package Chemistry::File::SMILES;

$VERSION = "0.47";
# $Id: SMILES.pm,v 1.16 2009/05/10 20:31:08 itubert Exp $

use 5.006;
use strict;
use warnings;
no warnings 'recursion';
use base "Chemistry::File";
use Chemistry::Mol;
use Chemistry::Bond::Find 'assign_bond_orders';
use List::Util 'first';
use Carp;


=head1 NAME

Chemistry::File::SMILES - SMILES linear notation parser/writer

=head1 SYNOPSYS

    #!/usr/bin/perl
    use Chemistry::File::SMILES;

    # parse a SMILES string
    my $s = 'C1CC1(=O)[O-]';
    my $mol = Chemistry::Mol->parse($s, format => 'smiles');

    # print a SMILES string
    print $mol->print(format => 'smiles');

    # print a unique (canonical) SMILES string
    print $mol->print(format => 'smiles', unique => 1);

    # parse a SMILES file
    my @mols = Chemistry::Mol->read("file.smi", format => 'smiles');

    # write a multiline SMILES file
    Chemistry::Mol->write("file.smi", mols => \@mols);


=head1 DESCRIPTION

This module parses a SMILES (Simplified Molecular Input Line Entry
Specification) string. This is a File I/O driver for the PerlMol project.
L<http://www.perlmol.org/>. It registers the 'smiles' format with
Chemistry::Mol.

This parser interprets anything after whitespace as the molecule's name;
for example, when the following SMILES string is parsed, $mol->name will be
set to "Methyl chloride":

    CCl	 Methyl chloride

The name is not included by default on output. However, if the C<name> option
is defined, the name will be included after the SMILES string, separated by a
tab.

    print $mol->print(format => 'smiles', name => 1);

=head2 Multiline SMILES and SMILES files

A file or string can contain multiple molecules, one per line.

    CCl	 Methyl chloride
    CO	 Methanol

Files with the extension '.smi' are assumed to have this format.

=head2 Atom Mapping Numbers

As an extension for reaction processing, SMILES strings may have atom mapping
numbers, which are introduced after a colon in a bracketed atom. For example,
[C:1]. The mapping number need not be unique. This module reads the mapping
numbers and stores them as the name of the atom ($atom->name).

On output, atom names are not included by default. See the C<number> and
C<auto_number> options below for ways of including them.

head1 OPTIONS

The following options are supported in addition to the options mentioned for
L<Chemistry::File>, such as C<mol_class>, C<format>, and C<fatal>.

=over

=item aromatic

On output, detect aromatic atoms and bonds by means of the Chemistry::Ring
module, and represent the organic aromatic atoms with lowercase symbols.

=item unique

When used on output, canonicalize the structure if it hasn't been canonicalized
already and generate a unique SMILES string. This option implies "aromatic".

=item number

For atoms that have a defined name, print the name as the "atom number". For
example, if an ethanol molecule has the name "42" for the oxygen atom and the
other atoms have undefined names, the output would be:

    CC[OH:42]

=item auto_number

When used on output, number all the atoms explicitly and sequentially. The
output for ethanol would look something like this:

    [CH3:1][CH2:2][OH:3]

=item name

Include the molecule name on output, as described in the previous section.

=item kekulize

When used on input, assign single or double bond orders to "aromatic" or
otherwise unspecified bonds (i.e., generate the Kekule structure). If false,
the bond orders will remain single. This option is true by default. This uses
C<assign_bond_orders> from the L<Chemistry::Bond::Find> module.

=back

=cut

# INITIALIZATION
Chemistry::Mol->register_format('smiles');
my $Smiles_parser = __PACKAGE__->new_parser;

#=begin comment
#
#=over
#
#=cut

sub file_is {
    my $self = shift;
    $self->name_is(@_);
}

sub name_is {
    my ($self, $name) = @_;
    $name =~ /\.smi/;
}

sub slurp_mol { 
    my ($self, $fh) = @_;
    scalar <$fh>;
}

sub read_mol {
    my ($self, $fh, %opts) = @_;
    %opts = (kekulize => 1, %opts);
    my $mol_class = $opts{mol_class} || "Chemistry::Mol";

    my $line = <$fh>;
    return unless defined $line;
    $line =~ tr/\r\n//d; 
    my ($smiles, $name) = split " ", $line, 2;

    my $mol = $mol_class->new;
    unless ($Smiles_parser->parse($smiles, $mol, \%opts)) {
        warn "error parsing SMILES line '$line'\n";
        $mol = $mol_class->new;
    }
    $mol->name($name);
    $self->add_implicit_hydrogens($mol);
    if ($opts{kekulize}) {
        assign_bond_orders($mol, method => "itub", use_coords => 0, 
            scratch => 0, charges => 0);
    }
    $mol;
}


