inserts
exactly I<count> characters (not necessarily consecutive) into
an anagram of the input pattern. May require C<use re 'eval'>.
=for testing
my $re = insert(2, 'ab[cd]{1,2}[^ef].');
print "zzedcba" =~ $re ? "ok 1" : "not ok 1";
print "zzeddcba" =~ $re ? "ok 2" : "not ok 2";
print "zzedddcba" !~ $re ? "ok 3" : "not ok 3";
print "zzebca" !~ $re ? "ok 4" : "not ok 4";
=item delete ( count, pattern )
Returns a regular expression that matches any word that deletes
exactly I<count> characters (not necessarily consecutive) from the
input pattern.
=for testing
my $re = delete(2, 'ab[cd]{1,2}[^ef].');
print "ab" !~ $re ? "ok 1" : "not ok 1";
print "abc" =~ $re ? "ok 2" : "not ok 2";
print "abcd" =~ $re ? "ok 3" : "not ok 3";
print "abcde" !~ $re ? "ok 4" : "not ok 4";
=item delete_anagram ( count, pattern )
Returns a regular expression that matches any word that deletes
exactly I<count> characters (not necessarily consecutive) from
an anagram of the input pattern. May require C<use re 'eval'>.
=for testing
my $re = delete(2, 'ab[cd]{1,2}[^ef].');
print "ba" !~ $re ? "ok 1" : "not ok 1";
print "cba" =~ $re ? "ok 2" : "not ok 2";
print "dcba" =~ $re ? "ok 3" : "not ok 3";
print "edcba" !~ $re ? "ok 4" : "not ok 4";
=item latent ( pattern )
Returns a regular expression that matches any word that, after
removing all occurrences of one character, matches the input
pattern.
=for testing
my $re = latent('ab[cd]{1,2}[^ef].');
print "zazbzczdzez" =~ $re ? "ok 1" : "not ok 1";
print "abccdgh" =~ $re ? "ok 2" : "not ok 2";
print "abdddgh" !~ $re ? "ok 3" : "not ok 3";
=item latent_anagram ( pattern )
Returns a regular expression that matches any word that, after
removing all occurrences of one character, matches an anagram
of the input pattern. May require C<use re 'eval'>.
=for testing
my $re = latent_anagram('ab[cd]{1,2}[^ef].');
print "zezdzczbzaz" =~ $re ? "ok 1" : "not ok 1";
print "hgdccba" =~ $re ? "ok 2" : "not ok 2";
print "hgdddba" !~ $re ? "ok 3" : "not ok 3";
=head1 AUTHOR
Hugo van der Sanden, E<lt>[EMAIL PROTECTED]<gt>
=head1 SEE ALSO
L<perlre>, L<Regexp::Common>.
=cut
