Better approach to altering initial content of output dataset if necessary. The upload tool will now call the data type's groom_output_dataset() method ( a better name than before_setting_metadata since it is not related to metadata ). This will now also run on the cluster.

tools/data_source/hbvar_filter.py

@@ -46,7 +46,6 @@ def exec_after_process(app, inp_data, out_data, param_dict, tool=None, stdout=No
     fp.close()
     #Set meta data, format file to be valid interval type
     if isinstance(data.datatype, datatypes.interval.Interval):
-        data.datatype.before_setting_metadata( data )
         data.set_meta(first_line_is_header=True)
         #check for missing meta data, if all there, comment first line and process file
         if not data.missing_meta():

tools/data_source/upload.py

@@ -10,6 +10,7 @@ from galaxy import eggs
 import galaxy.model
 from galaxy.datatypes import sniff
 from galaxy.datatypes.binary import *
+from galaxy.datatypes.registry import Registry
 from galaxy import util
 from galaxy.util.json import *
 
@@ -264,6 +265,9 @@ def add_file( dataset, json_file, output_path ):
                  name = dataset.name,
                  line_count = line_count )
     json_file.write( to_json_string( info ) + "\n" )
+    # Groom the dataset content if necessary
+    datatype = Registry().get_datatype_by_extension( ext )
+    datatype.groom_dataset_content( output_path )
 
 def add_composite_file( dataset, json_file, output_path, files_path ):
         if dataset.composite_files: