diff --git a/misc-scripts/ontology/scripts/load_OBO_file.pl b/misc-scripts/ontology/scripts/load_OBO_file.pl
index fd127527a235f7d5207e90e2672a5b005d747369..1ea10e04f7d38ecc9cb628b05884c2711d124ba1 100755
--- a/misc-scripts/ontology/scripts/load_OBO_file.pl
+++ b/misc-scripts/ontology/scripts/load_OBO_file.pl
@@ -2,12 +2,12 @@
use strict;
use warnings;
+#TODO Do we need this?
use FindBin;
use lib "$FindBin::Bin/../../../../ONTO-PERL-1.31/lib";
use DBI qw( :sql_types );
use Getopt::Long qw( :config no_ignore_case );
-use IO::File;
use OBO::Core::Ontology;
use OBO::Core::Term;
use OBO::Core::Relationship;
@@ -15,6 +15,7 @@ use OBO::Core::RelationshipType;
use OBO::Core::SynonymTypeDef;
use OBO::Parser::OBOParser;
use OBO::Util::TermSet;
+use POSIX qw/strftime/;
#-----------------------------------------------------------------------
@@ -257,7 +258,7 @@ sub write_term {
$existing_term_sth->execute($term->{'accession'});
my ($existing_term_id, $ontology_id) = $existing_term_sth->fetchrow_array;
$id = $existing_term_id;
- printf(STDERR "CLASH: TERM '%s' already exists in this database. Reusing ID %d\n", $term->{accession}, $existing_term_id);
+ printf(STDERR "DUPLICATION: TERM '%s' already exists in this database. Reusing ID %d\n", $term->{accession}, $existing_term_id);
$reuse = 1;
}
$term->{'id'} = $id;
@@ -267,7 +268,7 @@ sub write_term {
if($term->{synonyms} && @{$term->{synonyms}}) {
if($reuse) {
- print STDERR "REUSE: Skipping synonym writing as term already exists in this database\n";
+ print STDERR "REUSE: SKIPPING SYNONYM writing as term already exists in this database\n";
}
else {
foreach my $syn (@{$term->{'synonyms'}}) {
@@ -349,7 +350,7 @@ sub write_relation {
$dbh->do("LOCK TABLE relation WRITE");
- my $statement = "INSERT INTO relation " . "(child_term_id, parent_term_id, relation_type_id, intersection_of) " . "VALUES (?,?,?,?)";
+ my $statement = "INSERT IGNORE INTO relation " . "(child_term_id, parent_term_id, relation_type_id, intersection_of) " . "VALUES (?,?,?,?)";
my $sth = $dbh->prepare($statement);
@@ -379,6 +380,10 @@ sub write_relation {
$sth->bind_param(3, $relation_types->{$relation_type}{'id'}, SQL_INTEGER);
$sth->bind_param(4, 0, SQL_INTEGER);
$sth->execute();
+
+ if(! $dbh->last_insert_id(undef, undef, 'relation', 'relation_id')) {
+ printf "DUPLICATION: RELATION %s (child) %s %s (parent) has been repeated. Ignoring this insert\n", $child_term->{accession}, $relation_type, $parent_acc;
+ }
}
}
}
@@ -399,6 +404,9 @@ sub write_relation {
$sth->bind_param(3, $relation_types->{$relation_type}{'id'}, SQL_INTEGER);
$sth->bind_param(4, 1, SQL_INTEGER);
$sth->execute();
+ if(! $dbh->last_insert_id(undef, undef, 'relation', 'relation_id')) {
+ printf "DUPLICATION: RELATION %s (child) %s %s (parent) has been repeated. Ignoring this insert\n", $child_term->{accession}, $relation_type, $parent_acc;
+ }
}
}
}
@@ -478,31 +486,58 @@ if ($delete_unknown) {
}
#if parsing an EFO obo file delete xref lines - not compatible with OBO:Parser
-
+#requires correct default-namespace defined
+my $efo_default_namespace;
my @returncode = `grep "default-namespace: efo" $obo_file_name`;
+@returncode = `grep 'property_value: "hasDefaultNamespace" "EFO"' $obo_file_name` if scalar(@returncode) == 0;
my $returncode = @returncode;
if ($returncode > 0) {
- open(OBO_FILE, "$obo_file_name") or die;
- my $terminator = $/;
- undef $/;
- my $buf = <OBO_FILE>;
- $buf =~ s/\cM\cJ/\n/g;
- $/ = $terminator;
- my @file_lines = split(/\n/, $buf);
- my $no_of_lines = @file_lines;
- my $new_obo_file_name = "new" . $obo_file_name;
- open(NEW_OBO_FILE, ">$new_obo_file_name");
-
- for (my $i = 0 ; $i < $no_of_lines ; $i++) {
- my $line = shift(@file_lines);
- if ($line !~ /^xref/) {
- print NEW_OBO_FILE $line, "\n";
+ printf STDERR "Detected an EFO file; performing post-processing\n";
+ my $new_obo_file_name = "${obo_file_name}.new";
+ open my $obo_fh, '<', $obo_file_name or die "Cannot open the file ${obo_file_name}: $!";
+ open my $new_obo_fh, '>', $new_obo_file_name or die "Cannot open target file ${new_obo_file_name}: $!";
+ my $current_state = q{};
+ while(my $line = <$obo_fh>) {
+ chomp($line);
+ my $skip = 0;
+
+ if($line =~/^\[( (?:Term|Typedef|Instance) )\]$/xms) {
+ $current_state = $1;
+ }
+
+ if($current_state eq q{} && $line =~ /^property_value/) {
+ $skip = 1;
+ if($line =~/"hasDefaultNamespace" "(EFO)"/) {
+ $efo_default_namespace = $1;
+ }
+ }
+
+ #Original skip by mk8
+ if($line =~ /^xref/) {
+ $skip = 1;
+ }
+
+ #1.4 term type
+ if($current_state eq 'Term' && $line =~ /^equivalent_to/) {
+ $skip = 1;
}
+
+ if($current_state eq 'Typedef' && $line =~ /^is_ (?:inverse_)? functional/xms) {
+ $skip = 1;
+ }
+
+ #1.2 does not allow this in a typedef field
+ if($current_state eq 'Typedef' && $line =~ /^property_value/) {
+ $skip = 1;
+ }
+
+ print $new_obo_fh $line, "\n" if ! $skip;
}
- close OBO_FILE;
- close NEW_OBO_FILE;
+ close($obo_fh);
+ close($new_obo_fh);
+
+ printf STDERR "Switching to loading the file %s which is a post-processed file\n", $new_obo_file_name;
$obo_file_name = $new_obo_file_name;
-
}
my $my_parser = OBO::Parser::OBOParser->new;
@@ -511,6 +546,16 @@ printf("Reading OBO file '%s'...\n", $obo_file_name);
my $ontology = $my_parser->work($obo_file_name) or die;
+#Reset default namespace if it was EFO and they encoded it as a property
+$ontology->default_namespace($efo_default_namespace) if $efo_default_namespace && ! $ontology->default_namespace();
+
+#Set date to today if one was not present
+if(! $ontology->date()) {
+ print STDERR "No date found; generating one\n";
+ my $date = strftime('%d:%m:%Y %H:%M', localtime());
+ $ontology->date($date);
+}
+
my $statement = "SELECT name from ontology where name = ? group by name";
my $select_sth = $dbh->prepare($statement);