From 17cfbb920f818e239dcb5abebccdbae55df481c4 Mon Sep 17 00:00:00 2001
From: Keiran Raine <kr2@sanger.ac.uk>
Date: Mon, 31 Oct 2016 16:13:56 +0000
Subject: [PATCH 1/3] Allows use of ensemblgenomes.org as a datasource

---
 bin/Admin_EnsemblReferenceFileGenerator.pl | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/bin/Admin_EnsemblReferenceFileGenerator.pl b/bin/Admin_EnsemblReferenceFileGenerator.pl
index d704931..074cba0 100755
--- a/bin/Admin_EnsemblReferenceFileGenerator.pl
+++ b/bin/Admin_EnsemblReferenceFileGenerator.pl
@@ -53,7 +53,7 @@
 const my $CACHE_SUFFIX_RAW => 'vagrent.cache.raw';
 const my @TRANSCRIPT_BIOTYPES => qw(protein_coding lincRNA miRNA snoRNA rRNA snRNA);
 const my $ENSEMBL_SPECIES_ASSEMBLY => qr/([^\.]+?)\.(.+?)\./;
-const my $ENSEMBL_VERSION_PATTERN => qr/^ftp\:\/\/ftp\.ensembl\.org\/pub\/release\-(\d+?)\//;
+const my $ENSEMBL_VERSION_PATTERN => qr/^ftp\:\/\/ftp\.ensembl(?:genomes)?\.org\/pub\/release\-(\d+?)\//;
 
 try {
   my $opts = option_builder();

From 5556199713e290e68a61e5845eaaaeada8cfe96c Mon Sep 17 00:00:00 2001
From: Keiran Raine <kr2@sanger.ac.uk>
Date: Mon, 31 Oct 2016 16:15:28 +0000
Subject: [PATCH 2/3] Handle genes without names, and give more useful error
 message

---
 bin/Admin_EnsemblGtf2CacheConverter.pl | 21 ++++++++++++++++++---
 1 file changed, 18 insertions(+), 3 deletions(-)

diff --git a/bin/Admin_EnsemblGtf2CacheConverter.pl b/bin/Admin_EnsemblGtf2CacheConverter.pl
index bc11c66..6f78393 100755
--- a/bin/Admin_EnsemblGtf2CacheConverter.pl
+++ b/bin/Admin_EnsemblGtf2CacheConverter.pl
@@ -104,6 +104,7 @@ sub convertGtf {
     if(exists $attr{'transcript_id'} && defined $attr{'transcript_id'}){
 			$acc = unquoteValue($attr{'transcript_id'});
 			unless(exists $lookup->{$acc}) {
+        next unless(exists $attr{'transcript_version'});
 			  $acc .= '.'.unquoteValue($attr{'transcript_version'});
 			  next unless exists $lookup->{$acc};
 			}
@@ -125,7 +126,15 @@ sub convertGtf {
       $c++;
       $wip->{$acc}->{'type'} = $bioType;
 			$wip->{$acc}->{'acc'} = $acc;
-			$wip->{$acc}->{'gene'} = unquoteValue($attr{'gene_name'});
+			if(exists $attr{'gene_name'}) {
+			  $wip->{$acc}->{'gene'} = unquoteValue($attr{'gene_name'});
+			}
+			elsif(exists $attr{'gene_id'}) {
+			  $wip->{$acc}->{'gene'} = unquoteValue($attr{'gene_id'});
+			}
+			else {
+			  croak "Cannot identify gene name or ID for structure: ".Dumper(\%attr);
+			}
       $wip->{$acc}->{'CCDS'} = unquoteValue($attr{'ccds_id'}) if exists $attr{'ccds_id'};
     }
     if($lineType eq $CDS_TYPE && !defined $wip->{$acc}->{'protacc'}){
@@ -257,8 +266,14 @@ sub convertTranscript {
 
 sub writeTranscript {
 	my ($fh,$t,$rawT) = @_;
-	print $fh join("\t",$rawT->{'lines'}->{$EXON_TYPE}->[0]->[0],$t->getGenomicMinPos - 1,
-											$t->getGenomicMaxPos,$t->getAccession,$t->getGeneName,length $t->getcDNASeq);
+	eval {
+    print $fh join("\t",$rawT->{'lines'}->{$EXON_TYPE}->[0]->[0],$t->getGenomicMinPos - 1,
+                        $t->getGenomicMaxPos,$t->getAccession,$t->getGeneName,length $t->getcDNASeq);
+    1;
+	};
+	if($@) {
+	  die "\nTranscript Object: ".Dumper($t)."\n\nExon_Type layer: ".Dumper($rawT->{'lines'}->{$EXON_TYPE})."\n\nERROR: Abandon hope, Ensemble structure has changed\n";
+	}
 	$t->{_cdnaseq} = undef;
 	print $fh "\t",Dumper($t),"\n";
 }

From 7f92d5248ac03b1f75fb173fc50363dfdf3d929a Mon Sep 17 00:00:00 2001
From: Keiran Raine <kr2@sanger.ac.uk>
Date: Mon, 31 Oct 2016 16:18:53 +0000
Subject: [PATCH 3/3] Ensure an error code is emmitted on failure

---
 bin/Admin_EnsemblGtf2CacheConverter.pl | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/bin/Admin_EnsemblGtf2CacheConverter.pl b/bin/Admin_EnsemblGtf2CacheConverter.pl
index 6f78393..83d9e3c 100755
--- a/bin/Admin_EnsemblGtf2CacheConverter.pl
+++ b/bin/Admin_EnsemblGtf2CacheConverter.pl
@@ -62,7 +62,7 @@
 	$ccds = parseCCDSFile($opts->{'c'}) if defined $opts->{'c'};
 	convertGtf($opts,$lookup,$ccds);
 } catch {
-  warn "An error occurred while building reference support files\:\n\t$_"; # not $@
+  die "An error occurred while building reference support files\:\n\t$_"; # not $@
 };
 
 sub convertGtf {