[galaxy-dev] [hg] galaxy 2568: Added two new datatypes (sam and bam) and an a...

24 Aug 2009

details:   http://www.bx.psu.edu/hg/galaxy/rev/670f8800a2bf
changeset: 2568:670f8800a2bf
user:      Kelly Vincent <kpvincent@bx.psu.edu>
date:      Wed Aug 19 09:13:21 2009 -0400
description:
Added two new datatypes (sam and bam) and an associated test

7 file(s) affected in this change:

datatypes_conf.xml.sample
lib/galaxy/datatypes/images.py
lib/galaxy/datatypes/registry.py
lib/galaxy/datatypes/tabular.py
lib/galaxy/datatypes/test/1.sam
test-data/1.sam
test/functional/test_sniffing_and_metadata_settings.py

diffs (362 lines):

diff -r b56291fad13d -r 670f8800a2bf datatypes_conf.xml.sample

--- a/datatypes_conf.xml.sample	Mon Aug 17 09:54:57 2009 -0400
+++ b/datatypes_conf.xml.sample	Wed Aug 19 09:13:21 2009 -0400
@@ -3,6 +3,7 @@
   <registration converters_path="lib/galaxy/datatypes/converters">
     <datatype extension="ab1" type="galaxy.datatypes.images:Ab1" mimetype="application/octet-stream" display_in_upload="true"/>
     <datatype extension="axt" type="galaxy.datatypes.sequence:Axt" display_in_upload="true"/>
+    <datatype extension="bam" type="galaxy.datatypes.images:Bam" mimetype="application/octet-stream"/>
     <datatype extension="bed" type="galaxy.datatypes.interval:Bed" display_in_upload="true">
       <converter file="bed_to_gff_converter.xml" target_datatype="gff"/>
       <converter file="interval_to_coverage.xml" target_datatype="coverage"/>
@@ -49,6 +50,7 @@
     <datatype extension="qualsolexa" type="galaxy.datatypes.qualityscore:QualityScoreSolexa" display_in_upload="true"/>
     <datatype extension="qualsolid" type="galaxy.datatypes.qualityscore:QualityScoreSOLiD" display_in_upload="true"/>
     <datatype extension="qual454" type="galaxy.datatypes.qualityscore:QualityScore454" display_in_upload="true"/>
+    <datatype extension="sam" type="galaxy.datatypes.tabular:Sam" display_in_upload="true"/>
     <datatype extension="scf" type="galaxy.datatypes.images:Scf" mimetype="application/octet-stream" display_in_upload="true"/>
     <datatype extension="taxonomy" type="galaxy.datatypes.tabular:Taxonomy" display_in_upload="true"/>
     <datatype extension="tabular" type="galaxy.datatypes.tabular:Tabular" display_in_upload="true"/>
@@ -205,5 +207,6 @@
     <sniffer type="galaxy.datatypes.interval:Gff"/>
     <sniffer type="galaxy.datatypes.interval:Gff3"/>
     <sniffer type="galaxy.datatypes.interval:Interval"/>
+    <sniffer type="galaxy.datatypes.tabular:Sam"/>
   </sniffers>
 </datatypes>
diff -r b56291fad13d -r 670f8800a2bf lib/galaxy/datatypes/images.py
--- a/lib/galaxy/datatypes/images.py	Mon Aug 17 09:54:57 2009 -0400
+++ b/lib/galaxy/datatypes/images.py	Wed Aug 19 09:13:21 2009 -0400
@@ -4,6 +4,8 @@
 
 import data
 import logging
+from galaxy.datatypes.metadata import MetadataElement
+from galaxy.datatypes import metadata
 from galaxy.datatypes.sniff import *
 from urllib import urlencode, quote_plus
 import zipfile
@@ -187,7 +189,7 @@
         return 'text/html'
     def sniff( self, filename ):
         """
-        Determines wether the file is in html format
+        Determines whether the file is in html format
 
         >>> fname = get_test_fname( 'complete.bed' )
         >>> Html().sniff( fname )
@@ -233,3 +235,25 @@
             return dataset.peek
         except:
             return "peek unavailable"
+
+class Bam( data.Binary ):
+    """Class describing a BAM binary file"""
+    file_ext = "bam"
+    MetadataElement( name="bam_index", desc="BAM Index File", param=metadata.FileParameter, readonly=True, no_value=None, visible=False, optional=True )    
+    def set_peek( self, dataset ):
+        if not dataset.dataset.purged:
+            export_url = "/history_add_to?" + urlencode({'history_id':dataset.history_id,'ext':'bam','name':'bam alignments','info':'Alignments file','dbkey':dataset.dbkey})
+            dataset.peek  = "Binary bam alignments file" 
+            dataset.blurb = data.nice_size( dataset.get_size() )
+        else:
+            dataset.peek = 'file does not exist'
+            dataset.blurb = 'file purged from disk'
+    def display_peek(self, dataset):
+        try:
+            return dataset.peek
+        except:
+            return "Binary bam alignments file (%s)" % ( data.nice_size( dataset.get_size() ) )
+    def get_mime(self):
+        """Returns the mime type of the datatype"""
+        return 'application/octet-stream'
+        
\ No newline at end of file
diff -r b56291fad13d -r 670f8800a2bf lib/galaxy/datatypes/registry.py
--- a/lib/galaxy/datatypes/registry.py	Mon Aug 17 09:54:57 2009 -0400
+++ b/lib/galaxy/datatypes/registry.py	Wed Aug 19 09:13:21 2009 -0400
@@ -111,6 +111,7 @@
             self.datatypes_by_extension = { 
                 'ab1'         : images.Ab1(),
                 'axt'         : sequence.Axt(),
+                'bam'         : images.Bam(),
                 'bed'         : interval.Bed(), 
                 'binseq.zip'  : images.Binseq(),
                 'blastxml'    : xml.BlastXml(),
@@ -130,6 +131,7 @@
                 'qualsolid'   : qualityscore.QualityScoreSOLiD(),
                 'qualsolexa'  : qualityscore.QualityScoreSolexa(),
                 'qual454'     : qualityscore.QualityScore454(),
+                'sam'         : tabular.Sam(), 
                 'scf'         : images.Scf(),
                 'tabular'     : tabular.Tabular(),
                 'taxonomy'    : tabular.Taxonomy(),
@@ -140,6 +142,7 @@
             self.mimetypes_by_extension = { 
                 'ab1'         : 'application/octet-stream',
                 'axt'         : 'text/plain',
+                'bam'         : 'application/octet-stream',
                 'bed'         : 'text/plain', 
                 'binseq.zip'  : 'application/zip',
                 'blastxml'    : 'text/plain', 
@@ -157,6 +160,7 @@
                 'qualsolid'   : 'text/plain',
                 'qualsolexa'  : 'text/plain',
                 'qual454'     : 'text/plain',
+                'sam'         : 'text/plain',
                 'scf'         : 'application/octet-stream',
                 'tabular'     : 'text/plain',
                 'taxonomy'    : 'text/plain',
@@ -184,7 +188,8 @@
                 interval.CustomTrack(),
                 interval.Gff(),
                 interval.Gff3(),
-                interval.Interval()
+                interval.Interval(),
+                tabular.Sam()
             ]
         def append_to_sniff_order():
             # Just in case any supported data types are not included in the config's sniff_order section.
diff -r b56291fad13d -r 670f8800a2bf lib/galaxy/datatypes/tabular.py
--- a/lib/galaxy/datatypes/tabular.py	Mon Aug 17 09:54:57 2009 -0400
+++ b/lib/galaxy/datatypes/tabular.py	Wed Aug 19 09:13:21 2009 -0400
@@ -236,3 +236,84 @@
             out = "Can't create peek %s" % exc
         return out
 
+class Sam( Tabular ):
+    file_ext = 'sam'
+    def __init__(self, **kwd):
+        """Initialize taxonomy datatype"""
+        Tabular.__init__( self, **kwd )
+        self.column_names = ['QNAME', 'FLAG', 'RNAME', 'POS', 'MAPQ', 'CIGAR',
+                             'MRNM', 'MPOS', 'ISIZE', 'SEQ', 'QUAL', 'OPT'
+                             ]
+    def make_html_table( self, dataset, skipchars=[] ):
+        """Create HTML table, used for displaying peek"""
+        out = ['<table cellspacing="0" cellpadding="3">']
+        try:
+            # Generate column header
+            out.append( '<tr>' )
+            for i, name in enumerate( self.column_names ):
+                out.append( '<th>%s.%s</th>' % ( str( i+1 ), name ) )
+            # This data type requires at least 11 columns in the data
+            if dataset.metadata.columns - len( self.column_names ) > 0:
+                for i in range( len( self.column_names ), dataset.metadata.columns ):
+                    out.append( '<th>%s</th>' % str( i+1 ) )
+                out.append( '</tr>' )
+            out.append( self.make_html_peek_rows( dataset, skipchars=skipchars ) )
+            out.append( '</table>' )
+            out = "".join( out )
+        except Exception, exc:
+            out = "Can't create peek %s" % exc
+        return out
+    def sniff( self, filename ):
+        """
+        Determines whether the file is in SAM format
+        
+        A file in SAM format consists of lines of tab-separated data.
+        The following header line may be the first line:
+        @QNAME  FLAG    RNAME   POS     MAPQ    CIGAR   MRNM    MPOS    ISIZE   SEQ     QUAL
+        or
+        @QNAME  FLAG    RNAME   POS     MAPQ    CIGAR   MRNM    MPOS    ISIZE   SEQ     QUAL    OPT
+        Data in the OPT column is optional and can consist of tab-separated data
+
+        For complete details see http://samtools.sourceforge.net/SAM1.pdf
+        
+        Rules for sniffing as True:
+            There must be 11 or more columns of data on each line
+            Columns 2 (FLAG), 4(POS), 5 (MAPQ), 8 (MPOS), and 9 (ISIZE) must be numbers (9 can be negative)
+            We will only check that up to the first 5 alignments are correctly formatted.
+        
+        >>> fname = get_test_fname( 'sequence.maf' )
+        >>> Sam().sniff( fname )
+        False
+        >>> fname = get_test_fname( '1.sam' )
+        >>> Sam().sniff( fname )
+        True
+        """
+        try:
+            fh = open( filename )
+            count = 0
+            while True:
+                line = fh.readline()
+                line = line.strip()
+                if not line:
+                    break #EOF
+                if line: 
+                    if line[0] != '@':
+                        linePieces = line.split('\t')
+                        if len(linePieces) < 11:
+                            return False
+                        try:
+                            check = int(linePieces[1])
+                            check = int(linePieces[3])
+                            check = int(linePieces[4])
+                            check = int(linePieces[7])
+                            check = int(linePieces[8])
+                        except ValueError:
+                            return False
+                        count += 1
+                        if count == 5:
+                            return True
+            if count < 5 and count > 0:
+                return True
+        except:
+            pass
+        return False
diff -r b56291fad13d -r 670f8800a2bf lib/galaxy/datatypes/test/1.sam
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/lib/galaxy/datatypes/test/1.sam	Wed Aug 19 09:13:21 2009 -0400
@@ -0,0 +1,97 @@
+@QNAME	FLAG	RNAME	POS	MAPQ	CIGAR	MRNM	MPOS	ISIZE	SEQ	QUAL	OPT
+1378_11_329	69	*	0	0	*	*	0	0	AGACCGGGCGGGGTGGCGTTCGGT	%##+'#######%###$#$##$(#
+1378_11_329	133	*	0	0	*	*	0	0	GTTCGTGGCCGGTGGGTGTTTGGG	###$$#$#$&#####$'$#$###$
+1378_17_1788	69	*	0	0	*	*	0	0	TGCCGTGTCTTGCTAACGCCGATT	#'#$$#$###%%##$$$$######
+1378_17_1788	133	*	0	0	*	*	0	0	TGGGTGGATGTGTTGTCGTTCATG	#$#$###$#$#######$#$####
+1378_25_2035	69	*	0	0	*	*	0	0	CTGCGTGTTGGTGTCTACTGGGGT	#%#'##$#$##&%#%$$$%#%#'#
+1378_25_2035	133	*	0	0	*	*	0	0	GTGCGTCGGGGAGGGTGCTGTCGG	######%#$%#$$###($###&&%
+1378_28_770	89	chr11.nib:1-134452384	72131356	37	17M1I5M	=	72131356	0	CACACTGTGACAGACAGCGCAGC	00/02!!0//1200210!!44/1	XT:A:U	CM:i:2	SM:i:37	AM:i:0	X0:i:1	X1:i:0	XM:i:1	XO:i:1	XG:i:1	MD:Z:22
