#!perl
use strict;
use Data::Dumper;
use Carp;

#
# This is a SAS Component
#


=head1 query_entity_Publication

Query the entity Publication.

Annotators attach publications to ProteinSequences.  The criteria we have used
to gather such connections is a bit nonstandard.  We have sought to attach publications
to ProteinSequences when the publication includes an expert asserting a belief or estimate
of function.  The paper may not be the original characterization.  Further, it may not
even discuss a sequence protein (much of the literature is very valuable, but reports
work on proteins in strains that have not yet been sequenced).  On the other hand,
reports of sequencing regions of a chromosome (with no specific assertion of a
clear function) should not be attached.  The attached publications give an ID (usually a
Pubmed ID),  a URL to the paper (when we have it), and a title (when we have it).


Example:

    query_entity_Publication -a 

=head2 Related entities

The Publication entity has the following relationship links:

=over 4
    
=item Concerns ProteinSequence


=back


=head2 Command-Line Options

=over 4

=item -is field,value

Limit the results to entities where the given field has the given value.

=item -like field,value

Limit the results to entities where the given field is LIKE (in the sql sense) the given value.

=item -op operator,field,value

Limit the results to entities where the given field is related to the given value based on the given operator.

The operators supported are as follows. We provide text based alternatives to the comparison
operators so that extra quoting is not required to keep the command-line shell from 
confusing them with shell I/O redirection operators.

=over 4

=item < or lt

=item > or gt

=item <=  or le

=item >= or ge

=item =

=item LIKE

=back

=item -a

Return all fields.

=item -h

Display a list of the fields available for use.

=item -fields field-list

Choose a set of fields to return. Field-list is a comma-separated list of 
strings. The following fields are available:

=over 4

=item title

=item link

=item pubdate

=back    
   
=back

=head2 Output Format

The standard output is a tab-delimited file containing a column
for each requested field.

=cut

use Bio::KBase::CDMI::CDMIClient;
use Getopt::Long;

#Default fields

my @all_fields = ( 'title', 'link', 'pubdate' );
my %all_fields = map { $_ => 1 } @all_fields, 'id';

my $usage = "usage: query_entity_Publication [-is field,value] [-like field,value] [-op operator,field,value] [-show-fields] [-a | -f field list] > entity.data";

my $a;
my $f;
my @fields;
my $show_fields;
my @query_is;
my @query_like;
my @query_op;

my %op_map = ('>', '>',
	      'gt', '>',
	      '<', '<',
	      'lt', '<',
	      '>=', '>=',
	      'ge', '>=',
	      '<=', '<=',
	      'le', '<=',
	      'like', 'LIKE',
	      );

my $geO = Bio::KBase::CDMI::CDMIClient->new_get_entity_for_script("a" 		=> \$a,
								  "show-fields" => \$show_fields,
								  "h" 		=> \$show_fields,
								  "is=s"	=> \@query_is,
								  "like=s"	=> \@query_like,
								  "op=s"	=> \@query_op,
								  "fields=s"    => \$f);

if ($show_fields)
{
    print STDERR "Available fields: @all_fields\n";
    exit 0;
}

if (@ARGV != 0 || ($a && $f))
{
    print STDERR $usage, "\n";
    exit 1;
}

if ($a)
{
    @fields = @all_fields;
}
elsif ($f) {
    my @err;
    for my $field (split(",", $f))
    {
	if (!$all_fields{$field})
	{
	    push(@err, $field);
	}
	else
	{
	    push(@fields, $field);
	}
    }
    if (@err)
    {
	print STDERR "all_entities_Publication: unknown fields @err. Valid fields are: @all_fields\n";
	exit 1;
    }
}

my @qry;

for my $ent (@query_is)
{
    my($field,$value) = split(/,/, $ent, 2);
    if (!$all_fields{$field})
    {
	die "$field is not a valid field\n";
    }
    
    push(@qry, [$field, '=', $value]);
}

for my $ent (@query_like)
{
    my($field,$value) = split(/,/, $ent, 2);
    if (!$all_fields{$field})
    {
	die "$field is not a valid field\n";
    }
    
    push(@qry, [$field, 'LIKE', $value]);
}

for my $ent (@query_op)
{
    my($op,$field,$value) = split(/,/, $ent, 3);

    if (!$all_fields{$field})
    {
	die "$field is not a valid field\n";
    }
    my $mapped_op = $op_map{lc($op)};
    if (!$mapped_op)
    {
	die "$op is not a valid operator\n";
    }
    
    push(@qry, [$field, $mapped_op, $value]);
}

my $h = $geO->query_entity_Publication(\@qry, \@fields );

while (my($k, $v) = each %$h)
{
    print join("\t", $k, map { ref($_) eq 'ARRAY' ? join(",", @$_) : $_ } @$v{@fields}), "\n";
}