### The contents of the original Chemistry::Smiles module start below

my $Symbol = qr/
    s|p|o|n|c|b|Zr|Zn|Yb|Y|Xe|W|V|U|Tm|Tl|Ti|Th|
    Te|Tc|Tb|Ta|Sr|Sn|Sm|Si|Sg|Se|Sc|Sb|S|Ru|Rn|Rh|Rf|Re|Rb|Ra|
    Pu|Pt|Pr|Po|Pm|Pd|Pb|Pa|P|Os|O|Np|No|Ni|Ne|Nd|Nb|Na|N|Mt|Mt|
    Mo|Mn|Mg|Md|Lu|Lr|Li|La|Kr|K|Ir|In|I|Hs|Hs|Ho|Hg|Hf|He|H|Ge|
    Gd|Ga|Fr|Fm|Fe|F|Eu|Es|Er|Dy|Ds|Db|Cu|Cs|Cr|Co|Cm|Cl|Cf|Ce|
    Cd|Ca|C|Br|Bk|Bi|Bh|Be|Ba|B|Au|At|As|Ar|Am|Al|Ag|Ac|\*|R|X
/x; # Order is reverse alphabetical to ensure longest match

my $Simple_symbol = qr/Br|Cl|B|C|N|O|P|S|F|I|H|s|p|o|n|c|b/;

my $Bond = qr/(?:[-=#:.\/\\])?/; 
my $Simple_atom = qr/($Simple_symbol)/;   #3
my $Complex_atom = qr/
    (?:
        \[                          #begin atom
        (\d*)                       #4 isotope
        ($Symbol)                   #5 symbol
        (\@{0,2})                   #6 chirality
        (?:(H\d*))?                 #7 H-count
        (\+{2,}|-{2,}|\+\d*|-\d*)?  #8 charge
        (?::(\d+))?                 #9 name
        \]                          #end atom 
    )
/x;

my $Digits = qr/(?:($Bond)(?:\d|%\d\d))*/; 
my $Chain = qr/
    \G(                                     #1
        (?: 
            ($Bond)                         #2
            (?:$Simple_atom|$Complex_atom)  #3-9
            ($Digits)                       #10
        ) 
        |\( 
        |\)
        |.+
    )
/x;

my $digits_re = qr/($Bond)(\%\d\d|\d)/;

my %type_to_order = (
    '-' => 1,
    '=' => 2,
    '#' => 3,
    '/' => 1,
    '\\' => 1,
    '' => 1, # not strictly true
    '.' => 0,
);

my %ORGANIC_ELEMS = (
    Br => 1, Cl => 1, B => 3, C => 4, N => 3, O => 2, P => 3, S => 2, 
    F => 1, I => 1, s => 1, p => 1, o => 1, n => 1, c => 1, b => 1,
);

#=item Chemistry::Smiles->new([add_atom => \&sub1, add_bond => \&sub2])
#
#Create a SMILES parser. If the add_atom and add_bond subroutine references
#are given, they will be called whenever an atom or a bond needs to be added
#to the molecule. If they are not specified, default methods, which
#create a Chemistry::Mol object, will be used.
#
#=cut

sub new_parser {
    my $class = shift;
    my %opts = @_;
    my $self = bless {
        add_atom => $opts{add_atom} || \&add_atom,
        add_bond => $opts{add_bond} || \&add_bond,
    }, $class;
}

#=item $obj->parse($string, $mol)
#
#Parse a Smiles $string. $mol is a "molecule state object". It can be anything;
#the parser doesn't do anything with it except sending it as the first parameter
#to the callback functions. If callback functions were not provided when
#constructing the parser object, $mol must be a Chemistry::Mol object, because
#that's what the default callback functions require.
#
#=cut

sub parse {
    my ($self, $s, $mol, $opts) = @_;
    $self->{stack} = [ undef ];
    $self->{digits} = {};