+1378_28_770	181	chr11.nib:1-134452384	72131356	0	24M	=	72131356	0	TTGGTGCGCGCGGTTGAGGGTTGG	$$(#%%#$%#%####$%%##$###
+1378_33_1945	113	chr2.nib:1-242951149	181247988	0	23M	chr12.nib:1-132349534	41710908	0	GAGAGAGAGAGAGAGAGAGAGAG	PQRVUMNXYRPUXYXWXSOSZ]M	XT:A:R	CM:i:0	SM:i:0	AM:i:0	X0:i:163148	XM:i:0	XO:i:0	XG:i:0	MD:Z:23
+1378_33_1945	177	chr12.nib:1-132349534	41710908	0	23M	chr2.nib:1-242951149	181247988	0	AGAGAGAGAGAGAGAGAGAGAGA	SQQWZYURVYWX]]YXTSY]]ZM	XT:A:R	CM:i:0	SM:i:0	AM:i:0	X0:i:163148	XM:i:0	XO:i:0	XG:i:0	MD:Z:23
+1378_34_789	69	*	0	0	*	*	0	0	ATGGTGGCTGACGCGTTTGACTGT	#$##%#$##$&$#%##$##$###$
+1378_34_789	133	*	0	0	*	*	0	0	GGGCTTGCGTTAGTGAGAGGTTGT	###%$%$%%###$####$###$#&
+1378_35_263	115	chr16.nib:1-88827254	19671878	0	23M	=	19671877	-1	AGAGAGAGAGAGAGAGAGAGTCT	77543:<55#"4!&=964518A>	XT:A:R	CM:i:2	SM:i:0	AM:i:0	X0:i:4	X1:i:137	XM:i:2	XO:i:0	XG:i:0	MD:Z:23
+1378_35_263	179	chr16.nib:1-88827254	19671877	0	23M	=	19671878	1	GAGAGAGAGAGAGAGAGAGAGTC	LE7402DD34FL:27AKE>;432	XT:A:R	CM:i:0	SM:i:0	AM:i:0	X0:i:265	XM:i:0	XO:i:0	XG:i:0	MD:Z:23
+1378_43_186	69	*	0	0	*	*	0	0	ATACTAGTTGGGACGCGTTGTGCT	#$(4%$########$#$###$$$#
+1378_43_186	133	*	0	0	*	*	0	0	GCTAGGGTTTGGGTTTGCGGTGGG	$%#$########%##%#$###'#'
+1378_51_1671	117	chr2.nib:1-242951149	190342418	0	24M	=	190342418	0	CTGGCGTTCTCGGCGTGGATGGGT	#####$$##$#%#%%###%$#$##
+1378_51_1671	153	chr2.nib:1-242951149	190342418	37	16M1I6M	=	190342418	0	TCTAACTTAGCCTCATAATAGCT	/<<!"0///////00/!!0121/	XT:A:U	CM:i:2	SM:i:37	AM:i:0	X0:i:1	X1:i:0	XM:i:1	XO:i:1	XG:i:1	MD:Z:22
+1378_56_324	117	chr2.nib:1-242951149	80324999	0	24M	=	80324999	0	TCCAGTCGCGTTGTTAGGTTCGGA	#$#$$$#####%##%%###**#+/
+1378_56_324	153	chr2.nib:1-242951149	80324999	37	8M1I14M	=	80324999	0	TTTAGCCCGAAATGCCTAGAGCA	4;6//11!"11100110////00	XT:A:U	CM:i:2	SM:i:37	AM:i:0	X0:i:1	X1:i:0	XM:i:1	XO:i:1	XG:i:1	MD:Z:22
+1378_56_773	69	*	0	0	*	*	0	0	TGTCGTGAGGTCACTTATCCCCAT	&%#%##%%#####&#$%##$%##$
+1378_56_773	133	*	0	0	*	*	0	0	TCTGGTCGGTTTCGGGGAGTGGAA	##%%#&$###$#$##%$####%%$
+1378_62_2027	69	*	0	0	*	*	0	0	CTTCCACGATCTGCTCGCTGTGGT	(#&&$##$$#$%#%$$$#$###'#
+1378_62_2027	133	*	0	0	*	*	0	0	GTTGGCCTGGCCTGCCGTGCTGCG	*##),/%##$)#%##1$#'%.#&#
+1378_62_2029	69	*	0	0	*	*	0	0	TCTGGGCTGTCTTCGGGTCGGTGT	$%$$####$##$$#)##%%#$###
+1378_62_2029	133	*	0	0	*	*	0	0	GGCGGTGTGTGGTGCGGCTGTGCG	/$$$=(####%####)$$%$-&%#
+1378_67_1795	81	chr16.nib:1-88827254	26739130	0	23M	chrY.nib:1-57772954	57401793	0	TGGCATTCCTGTAGGCAGAGAGG	AZWWZS]!"QNXZ]VQ]]]/2]]	XT:A:R	CM:i:2	SM:i:0	AM:i:0	X0:i:3	X1:i:0	XM:i:2	XO:i:0	XG:i:0	MD:Z:23
+1378_67_1795	161	chrY.nib:1-57772954	57401793	37	23M	chr16.nib:1-88827254	26739130	0	GATCACCCAGGTGATGTAACTCC	]WV]]]]WW]]]]]]]]]]PU]]	XT:A:U	CM:i:0	SM:i:37	AM:i:0	X0:i:1	X1:i:0	XM:i:0	XO:i:0	XG:i:0	MD:Z:23
+1378_68_466	69	*	0	0	*	*	0	0	GTGATCGTCGGTGCCAGTCCCTGT	#(%)+##$#$#%#+$%##$#####
+1378_68_466	133	*	0	0	*	*	0	0	GTGTCATCTGAGGTAAAGCATTGT	/##$09#$#.=$#$76+$%1'###
+1378_68_1692	117	chr13.nib:1-114142980	36365609	0	24M	=	36365609	0	TTGAACCGGGCACGGGTCTTCTGG	#$#######%###$##%&'%)###
+1378_68_1692	153	chr13.nib:1-114142980	36365609	37	10M1D13M	=	36365609	0	CTGCACATACAGAATATTCATAG	0010/!"0/!!021/132231//	XT:A:U	CM:i:2	SM:i:37	AM:i:0	X0:i:1	X1:i:0	XM:i:1	XO:i:1	XG:i:1	MD:Z:10^T13
+1378_80_664	69	*	0	0	*	*	0	0	CTGCTTTGATCCCCGGTGGAGCAC	7#%###$$6#######$##$$$##
+1378_80_664	133	*	0	0	*	*	0	0	TGTCTGCGTTGTATCTCTGGTGTA	%##%,%$$#&$$###$#$%##'%#
+1378_85_1786	69	*	0	0	*	*	0	0	ATACTATGTCGATCTGTAAAAAAA	)&.)#3%@$&%-,2#&+.-%0&./
+1378_85_1786	133	*	0	0	*	*	0	0	CCCTAGGAGCGTATACCGGACGAG	,'&/%/@,&1,&'/)&,6&&1)((
+1378_86_1011	69	*	0	0	*	*	0	0	CTACGTTATTGCTCTGTTTGTCCT	######$%##$$$%###%#$####
+1378_86_1011	133	*	0	0	*	*	0	0	AGGCGATGGGATATTATTTTACTT	:$###)%##$9$###1$$#$2###
+1378_86_1789	89	chr12.nib:1-132349534	39007065	37	23M	=	39007065	0	GCTTTCCATAGATGTGTAATTTC	J2K]]Z5!GN?@U]]]VX]UYYP	XT:A:U	CM:i:1	SM:i:37	AM:i:0	X0:i:1	X1:i:0	XM:i:1	XO:i:0	XG:i:0	MD:Z:23
+1378_86_1789	181	chr12.nib:1-132349534	39007065	0	24M	=	39007065	0	ACAACTTAAATAATCATGGACCGG	02,5$$0&6#%?*,$'#%&/15.1
+1378_91_1596	69	*	0	0	*	*	0	0	TTAGCGGTTGACTATCTGCTGACA	*&+'#9'(%*'#//,&<),/)'*#
+1378_91_1596	133	*	0	0	*	*	0	0	GCTTTTTCATTCGGTGCCTTTGGA	'>%/3%=()8'#.%?50$&5>%)%
+1378_94_1595	69	chr7.nib:1-158821424	127518258	0	24M	=	127518258	0	CGTGCGACAGCCCATGTTTTCAGA	-=..5,3826&*+.+#+#%%6;%#
+1378_94_1595	137	chr7.nib:1-158821424	127518258	37	23M	=	127518258	0	TGAGATAAACACCTAACATGCTC	M]]FN]]\V]]]Q>T]KIG:LVN	XT:A:U	CM:i:0	SM:i:37	AM:i:0	X0:i:1	X1:i:0	XM:i:0	XO:i:0	XG:i:0	MD:Z:23
+1378_95_1039	69	*	0	0	*	*	0	0	CGGCGTCCATCTTCGCCTTGAGAT	$##.#$##$$#%$#$%%$###$)$
+1378_95_1039	133	*	0	0	*	*	0	0	GTTCTGTGCCAGGTGAGGTACGGA	&##,./#$&)6##+,'#$$0(##$
+1378_95_1767	65	chr11.nib:1-134452384	65333552	25	23M	chr3.nib:1-199501827	123725482	0	CAACTGGTGGCATCTGGACAAAC	W[[TZYY]]RO<BI7!!:!!>@2	XT:A:U	CM:i:2	SM:i:25	AM:i:25	X0:i:1	X1:i:0	XM:i:2	XO:i:0	XG:i:0	MD:Z:23
+1378_95_1767	129	chr3.nib:1-199501827	123725482	37	6M1I16M	chr11.nib:1-134452384	65333552	0	ATTTATCTGTCTCATTCATTATT	<AGB8B"!V]]UO/&JB4DE88E	XT:A:U	CM:i:2	SM:i:37	AM:i:25	X0:i:1	X1:i:0	XM:i:1	XO:i:1	XG:i:1	MD:Z:22
+1378_96_1037	69	*	0	0	*	*	0	0	ATCCCCCAAGATGCCTGTTGATTG	$#$'##$$$#%$$#%###+##$#$
+1378_96_1037	133	*	0	0	*	*	0	0	CTGCTGGGCCATTTGACTTACTCA	'$#+#(##-%5##+*&###-.$$$
+1378_96_1764	81	chr15.nib:1-100338915	89251272	25	23M	chr7.nib:1-158821424	19412615	0	AGAAATGGTCGCACCCTCTGGTT	E*2ZEHX\SN]O>SYRL):LIOL	XT:A:U	CM:i:2	SM:i:25	AM:i:25	X0:i:1	X1:i:0	XM:i:2	XO:i:0	XG:i:0	MD:Z:23
+1378_96_1764	161	chr7.nib:1-158821424	19412615	37	23M	chr15.nib:1-100338915	89251272	0	GTATAGCCCACAACGCCTAATAT	ZMBS]UW]UYR\]QPZ[SMYL7C	XT:A:U	CM:i:0	SM:i:37	AM:i:25	X0:i:1	X1:i:0	XM:i:0	XO:i:0	XG:i:0	MD:Z:23
+1378_98_1574	69	*	0	0	*	*	0	0	GTTCTGCCGGTGTCTGTGGCGGGC	$$#+&$$####%$$$###$%#%%#
+1378_98_1574	133	*	0	0	*	*	0	0	AGGCGAGTGTGGGGGTTGTTTGAG	+%%$#)##%##$####%###$%$#
+1378_107_1647	69	*	0	0	*	*	0	0	AGGCCTACTACGCGTCATTGATAG	&#$$#$(.#%#$$####&$%##($
+1378_107_1647	133	*	0	0	*	*	0	0	GGTCTGGTTCTATGTTGGTCGACT	###'$$#$$$(#%###(#$##$%#
+1378_111_829	69	chr9.nib:1-140273252	82506894	0	24M	=	82506894	0	TGCGGCACTTGCTTCTTCGTATTT	%#%##%#$%#$#%###$$##&#$$
+1378_111_829	137	chr9.nib:1-140273252	82506894	37	4M1I18M	=	82506894	0	GATGCGTAATCTAGTAAAATAAG	0/362//00/5516500210451	XT:A:U	CM:i:2	SM:i:37	AM:i:0	X0:i:1	X1:i:0	XM:i:1	XO:i:1	XG:i:1	MD:Z:22
+1378_111_1900	69	*	0	0	*	*	0	0	TCCCCTCGCTCGGCTCTGTGCTGT	$&%*$#(#)##$#'##%(##$#$%
+1378_111_1900	133	*	0	0	*	*	0	0	GCACGCCTTTGGGCTAAGCCGTAA	)$)'#%$########$'#&%$#(#
+1378_112_1483	69	*	0	0	*	*	0	0	TGTCCAGCTATGCGGCTTCCTCCT	%#$+#%#&#$#####%####%$##
+1378_112_1483	133	*	0	0	*	*	0	0	TGGAGTGGTGTGTTTGCTGAGCCA	#$#)#############$#%#%'%
+1378_125_1287	69	*	0	0	*	*	0	0	TGTCTCTGGGGGGCCTGGTTAGGT	$##13$'%#$###$$###$$$#&#
+1378_125_1287	133	*	0	0	*	*	0	0	TGACGTGGGTTGTCCCGTGAGATT	##$%%#$###$##$$#&%##$(%%
+1378_126_468	117	chr11.nib:1-134452384	72541052	0	24M	=	72541052	0	TGCCTCTATACAGATTAGTCCTCT	)7,7..?97594@8=,=?813@>7
+1378_126_468	153	chr11.nib:1-134452384	72541052	0	23M	=	72541052	0	AGGCAAGACTCTGTCTCAAAAAA	PK5G]]PDT\]SEXY[]]]]]]]	XT:A:R	CM:i:0	SM:i:0	AM:i:0	X0:i:4	X1:i:15713	XM:i:0	XO:i:0	XG:i:0	MD:Z:23
+1378_127_664	69	*	0	0	*	*	0	0	AGAGGTTGGTGTCTTGTCGCAGCT	##'#$######$$%######$$$#
