Fixes for Biomart and UCSC to better ensure the data type is correctly set. A data type of "tabular" is now the default, and data will be sniffed if this is the data type.

lib/galaxy/tools/__init__.py

@@ -1240,8 +1240,6 @@ class Tool:
 
     def exec_before_job( self, app, inp_data, out_data, param_dict={} ):
         if self.tool_type == 'data_source':
-            # List for converting UCSC to Galaxy exts, if not in following dictionary, use provided datatype
-            data_type_to_ext = { 'wigdata':'wig', 'tab':'interval', 'hyperlinks':'html', 'sequence':'fasta' }
             dbkey = param_dict.get( 'dbkey' )
             organism = param_dict.get( 'organism' )
             table = param_dict.get( 'table' )
@@ -1263,12 +1261,9 @@ class Tool:
                     data.name = '%s on %s' % ( data.name, gb_landmark_region )
                 data.info = info
                 data.dbkey = dbkey
-                try:
-                    data_type = data_type_to_ext[ data_type ]
-                except: 
-                    pass
-                if data_type not in app.datatypes_registry.datatypes_by_extension: 
-                    data_type = 'interval'
+                if data_type not in app.datatypes_registry.datatypes_by_extension:
+                    # Setting data_type to tabular will force the data to be sniffed in exec_after_process()
+                    data_type = 'tabular'
                 data = app.datatypes_registry.change_datatype( data, data_type )
                 # Store external data source's request parameters temporarily in output file.
                 # In case the config setting for "outputs_to_working_directory" is True, we must write to
@@ -1284,8 +1279,6 @@ class Tool:
             return out_data
 
     def exec_after_process( self, app, inp_data, out_data, param_dict ):
-        # TODO: for data_source tools at least, this code can probably be handled more optimally by adding a new
-        # tag set in the tool config.
         if self.tool_type == 'data_source':
             name, data = out_data.items()[0]
             data.set_size()
@@ -1294,9 +1287,10 @@ class Tool:
                 data.info = param_dict.get( 'info', data.name )
                 data.dbkey = param_dict.get( 'dbkey', data.dbkey )
                 data.extension = param_dict.get( 'data_type', data.extension )
-            if data.extension == 'txt':
+            if data.extension in [ 'txt', 'tabular' ]:
                 data_type = sniff.guess_ext( data.file_name, sniff_order=app.datatypes_registry.sniff_order )
-                data = app.datatypes_registry.change_datatype( data, data_type )
+                if data.extension != data_type:
+                    data = app.datatypes_registry.change_datatype( data, data_type )
             elif not isinstance( data.datatype, datatypes.interval.Bed ) and isinstance( data.datatype, datatypes.interval.Interval ):
                 data.set_meta()
                 if data.missing_meta(): 

tools/data_source/biomart.xml

@@ -21,17 +21,21 @@
                 <param_from_source name="GALAXY_URL" missing="0" />
             </add_to_url>
         </request_param>
+        <request_param galaxy_name="data_type" remote_name="exportView_outputformat" missing="tabular" >
+            <data_type_translation>
+                <format galaxy_format="tabular" remote_format="TSV" />
+            </data_type_translation>
+        </request_param>
         <request_param galaxy_name="dbkey" remote_name="dbkey" missing="?" />
         <request_param galaxy_name="organism" remote_name="organism" missing="" />
         <request_param galaxy_name="table" remote_name="table" missing="" />
         <request_param galaxy_name="description" remote_name="description" missing="" />
         <request_param galaxy_name="name" remote_name="name" missing="Biomart query" />
         <request_param galaxy_name="info" remote_name="info" missing="" />
-        <request_param galaxy_name="data_type" remote_name="type" missing="txt" />
     </request_param_translation>
 	<uihints minwidth="800"/>
 	<outputs>
-		<data name="output" format="txt" />
+		<data name="output" format="tabular" />
 	</outputs>
 	<options sanitize="False" refresh="True"/>
 </tool>

