cram datatype: support for crai indexes via pysam

lib/galaxy/datatypes/binary.py

@@ -461,14 +461,44 @@ class CRAM( Binary ):
     edam_format = "format_3462"
 
     MetadataElement( name="cram_version", default=None, desc="CRAM Version", param=MetadataParameter, readonly=True, visible=False, optional=False, no_value=None )
+    MetadataElement( name="cram_index", desc="CRAM Index File", param=metadata.FileParameter, file_ext="crai", readonly=True, no_value=None, visible=False, optional=True )
 
     def set_meta( self, dataset, overwrite=True, **kwd ):
+        major_version, minor_version = self.get_cram_version( dataset.file_name )
+        if major_version > 0:
+            dataset.metadata.cram_version = str(major_version) + "." + str(minor_version)
+
+        if not dataset.metadata.bam_index:
+            index_file = dataset.metadata.spec['cram_index'].param.new_file( dataset=dataset )
+            self.set_index_file(dataset, index_file)
+
+    def get_cram_version( self, filename):
         try:
-            with open(dataset.file_name, "r") as fh:
+            with open(, "r") as fh:
                 header = fh.read(6)
-                dataset.metadata.cram_version = str(ord(header[4])) + "." + str(ord(header[5]))
-        except Exception as exc:
-            log.warn( '%s, set_meta Exception: %s', self, exc )
+                return ord(header[4]), ord(header[5])
+            except Exception as exc:
+                log.warn( '%s, set_meta Exception: %s', self, exc )
+                return -1, -1
+
+    def get_index_file(self, dataset, index_file):
+        try:
+            ## @todo when pysam 1.2.1 or pysam 1.3.0 gets released use pysam.index(alignment, target_idx)
+            ## This currently gives coredump in the current release but is fixed in the dev branch
+            ## xref: https://github.com/samtools/samtools/issues/199
+
+            dataset_symlink = os.path.join( os.path.dirname( index_file.file_name ), '__dataset_%d_%s' % ( dataset.id, os.path.basename( index_file.file_name ) ) )
+            os.symlink( dataset.file_name, dataset_symlink )            
+            pysam.index(bam_file)
+
+            # if file dataset_symlink + ".bai" exists
+            #     shutil.move( dataset_symlink + '.bai', index_file.file_name ) 
+            # else
+            #     os.unlink( dataset_symlink )
+            #     return False
+        except:
+            # Can not read the bam file for some reason
+            return False
 
     def set_peek( self, dataset, is_multi_byte=False ):
         if not dataset.dataset.purged: