From f2286c240a328431e8aec65f5ff67a2563fa3ccb Mon Sep 17 00:00:00 2001
From: Kanwei Li <kanwei@gmail.com>
Date: Thu, 2 Jun 2011 17:36:38 -0400
Subject: [PATCH] BLAST: Fix sniffer for some dataset. Closes #567. Set
 mimetype to application/xml. Closes #569

---
 datatypes_conf.xml.sample        |  4 ++--
 lib/galaxy/datatypes/registry.py |  3 ++-
 lib/galaxy/datatypes/xml.py      | 30 +++++++++++++++++++++---------
 3 files changed, 25 insertions(+), 12 deletions(-)

diff --git a/datatypes_conf.xml.sample b/datatypes_conf.xml.sample
index ed57048ab2a..a8b3c08e5f6 100644
--- a/datatypes_conf.xml.sample
+++ b/datatypes_conf.xml.sample
@@ -117,8 +117,8 @@
         <datatype extension="taxonomy" type="galaxy.datatypes.tabular:Taxonomy" display_in_upload="true"/>
         <datatype extension="tabular" type="galaxy.datatypes.tabular:Tabular" display_in_upload="true"/>
         <datatype extension="txt" type="galaxy.datatypes.data:Text" display_in_upload="true"/>
-        <datatype extension="memexml" type="galaxy.datatypes.xml:MEMEXml" display_in_upload="true"/>
-        <datatype extension="blastxml" type="galaxy.datatypes.xml:BlastXml" display_in_upload="true"/>
+        <datatype extension="memexml" type="galaxy.datatypes.xml:MEMEXml" mimetype="application/xml" display_in_upload="true"/>
+        <datatype extension="blastxml" type="galaxy.datatypes.xml:BlastXml" mimetype="application/xml" display_in_upload="true"/>
         <datatype extension="vcf" type="galaxy.datatypes.tabular:Vcf" display_in_upload="true">
             <converter file="vcf_to_bgzip_converter.xml" target_datatype="bgzip"/>
             <converter file="vcf_to_tabix_converter.xml" target_datatype="tabix" depends_on="bgzip"/>
diff --git a/lib/galaxy/datatypes/registry.py b/lib/galaxy/datatypes/registry.py
index 7e7b18bc858..838450296b6 100644
--- a/lib/galaxy/datatypes/registry.py
+++ b/lib/galaxy/datatypes/registry.py
@@ -187,7 +187,7 @@ class Registry( object ):
                 'axt'         : 'text/plain',
                 'bam'         : 'application/octet-stream',
                 'bed'         : 'text/plain', 
-                'blastxml'    : 'text/plain', 
+                'blastxml'    : 'application/xml', 
                 'customtrack' : 'text/plain',
                 'csfasta'     : 'text/plain',
                 'fasta'       : 'text/plain',
@@ -200,6 +200,7 @@ class Registry( object ):
                 'laj'         : 'text/plain',
                 'lav'         : 'text/plain',
                 'maf'         : 'text/plain',
+                'memexml'     : 'application/xml',
                 'pileup'      : 'text/plain',
                 'qualsolid'   : 'text/plain',
                 'qualsolexa'  : 'text/plain',
diff --git a/lib/galaxy/datatypes/xml.py b/lib/galaxy/datatypes/xml.py
index 6fc25c34320..2195170f621 100644
--- a/lib/galaxy/datatypes/xml.py
+++ b/lib/galaxy/datatypes/xml.py
@@ -26,19 +26,31 @@ class BlastXml( data.Text ):
         >>> fname = get_test_fname( 'megablast_xml_parser_test1.blastxml' )
         >>> BlastXml().sniff( fname )
         True
+        >>> fname = get_test_fname( 'tblastn_four_human_vs_rhodopsin.xml' )
+        >>> BlastXml().sniff( fname )
+        True
         >>> fname = get_test_fname( 'interval.interval' )
         >>> BlastXml().sniff( fname )
         False
         """
-        blastxml_header = [ '<?xml version="1.0"?>',
-                            '<!DOCTYPE BlastOutput PUBLIC "-//NCBI//NCBI BlastOutput/EN" "http://www.ncbi.nlm.nih.gov/dtd/NCBI_BlastOutput.dtd">',
-                            '<BlastOutput>' ]
-        for i, line in enumerate( file( filename ) ):
-            if i >= len( blastxml_header ):
-                return True
-            line = line.rstrip( '\n\r' )
-            if line != blastxml_header[ i ]:
-                return False
+        #TODO - Use a context manager on Python 2.5+ to close handle
+        handle = open(filename)
+        line = handle.readline()
+        if line.strip() != '<?xml version="1.0"?>':
+            handle.close()
+            return False
+        line = handle.readline()
+        if line.strip() not in ['<!DOCTYPE BlastOutput PUBLIC "-//NCBI//NCBI BlastOutput/EN" "http://www.ncbi.nlm.nih.gov/dtd/NCBI_BlastOutput.dtd">',
+                                '<!DOCTYPE BlastOutput PUBLIC "-//NCBI//NCBI BlastOutput/EN" "NCBI_BlastOutput.dtd">']:
+            handle.close()
+            return False
+        line = handle.readline()
+        if line.strip() != '<BlastOutput>':
+            handle.close()
+            return False
+        handle.close()
+        return True
+        
 
 class MEMEXml( data.Text ):
     """MEME XML Output data"""