    eval {
        while ($s =~ /$Chain/g) {
            #my @a = ($1, $2, $3, $4, $5, $6, $7, $8);
            #print Dumper(\@a);
            my ($all, $bnd, $sym, $iso, $sym2, $chir, $hcnt, $chg, $name, $dig) 
                = ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10);
            if ($all eq '(') {
                $self->start_branch();
            } elsif ($all eq ')') {
                $self->end_branch();
            } elsif ($sym) { # Simple atom
                no warnings;
                my @digs = parse_digits($dig);
                $self->atom($mol, $bnd, '', $sym, '', undef, '', \@digs);
            } elsif ($sym2) { # Complex atom
                no warnings;
                my @digs = parse_digits($dig);
                if ($hcnt eq 'H') { 
                    $hcnt = 1;
                } else {
                    $hcnt =~ s/H//;
                }
                unless ($chg =~ /\d/) {
                    $chg = ($chg =~ /-/) ? -length($chg) : length($chg);
                }
                $self->atom($mol, $bnd, $iso, $sym2, $chir, $hcnt || 0, 
                    $chg || 0, \@digs, $name);
            } else {
                die "SMILES ERROR: '$all in $s'\n";
            }
        }
    };
    # clean up to avoid memory leak
    $self->{stack} = undef;
    if ($@) {
        croak $@ if $opts->{fatal};
        return;
    }
    $mol;
}

sub parse_digits {
    my ($dig) = @_;
    my @digs;
    while ($dig && $dig =~ /$digits_re/g) {
        push @digs, {bnd=>$1, dig=>$2};
    }
    @digs;
}

sub atom {
    my $self = shift;
    my ($mol,$bnd,$iso,$sym,$chir,$hcount,$chg,$digs,$name) = @_;
    #{no warnings; local $" = ','; print "atom(@_)\n"}
    my $a = $self->{add_atom}($mol,$iso,$sym,$chir,$hcount,$chg,$name);
    if($self->{stack}[-1]) {
        $self->{add_bond}($mol, $bnd, $self->{stack}[-1], $a);
    }
    for my $dig (@$digs) {
        if ($self->{digits}{$dig->{dig}}) {
            if ($dig->{bnd} && $self->{digits}{$dig->{dig}}{bnd}
                &&  $dig->{bnd} ne $self->{digits}{$dig->{dig}}{bnd}){
                die "SMILES: Inconsistent ring closure\n";
            }
            $self->{add_bond}($mol, 
                $dig->{bnd} || $self->{digits}{$dig->{dig}}{bnd}, 
                $self->{digits}{$dig->{dig}}{atom}, $a);
            delete $self->{digits}{$dig->{dig}};
        } else {
            $self->{digits}{$dig->{dig}} = {atom=>$a, bnd=>$dig->{bnd}};
        }
    }
    $self->{stack}[-1] = $a;
}

#=back
#
#=head1 CALLBACK FUNCTIONS
#
#=over
#
#=item $atom = add_atom($mol, $iso, $sym, $chir, $hcount, $chg)
#
#Called by the parser whenever an atom is found. The first parameter is the
#state object given to $obj->parse(). The other parameters are the isotope,
#symbol, chirality, hydrogen count, and charge of the atom. Only the symbol is
#guaranteed to be defined. Mnemonic: the parameters are given in the same order
#that is used in a SMILES string (such as [18OH-]). This callback is expected to
#return something that uniquely identifies the atom that was created (it might
#be a number, a string, or an object).
#
#=cut

# Default add_atom callback 
sub add_atom {
    my ($mol, $iso, $sym, $chir, $hcount, $chg, $name) = @_;
    my $atom = $mol->new_atom(symbol => ucfirst $sym, name => $name);
    $iso && $atom->attr('smiles/isotope' => $iso);
    $iso && $atom->mass($iso);
    $chir && $atom->attr('smiles/chirality' => $chir);
    defined $hcount && $atom->hydrogens($hcount);
    $chg && $atom->formal_charge($chg);
    if ($sym =~ /^[a-z]/) {
        $atom->attr("smiles/aromatic", 1);
    }
    $atom;
}

#=item add_bond($mol, $type, $a1, $a2)
#
#Called by the parser whenever an bond needs to be created. The first parameter
#is the state object given to $obj->parse(). The other parameters are the bond
#type and the two atoms that need to be bonded. The atoms are identified using
#the return values from the add_atom() callback.
#
#=back
#
#=end comment
#
#=cut

# Default add_bond callback 
sub add_bond {
    my ($mol, $type, $a1, $a2) = @_;
    my $order = $type_to_order{$type} or return; # don't add bonds of order 0
    my $bond = $mol->new_bond(type=>$type, atoms=>[$a1, $a2], order=>$order);
    $bond->attr("smiles/type" => $type);
    $bond;
}

sub start_branch {
    my $self = shift;
    #print "start_branch\n";
    push @{$self->{stack}}, $self->{stack}[-1];
}

sub end_branch {
    my $self = shift;
    #print "end_branch\n";
    pop @{$self->{stack}};
}
 