tools/data_source/biomart_test.xml

@@ -21,17 +21,21 @@
                 <param_from_source name="GALAXY_URL" missing="0" />
             </add_to_url>
         </request_param>
+        <request_param galaxy_name="data_type" remote_name="exportView_outputformat" missing="tabular" >
+            <data_type_translation>
+                <format galaxy_format="tabular" remote_format="TSV" />
+            </data_type_translation>
+        </request_param>
         <request_param galaxy_name="dbkey" remote_name="dbkey" missing="?" />
         <request_param galaxy_name="organism" remote_name="organism" missing="" />
         <request_param galaxy_name="table" remote_name="table" missing="" />
         <request_param galaxy_name="description" remote_name="description" missing="" />
         <request_param galaxy_name="name" remote_name="name" missing="Biomart test query" />
         <request_param galaxy_name="info" remote_name="info" missing="" />
-        <request_param galaxy_name="data_type" remote_name="type" missing="txt" />
     </request_param_translation>
 	<uihints minwidth="800"/>		
 	<outputs>
-		<data name="output" format="txt" />
+		<data name="output" format="tabular" />
 	</outputs>
 	<options sanitize="False" refresh="True"/>
 </tool>

tools/data_source/ucsc_tablebrowser.xml

@@ -21,16 +21,20 @@
         <request_param galaxy_name="organism" remote_name="org" missing="unknown species" />
 	    <request_param galaxy_name="table" remote_name="hgta_track" missing="unknown table" />
 	    <request_param galaxy_name="description" remote_name="hgta_regionType" missing="no description" />
-	    <request_param galaxy_name="data_type" remote_name="hgta_outputType" missing="bed" >
+	    <request_param galaxy_name="data_type" remote_name="hgta_outputType" missing="tabular" >
             <data_type_translation>
                 <format galaxy_format="tabular" remote_format="primaryTable" />
                 <format galaxy_format="tabular" remote_format="selectedFields" />
+                <format galaxy_format="wig" remote_format="wigdata" />
+                <format galaxy_format="interval" remote_format="tab" />
+                <format galaxy_format="html" remote_format="hyperlinks" />
+                <format galaxy_format="fasta" remote_format="sequence" />
             </data_type_translation>
         </request_param>
 	</request_param_translation>
 	<uihints minwidth="800"/>
 	<outputs>
-		<data name="output" format="bed" />
+		<data name="output" format="tabular" />
 	</outputs>
 	<options sanitize="False" refresh="True"/>
 </tool>

tools/data_source/ucsc_tablebrowser_archaea.xml

@@ -21,16 +21,20 @@
         <request_param galaxy_name="organism" remote_name="org" missing="unknown species" />
         <request_param galaxy_name="table" remote_name="hgta_track" missing="" />
         <request_param galaxy_name="description" remote_name="hgta_regionType" missing="" />
-        <request_param galaxy_name="data_type" remote_name="hgta_outputType" missing="bed" >
+        <request_param galaxy_name="data_type" remote_name="hgta_outputType" missing="tabular" >
             <data_type_translation>
                 <format galaxy_format="tabular" remote_format="primaryTable" />
                 <format galaxy_format="tabular" remote_format="selectedFields" />
+                <format galaxy_format="wig" remote_format="wigdata" />
+                <format galaxy_format="interval" remote_format="tab" />
+                <format galaxy_format="html" remote_format="hyperlinks" />
+                <format galaxy_format="fasta" remote_format="sequence" />
             </data_type_translation>
         </request_param>
     </request_param_translation>
 	<uihints minwidth="800"/>
 	<outputs>
-		<data name="output" format="bed" />
+		<data name="output" format="tabular" />
 	</outputs>
 	<options sanitize="False" refresh="True"/>
 </tool>

tools/data_source/ucsc_tablebrowser_test.xml

@@ -21,16 +21,20 @@
         <request_param galaxy_name="organism" remote_name="org" missing="unknown species" />
         <request_param galaxy_name="table" remote_name="hgta_track" missing="" />
         <request_param galaxy_name="description" remote_name="hgta_regionType" missing="" />
-        <request_param galaxy_name="data_type" remote_name="hgta_outputType" missing="bed" >
+        <request_param galaxy_name="data_type" remote_name="hgta_outputType" missing="tabular" >
             <data_type_translation>
                 <format galaxy_format="tabular" remote_format="primaryTable" />
                 <format galaxy_format="tabular" remote_format="selectedFields" />
+                <format galaxy_format="wig" remote_format="wigdata" />
+                <format galaxy_format="interval" remote_format="tab" />
+                <format galaxy_format="html" remote_format="hyperlinks" />
+                <format galaxy_format="fasta" remote_format="sequence" />
             </data_type_translation>
         </request_param>
     </request_param_translation>
 	<uihints minwidth="800"/>
 	<outputs>
-		<data name="output" format="bed" />
+		<data name="output" format="tabular" />
 	</outputs>
 	<options sanitize="False" refresh="True"/>
 </tool>