calculate frequencies on the fly, fix bug in our char match regex

1 files changed, 24 insertions, 5 deletions

diff --git a/wordle.pl b/wordle.pl
index 7b67fd1..a44026c 100644
--- a/wordle.pl
+++ b/wordle.pl
@@ -33,15 +33,20 @@ my $wlen = length($answer);
 say "total dict size: " . scalar @words;
 my $total = scalar @words;
 
-# filter to word length  
+
 @words = grep( {$_ =~ /^.{$wlen}$/} @words);
 say "total $wlen letter words: " . scalar @words . "\n";
 
+my $frequency = letter_frequency(@words);
+say "calc: $frequency";
+
+# generate word values ahead of time as a cache for sort.
 my %wordvalues;
 say "building wordvalues\n";
 for (@words) {
   $wordvalues{$_} = wordvalue($_);
 }
+
 # In word letters
 my %iwl;
 # letters in correct position 
@@ -178,8 +183,9 @@ while($guess ne $answer ) {
 
   # filter for letters in the answer only
   if(scalar keys %iwl) {
-    my $pattern = '(' . join ('|', sort keys %iwl) . ')';
-    @words = grep { $_ =~ /$pattern/ } @words;
+    for my $c(keys %iwl) {
+      @words = grep { $_ =~ /$c/ } @words;
+    }
   }
   print scalar  @words;
   # sleep 1;
@@ -269,13 +275,15 @@ sub wordvalue {
 
   # my $freq = "etaoinsrhdlucmfywgpbvkxqjz";
   # reverse frequency, later numbers have higher value
-  my $f= "zjqxkvbpgwyfmculdhrsnioate";
+  # general language english:
+  # this should be calculated by length 
+  # my $f = "zjqxkvbpgwyfmculdhrsnioate";
 
   my $v = 1;
   for my $c (split('', $w)) {
     # this is basically bullshit, but sum the value against the index
     # position...
-    $v += (index($f, $c));
+    $v += (index($frequency, $c));
   }
   return $v;
 }
@@ -294,3 +302,14 @@ sub wordvalue_sort  {
     return 1;
   }
 }
+
+sub letter_frequency {
+  my $charfreqhash;
+  for (@_) {
+    for $char (split('', $_)) {
+      $charfreqhash{$char}++;
+    }
+  }
+  my @char_frequency = reverse sort { $charfreqhash{$b} <=> $charfreqhash{$a} } keys %charfreqhash;
+  return join('', @char_frequency);
+}