solving bugs

FAANG · Nov 29, 2015 · fd2a989 · fd2a989
1 parent d20a1c7
commit fd2a989
Show file tree

Hide file tree

Showing 6 changed files with 120 additions and 7 deletions.
diff --git a/json_schemas/BlueprintSample.schema.dev.json b/json_schemas/BlueprintSample.schema.dev.json
@@ -0,0 +1,108 @@
+{
+    "$schema": "http://json-schema.org/draft-04/schema#",
+    "description": "A representation of a Blueprint Sample XML",
+    "type": "object",
+    "properties": {
+	"id": {
+	    "type": "string",
+	    "pattern": "^S|^ERS",
+	     "minLength": 4
+	},
+	"entity_type": {
+	    "type": "string",
+	    "enum": ["EXPERIMENT","SAMPLE","RUN","STUDY"]
+
+	},
+	"attributes" : {
+	    "type": "array",
+	    "minItems": 2,
+	    "uniqueItems": true,
+	    "items": {
+		"oneOf": [ 
+		    {
+			"type": "object",
+			"properties": {
+			    "subject_id": {
+				"type": "string"
+			    },
+			    "MOLECULE": {
+				"type": "string",
+				"enum": ["total RNA", "polyA RNA", "cytoplasmic RNA", "nuclear RNA", "genomic DNA", "protein", "other"]
+			    },
+			    "DISEASE": {
+				"type": "string"
+			    },
+			    "BIOMATERIAL_PROVIDER": {
+				"type": "string",
+				"enum": ["NIHR Cambridge BioResource","Jose I. Martin-Subero","Prof: Lucia Altucci  (SECONDA UNIVERSITA' di NAPOLI- IT)","Prof.dr. E. Vellenga, University Medical Centre Groningen - Department of Hematology","Sanquin_Nijmegen"]
+			    },
+			    "BIOMATERIAL_TYPE": {
+				"type": "string",
+				"enum": ["Cell Line","Primary Cell","Pooled Primary Cell","Primary Tissue"]
+			    },
+			    "CELL_TYPE": {
+				"type": "string"
+			    },
+			    "TISSUE_TYPE": {
+				"type": "string"
+			    },
+			    "MARKERS": {
+				"type": "string",
+				"pattern": "^CD"
+			    },
+			    "DONOR_ID": {
+				"type": "string"
+			    },
+			    "DONOR_AGE": {
+				"type": "string",
+				"pattern": "[0-9]+-[0-9]+|NA"
+			    },
+			    "DONOR_HEALTH_STATUS": {
+				"type": "string"
+			    },
+			    "DONOR_SEX": {
+				"type": "string",
+				"enum": ["Male","Female","Unknown"]
+			    },
+			    "gender": {
+				"type": "string",
+				"enum": ["male","female","unknown"]
+			    },
+			    "DONOR_ETHNICITY": {
+				"type": "string",
+				"enum": ["arab","caucasian","dutch","nortern european","NA"]
+			    },
+			    "PASSAGE_IF_EXPANDED": {
+				"type": "string"
+			    },
+			    "SAMPLE_ONTOLOGY_URI": {
+				"type": "string",
+				"pattern": "http://purl.obolibrary.org/obo/+"
+			    },
+			    "DISEASE_ONTOLOGY_URI": {
+				"type": "string",
+				"pattern": "^http|NA"
+			    },
+			    "TREATMENT": {
+				"type": "string"
+			    },
+			    "phenotype": {
+				"type": "string",
+				"pattern": "C[0-9]+;CL_[0-9]+;UBERON_[0-9]+"
+			    },
+			    "TISSUE_DEPOT": {
+			        "type": "string"
+			    },
+			    "ENA-CHECKLIST": {
+			    	"type": "string",
+				"pattern": "ERC*"
+			    }			    
+			},
+			"additionalProperties":false
+		    }
+		]
+	    }
+	}
+    },
+    "required": ["id","entity_type","attributes"]
+}
diff --git a/lib/Bio/Metadata/Loader/XMLLoaderRole.pm b/lib/Bio/Metadata/Loader/XMLLoaderRole.pm
@@ -49,8 +49,9 @@ sub load {
             $o = $self->hash_to_object($xml_data);
         }
     elsif ( $root[0] =~ /(.*)_SET/ ) {
-      my $entities=$xml_data->{$root[0]}->{$1};
-      $o=$self->array_to_object($entities)
+      my $entity_type=$1;
+      my $entities=$xml_data->{$root[0]}->{$entity_type};
+      $o=$self->array_to_object($entities,$entity_type)
     }
   #  }
   #  catch {

diff --git a/lib/Bio/Metadata/Loader/XMLSampleLoader.pm b/lib/Bio/Metadata/Loader/XMLSampleLoader.pm
@@ -19,6 +19,7 @@ use warnings;
 
 use Carp;
 use Moose;
+use Data::Dumper;
 use namespace::autoclean;
 use Bio::Metadata::Entity;
 
@@ -53,7 +54,7 @@ sub hash_to_object {
 }
 
 sub array_to_object {
-  my ( $self, $array ) = @_;
+  my ( $self, $array, $type ) = @_;
 
   my @objects;
 
@@ -64,6 +65,10 @@ sub array_to_object {
     my $id=$sample->{'TITLE'};
     $o->id($id);
 
+    #set type
+    $o->entity_type($type);
+
+
     my $attrb_array=$sample->{'SAMPLE_ATTRIBUTES'}->{'SAMPLE_ATTRIBUTE'};
 
     foreach my $attrb (@$attrb_array) {

diff --git a/t/.#validate_xml_bp_sampleset.t b/t/.#validate_xml_bp_sampleset.t
diff --git a/t/data/sample_set.xml b/t/data/sample_set.xml
@@ -40,7 +40,7 @@
       </SAMPLE_ATTRIBUTE>
       <SAMPLE_ATTRIBUTE>
         <TAG>DONOR_AGE</TAG>
-        <VALUE>50 - 55</VALUE>
+        <VALUE>50-55</VALUE>
       </SAMPLE_ATTRIBUTE>
       <SAMPLE_ATTRIBUTE>
         <TAG>DONOR_HEALTH_STATUS</TAG>
@@ -108,7 +108,7 @@
       </SAMPLE_ATTRIBUTE>
       <SAMPLE_ATTRIBUTE>
         <TAG>DONOR_AGE</TAG>
-        <VALUE>50 - 55</VALUE>
+        <VALUE>50-55</VALUE>
       </SAMPLE_ATTRIBUTE>
       <SAMPLE_ATTRIBUTE>
         <TAG>DONOR_HEALTH_STATUS</TAG>

diff --git a/t/validate_xml_bp_sampleset.t b/t/validate_xml_bp_sampleset.t
@@ -13,7 +13,7 @@ use Data::Dumper;
 use Test::More;
 
 my $data_dir = "$Bin/data";
-my $schema_file="$Bin/../json_schemas/BlueprintSample.schema.json";
+my $schema_file="$Bin/../json_schemas/BlueprintSample.schema.dev.json";
 
 my $loader = Bio::Metadata::Loader::XMLSampleLoader->new();