# returns the number of hydrogens for an atom, assuming it has
# no charge or radical (because those require an explicit H-count anyway)
sub calc_implicit_hydrogens {
    my ($self, $atom) = @_;
    no warnings 'uninitialized';
    my $h_count = $ORGANIC_ELEMS{$atom->symbol} - $atom->valence;
    if ($atom->attr("smiles/aromatic") and $atom->symbol =~ /^[CN]$/) {
        $h_count--;
    }
    $h_count = 0 if $h_count < 0;
    $h_count;
}

# returns the number of hydrogens that an atom should have,
# taking into account that it may or may not have a few hydrogens 
# defined already. This assumes that the atom is neutral and not radical
sub calc_implicit_hydrogens_2 {
    my ($self, $atom) = @_;
    my $h_count = $ORGANIC_ELEMS{$atom->symbol} - $atom->valence 
        + $atom->total_hydrogens;
    $h_count = 0 if $h_count < 0;
    $h_count;
}

sub add_implicit_hydrogens {
    my ($self, $mol) = @_;
    for my $atom ($mol->atoms) {
        #print "H=".$atom->hydrogens."\n";
        unless (defined $atom->hydrogens) {
            my $h_count = $self->calc_implicit_hydrogens($atom);
            $atom->hydrogens($h_count);
        }
    }
}

##### SMILES WRITER ########

sub write_string {
    my ($self, $mol_ref, %opts) = @_;

    my $eol;
    my @mols;
    if ($opts{mols}) {
        @mols = @{$opts{mols}};
        $eol = "\n";
    } else {
        @mols = $mol_ref; 
        $eol = "";
    }

    my $smiles;
    for my $mol (@mols) {
        $mol = $mol->clone; 
        $mol->collapse_hydrogens;
        my @atoms = $mol->atoms; 

        if (@atoms) {
            my $i;
            if ($opts{auto_number}) {
                $_->name(++$i) for @atoms;
                $opts{number} = 1;
            }
            if ($opts{unique}) {
                unless ($atoms[0]->attr("canon/class")) {
                    require Chemistry::Canonicalize;
                    Chemistry::Canonicalize::canonicalize($mol);
                }
                $opts{aromatic} = 1; # all unique smiles have to be aromatic
                @atoms = sort {
                    $a->attr("canon/class") <=> $b->attr("canon/class")
                } @atoms;
            }

            if ($opts{aromatic}) {
                require Chemistry::Ring;
                Chemistry::Ring::aromatize_mol($mol);
            }

            my $visited = {};
            my @s;
            for my $atom (@atoms) {
                next if $visited->{$atom};
                my $ring_atoms = {};

                # first pass to find and number the ring bonds
                $self->find_ring_bonds($mol, \%opts, $atom, undef, {}, $ring_atoms);

                # second pass to actually generate the SMILES string
                push @s, $self->branch($mol, \%opts, $atom, undef, $visited, $ring_atoms);
            }
            $smiles .= join '.', @s;
        }

        if ($opts{name}) {
            $smiles .= "\t" . $mol->name;
        }
        $smiles .= $eol;
    }
    return $smiles;
}

sub find_ring_bonds {
    my ($self, $mol, $opts, $atom, $from_bond, $visited, $ring_atoms) = @_;

    $visited->{$atom}  = 1;
    for my $bn ($self->sorted_bonds_neighbors($atom, $opts)) {
        my $nei  = $bn->{to};
        my $bond = $bn->{bond};
        next if $visited->{$bond};
        $visited->{$bond}  = 1;
        if ($visited->{$nei}) { # closed ring
            #print "closing ring\n";
            $ring_atoms->{$nei}++;
        } else {
            $self->find_ring_bonds($mol, $opts, $nei, $bond, $visited, $ring_atoms);
        }
    }
}

sub branch {
    my ($self, $mol, $opts, $atom, $from_bond, $visited, $digits) = @_;

    my $prev_branch = "";
    my $smiles;
    $smiles .= $self->bond_symbol($from_bond, $opts);
    #$digits->{count}++;
    $smiles .= $self->format_atom($atom, $opts);
    if ($digits->{$atom}) {  # opening a ring
        my @d;
        for (1 .. $digits->{$atom}) {
            push @d, $self->next_digit($digits);
        }
        $digits->{$atom} = \@d;
        $smiles .= join "", map { $_ < 10 ? $_ : "%$_"} @d;
    }

