# Converts peptide sequences in fasta format to single strings.
# Strips comments and headings, and then prints each sequence, one to a line.
# We expect IUPAC peptide codes (upper or lower case).
# Two other character codes are common.
# * is often used for stop codons.
# X is often used to mean no other aa code.
# More search programs can handle X than *.

my $seq;

sub loadOneSeq( $ ) {
  ( my $identifier ) = @_;
# $seq =~ s/\*$//; # strip any trailing stop.
# $seq =~ s/X$//; # strip any trailing stop.
  print "$identifier\t";
  print $seq;
  print "\n";
  $seq = nil;
}

my $seqid;
while ( <STDIN> ) {
  chomp;
  if ( /^>(\S+)/ ) {
    if ( $seq ) { loadOneSeq( $seqid ) }
    $seqid = $1;
  }
  elsif ( /^;/ ) {
    # Skip comments.
  }
  # Note that it is slightly faster to check for an illegal character
  # than for a string of legal ones.
# elsif ( /[^AC-IK-NP-TVWY]/i ) { # Strict IUPAC codes only.
# elsif ( /[^A-IK-NP-TVWYZ]/i ) { # Allow IUPAC "ambiguity" codes B and Z.
  elsif ( /[^A-IK-NP-TV-Z*]/i ) { # Allow ambiguity codes and non-IUPAC codes X and *.
# elsif ( /[^A-IK-NP-Z*]/i ) { # Allow U also which is sometimes used for seleneomethionine.
# elsif ( /[^A-IK-Z*]/i ) { # Allow O also which is sometimes used for ornithine.
    print STDERR "ERROR in sequence $seqid\n";
    print STDERR "$_\n";
    # Should also null out seq and skip to next seq header.
    # Or should we just quit?
    die;
  }
  else {
    # Note this also matches blank lines. Is that OK?
#   s/\*/X/g; # Convert stops?
    $seq .= $_;
  }
}