Added support for gff3 datatype

lib/galaxy/datatypes/interval.py

@@ -330,6 +330,16 @@ class Gff( Tabular ):
                         ret_val.append( (site_name, link) )
         return ret_val
 
+class Gff3( Gff ):
+    """Tab delimited data in Gff3 format"""
+
+    """Add metadata elements"""
+    MetadataElement( name="column_types", default=['str','str','str','int','int','float','str','int','list'], desc="Column types", readonly=True )
+    
+    def __init__(self, **kwd):
+        """Initialize datatype, by adding GBrowse display app"""
+        Gff.__init__(self, **kwd)
+
 class Wiggle( Tabular ):
     """Tab delimited data in wiggle format"""
     MetadataElement( name="columns", default=3, desc="Number of columns", readonly=True )

lib/galaxy/datatypes/registry.py

@@ -48,6 +48,7 @@ class Registry( object ):
                 'maf'      : sequence.Maf(),
                 'axt'      : sequence.Axt(),
                 'gff'      : interval.Gff(),
+                'gff3'     : interval.Gff3(),  
                 'wig'      : interval.Wiggle(),
                 'gmaj.zip' : images.Gmaj(),
                 'laj'      : images.Laj(),
@@ -69,6 +70,7 @@ class Registry( object ):
                 'maf'      : 'text/plain',
                 'axt'      : 'text/plain',
                 'gff'      : 'text/plain',
+                'gff3'     : 'text/plain',
                 'wig'      : 'text/plain',
                 'gmaj.zip' : 'application/zip',
                 'laj'      : 'text/plain',

lib/galaxy/datatypes/sniff.py

@@ -599,8 +599,10 @@ def guess_ext(fname):
 
         if is_column_based(fname, '\t', 1):
             headers = get_headers(fname, '\t')
-            if is_gff(headers) or is_gff3(headers):
+            if is_gff(headers):
                 return 'gff'
+            if is_gff3(headers):
+                return 'gff3'
             elif is_interval(headers):
                 return 'interval'
             else:

lib/galaxy/datatypes/tabular.py

@@ -86,6 +86,26 @@ class Tabular( data.Text ):
                                     proceed = True
                                 except:
                                     pass
+                        elif format == 'gff3':
+                            valid_gff3_strand = ['+', '-', '.', '?']
+                            valid_start = False
+                            valid_end = False
+                            if elems_len == 9:
+                                try:
+                                    start = int(hdr[3])
+                                    valid_start = True                                    
+                                except:
+                                    if hdr[3] == '.':
+                                        valid_start = True                                        
+                                try:
+                                    end = int(hdr[4])
+                                    valid_end = True
+                                except:
+                                    if hdr[4] == '.':
+                                        valid_end = True
+                                    srand = hdr[6]
+                                if valid_start and valid_end and start < end and strand in valid_gff3_strand:
+                                    proceed = True                
                         elif format=='wig':
                             try:
                                 int( elems[0] )

lib/galaxy/util/__init__.py

@@ -70,7 +70,7 @@ mime_types = {
 
 text_types = sets.Set([ 
     'txt', 'text', 'wig', 'genbank', 'motif', 'acedb', 'nexus', 'fitch', 'meganon', 'codata', 'dbmotif', 
-    'table', 'fasta', 'txt', 'gff', 'pir', 'ig', 'seqtable', 'clustal', 'gcg', 'hennig86', 'excel', 'asn1', 
+    'table', 'fasta', 'txt', 'gff', 'gff3', 'pir', 'ig', 'seqtable', 'clustal', 'gcg', 'hennig86', 'excel', 'asn1', 
     'regions', 'simple', 'score', 'text', 'msf', 'selex', 'tagseq', 'embl', 'srspair', 'staden', 
     'strider', 'xbed', 'markx10', 'pair', 'markx1', 'markx0', 'markx3', 'markx2', 'jackknifer', 
     'ncbi', 'mega', 'fa', 'feattable', 'phylip', 'diffseq', 'bed', 'srs', 'jackknifernon', 'swiss', 

tools/data_source/ucsc_filter.py

@@ -22,6 +22,8 @@ def exec_before_job( app, inp_data, out_data, param_dict, tool=None):
             ext = "maf"
         elif outputType == 'gff':
             ext = "gff"
+        elif outputType == 'gff3':
+            ext = "gff3"
         else:
             if 'hgta_doPrintSelectedFields' in param_dict:
                 ext = "interval"

tools/data_source/upload.xml

@@ -25,6 +25,7 @@
 			<option value="axt">AXT</option>
 			<option value="tabular">Tab delimited</option>
 			<option value="gff">Gff</option>
+			<option value="gff3">Gff3</option>
 			<option value="lav">LAV</option>
 			<option value="html">HTML</option>
 			<option value="wig">Wiggle</option>
@@ -112,7 +113,13 @@ A sequence in FASTA format consists of a single-line description, followed by li
 
 **Gff**
 
-GFF lines have nine required fields that must be tab-separated.  Gff version 3 is also supported.
+GFF lines have nine required fields that must be tab-separated.
+
+-----
+
+**Gff3**
+
+The proposed GFF3 format addresses the most common extensions to GFF, while preserving backward compatibility with previous formats.
 
 -----
 

universe_wsgi.ini.sample

@@ -116,6 +116,7 @@ fasta = galaxy.datatypes.sequence:Fasta
 maf = galaxy.datatypes.sequence:Maf
 axt = galaxy.datatypes.sequence:Axt
 gff = galaxy.datatypes.interval:Gff
+gff3 = galaxy.datatypes.interval:Gff3
 wig = galaxy.datatypes.interval:Wiggle
 gmaj.zip = galaxy.datatypes.images:Gmaj,application/zip
 laj = galaxy.datatypes.images:Laj