    $visited->{$atom}  = 1;
    my @bns = $self->sorted_bonds_neighbors($atom, $opts);

    for my $bn (@bns) {
        my $nei  = $bn->{to};
        my $bond = $bn->{bond};
        next if $visited->{$bond};
        if ($visited->{$nei}) { # closed a ring
            my $digit = shift @{$digits->{$nei}};
            $smiles .= $self->bond_symbol($bond, $opts);
            $smiles .= $digit < 10 ? $digit : "%$digit";
            $digits->{used_digits}[$digit] = 0; # free for future use
            $visited->{$bond} = 1;
        } 
    }
    
    for my $bn (@bns) {
        my $nei  = $bn->{to};
        my $bond = $bn->{bond};
        next if $visited->{$bond};
        $visited->{$bond} = 1;
        unless ($visited->{$nei}) { 
            my $branch = $self->branch($mol, $opts, $nei, $bond, $visited, $digits);
            if ($prev_branch) {
                $smiles .= "($prev_branch)";
            }
            $prev_branch = $branch;
        }
    }
    $smiles .= "$prev_branch";
    $smiles;
}

sub next_digit {
    my ($self, $digits) = @_;
    for (my $i = 1; $i < 100; $i++) {
        unless ($digits->{used_digits}[$i]) {
            $digits->{used_digits}[$i] = 1;  # mark as used
            return $i;
        }
    }
    die "no more available smiles digits!";  # shouldn't happen
}

sub sorted_bonds_neighbors {
    my ($self, $atom, $opts) = @_;
    my @bn = $atom->bonds_neighbors;
    if ($opts->{unique}) {
        @bn = sort { 
            $a->{to}->attr("canon/class") <=> $b->{to}->attr("canon/class") 
        } @bn;
    }
    @bn;
}

my %ORDER_TO_TYPE = (
    2 => '=', 1 => '', 3 => '#',
);

sub bond_symbol {
    my ($self, $bond, $opts) = @_;
    return '' unless $bond;
    return '' if $opts->{aromatic} && $bond->aromatic;
    return $ORDER_TO_TYPE{$bond->order};
}

sub format_atom {
    my ($self, $atom, $opts) = @_;

    my $symbol  = $atom->symbol;
    $symbol = lc $symbol if $opts->{aromatic} && $atom->aromatic;
    my $s = $symbol;

    # unless atom is "simple"...
    if (!$ORGANIC_ELEMS{$atom->symbol} || $atom->formal_charge
        || $atom->total_hydrogens != $self->calc_implicit_hydrogens_2($atom)
        || ($opts->{number} && defined $atom->name)
    ) {
        # "complex atom"; bracketed
        my $h_count = $atom->hydrogens;
        my $charge  = $atom->formal_charge || '';
        my $iso     = $atom->attr("smiles/isotope") || '';
        my $number = '';

        if ($charge and abs($charge) > 1) {
            $charge = sprintf("%+d", $charge);
        } elsif ($charge) {
            $charge = $charge > 0 ? '+' : '-';
        }

        $h_count = $h_count ? ($h_count > 1 ? "H$h_count" : 'H') : '';

        $number = ':' . $atom->name if $opts->{number} and defined $atom->name;

        $s = "[$iso$symbol$h_count$charge$number]";
    }
    $s;
}


1;

=head1 CAVEATS

Stereochemistry is not supported! Stereochemical descriptors such as @, @@, /,
and \ will be silently ignored on input, and will certainly not be produced on
output.

Reading branches that start before an atom, such as (OC)C, which should be
equivalent to C(OC) and COC, according to some variants of the SMILES
specification. Many other tools don't implement this rule either.

The kekulize option works by increasing the bond orders of atoms that don't
have their usual valences satisfied. This may cause problems if you have atoms
with explicitly low hydrogen counts.

=head1 VERSION

0.47

=head1 SEE ALSO

L<Chemistry::Mol>, L<Chemistry::File>

The SMILES Home Page at http://www.daylight.com/dayhtml/smiles/

The Daylight Theory Manual at 
http://www.daylight.com/dayhtml/doc/theory/theory.smiles.html

The PerlMol website L<http://www.perlmol.org/>

=head1 AUTHOR

Ivan Tubert-Brohman E<lt>itub@cpan.orgE<gt>

=head1 COPYRIGHT

Copyright (c) 2009 Ivan Tubert-Brohman. All rights reserved. This program is
free software; you can redistribute it and/or modify it under the same terms as
Perl itself.

=cut