Merge pull request #4368 from mvdbeek/tie_down_collection_matcher

Implement default_identifier_source for outputs and sort input keys by default

lib/galaxy/dataset_collections/matching.py

@@ -75,7 +75,7 @@ class MatchingCollections( object ):
             return None
 
         matching_collections = MatchingCollections()
-        for input_key, to_match in collections_to_match.items():
+        for input_key, to_match in sorted(collections_to_match.items()):
             hdca = to_match.hdca
             collection_type_description = collection_type_descriptions.for_collection_type( hdca.collection.collection_type )
             subcollection_type = to_match.subcollection_type

lib/galaxy/tools/execute.py

@@ -157,7 +157,21 @@ class ToolExecutionTracker( object ):
                 log.warning( "Problem matching up datasets while attempting to create implicit dataset collections")
                 continue
             output = self.tool.outputs[ output_name ]
-            element_identifiers = structure.element_identifiers_for_outputs( trans, outputs )
+
+            element_identifiers = None
+            if hasattr(output, "default_identifier_source"):
+                # Switch the structure for outputs if the output specified a default_identifier_source
+                collection_type_descriptions = trans.app.dataset_collections_service.collection_type_descriptions
+
+                source_collection = self.collection_info.collections.get(output.default_identifier_source)
+                if source_collection:
+                    collection_type_description = collection_type_descriptions.for_collection_type(source_collection.collection.collection_type)
+                    _structure = structure.for_dataset_collection( source_collection.collection, collection_type_description=collection_type_description)
+                    if structure.can_match(_structure):
+                        element_identifiers = _structure.element_identifiers_for_outputs(trans, outputs)
+
+            if not element_identifiers:
+                element_identifiers = structure.element_identifiers_for_outputs( trans, outputs )
 
             implicit_collection_info = dict(
                 implicit_inputs=implicit_inputs,

lib/galaxy/tools/parser/xml.py

@@ -289,6 +289,7 @@ class XmlToolSource(ToolSource):
         output.format = output_format
         output.change_format = data_elem.findall("change_format")
         output.format_source = data_elem.get("format_source", default_format_source)
+        output.default_identifier_source = data_elem.get("default_identifier_source", 'None')
         output.metadata_source = data_elem.get("metadata_source", default_metadata_source)
         output.parent = data_elem.get("parent", None)
         output.label = xml_text( data_elem, "label" )

lib/galaxy/tools/parser/yaml.py

@@ -113,6 +113,7 @@ class YamlToolSource(ToolSource):
         output.format = output_dict.get("format", "data")
         output.change_format = []
         output.format_source = output_dict.get("format_source", None)
+        output.default_identifier_source = output_dict.get("default_identifier_source", None)
         output.metadata_source = output_dict.get("metadata_source", "")
         output.parent = output_dict.get("parent", None)
         output.label = output_dict.get( "label", None )

lib/galaxy/tools/xsd/galaxy.xsd

@@ -3693,6 +3693,12 @@ The valid values for format can be found in
         <xs:documentation xml:lang="en">This sets the data type of the output file to be the same format as that of a tool input dataset.</xs:documentation>
       </xs:annotation>
     </xs:attribute>
+    <xs:attribute name="default_identifier_source" type="xs:string">
+      <xs:annotation>
+        <xs:documentation xml:lang="en">Sets the source of element identifier to the specified input.
+This only applies to collections that are mapped over a non-collection input and that have equivalent structures.</xs:documentation>
+      </xs:annotation>
+    </xs:attribute>
     <xs:attribute name="metadata_source" type="xs:string">
       <xs:annotation>
         <xs:documentation xml:lang="en">This copies the metadata information

test/api/test_tools.py

@@ -1059,6 +1059,20 @@ class ToolsTestCase( api.ApiTestCase ):
         self.assertEquals( len( response_object[ 'jobs' ] ), 2 )
         self.assertEquals( len( response_object[ 'implicit_collections' ] ), 1 )
 
+    @skip_without_tool( "identifier_source" )
+    def test_default_identifier_source_map_over(self):
+        history_id = self.dataset_populator.new_history()
+        input_a_hdca_id = self.dataset_collection_populator.create_list_in_history( history_id, contents=[("A", "A content")]).json()['id']
+        input_b_hdca_id = self.dataset_collection_populator.create_list_in_history(history_id, contents=[("B", "B content")]).json()['id']
+        inputs = {
+            "inputA": { 'batch': True, 'values': [ dict( src="hdca", id=input_a_hdca_id ) ] },
+            "inputB": { 'batch': True, 'values': [ dict( src="hdca", id=input_b_hdca_id ) ] },
+        }
+        self.dataset_populator.wait_for_history( history_id, assert_ok=True )
+        create = self._run("identifier_source", history_id, inputs, assert_ok=True)
+        assert create['implicit_collections'][0]['elements'][0]['element_identifier'] == 'B'
+        assert create['implicit_collections'][1]['elements'][0]['element_identifier'] == 'A'
+
     @skip_without_tool( "collection_creates_pair" )
     def test_map_over_collection_output( self ):
         history_id = self.dataset_populator.new_history()

test/functional/tools/samples_tool_conf.xml

@@ -78,6 +78,7 @@
   <tool file="validation_empty_dataset.xml" />
   <tool file="implicit_conversion.xml" />
   <tool file="explicit_conversion.xml" />
+  <tool file="identifier_source.xml" />
   <tool file="identifier_single.xml" />
   <tool file="identifier_multiple.xml" />
   <tool file="identifier_multiple_in_conditional.xml" />