+1378_127_664	133	*	0	0	*	*	0	0	TCGCTTTGCCTATGTTTGTTCGGA	#%$%#&##$%#%%###$$###)-'
+1378_129_463	97	chr8.nib:1-146274826	29931771	37	23M	chr19.nib:1-63811651	5702213	0	GTAGCTCTGTTTCACATTAGGGG	J>AQ[G>C?NM:GD=)*PLORIF	XT:A:U	CM:i:1	SM:i:37	AM:i:0	X0:i:1	X1:i:0	XM:i:1	XO:i:0	XG:i:0	MD:Z:23
+1378_129_463	145	chr19.nib:1-63811651	5702213	0	23M	chr8.nib:1-146274826	29931771	0	AAAAAAAAAAAAAAAAAAAAAAA	JOI:AHGD==@KQB78HF>KA8>	XT:A:R	CM:i:0	SM:i:0	AM:i:0	X0:i:583698	XM:i:0	XO:i:0	XG:i:0	MD:Z:23
+1378_129_875	69	*	0	0	*	*	0	0	TTTCTATGGCTTACGCTGTCTGCC	#$($##%####%$#$#####$###
+1378_129_875	133	*	0	0	*	*	0	0	GACCTTTACGTATTGGGGGTTGGC	###)###+###$##$#&%##$,#$
+1378_140_1251	69	*	0	0	*	*	0	0	ATCCTAGCGCGGTGTCTTGGGGAC	#$%1#$$$##$##$#$#$##$%$$
+1378_140_1251	133	*	0	0	*	*	0	0	TTTCCTTCGTGTGCGTGCGGAGTG	#%#%$##$$$######.$$$%#%(
+1378_141_809	69	*	0	0	*	*	0	0	TGTCCTCCAGTGTCTGTTGGGTGT	%&,-##$$#(%###$#$$'###'#
+1378_141_809	133	*	0	0	*	*	0	0	TCTCGTGGTTTCTTTTTTATGTGT	##%)##$$#####%$#$#%%#'##
+1378_144_983	69	*	0	0	*	*	0	0	AGCGCCCGGTTGGTGCGGCTCGTC	-$(&%*$#*#))#$$$#%%$#$##
+1378_144_983	133	*	0	0	*	*	0	0	GTTCGTTCGTGGTGTACGAGGGTG	#(#%#####($#%##$$#%##%#)
+1378_153_270	69	*	0	0	*	*	0	0	AGTCCTTGTCCCCTGGGTTTTCCC	+''$#&%$%#$##&$$($#&#$$#
+1378_153_270	133	*	0	0	*	*	0	0	GGCCGTGTGCGGGTGTAGATTGGA	%$##($######&##$&$$$$%##
+1378_155_1689	65	chrX.nib:1-154913754	106941539	37	23M	=	106940385	-1154	ATCTCCTCTTCCTTCCATTCCAC	\]]]Y]]]]]UV]]]ZYZZ]]RV	XT:A:U	CM:i:0	SM:i:37	AM:i:37	X0:i:1	X1:i:0	XM:i:0	XO:i:0	XG:i:0	MD:Z:23
+1378_155_1689	129	chrX.nib:1-154913754	106940385	37	23M	=	106941539	1154	GACTATGAGGTTTTCATTCAACA	]]]]\\]]]YW]]]WRZ]]WIOK	XT:A:U	CM:i:0	SM:i:37	AM:i:37	X0:i:1	X1:i:0	XM:i:0	XO:i:0	XG:i:0	MD:Z:23
+1378_157_1580	69	*	0	0	*	*	0	0	TGGGCCTCGGTGCCCTTGGTCTGT	#%)$##'#$$$&#####%#$#$##
+1378_157_1580	133	*	0	0	*	*	0	0	GGGATTGAAGGGATGTATGCTAGG	#%$&%#$$'%$%#$##*#%$$$$#
+1378_161_317	69	*	0	0	*	*	0	0	TTGGCCGGCAACCCCGGTACCTAA	7<,<'@)@>.)2@/')'&(?/-<(
+1378_161_317	133	*	0	0	*	*	0	0	AATCCATACCCACAAAAGCAGGCC	.&%','(@''?7//+&)+2.+)0)
+1378_177_735	113	chr2.nib:1-242951149	222173182	25	23M	=	222173882	700	TTGTTCAGCGCCGATTGTCAATC	KPNICFMS]]]Z]]]]Y]]]]]]	XT:A:U	CM:i:2	SM:i:25	AM:i:25	X0:i:1	X1:i:0	XM:i:2	XO:i:0	XG:i:0	MD:Z:1G21
+1378_177_735	177	chr2.nib:1-242951149	222173882	37	23M	=	222173182	-700	AGAATTCCTAACAAAATGTGAAG	ES6-]]]]]]]]]]]]]]]]]]]	XT:A:U	CM:i:1	SM:i:37	AM:i:25	X0:i:1	X1:i:0	XM:i:1	XO:i:0	XG:i:0	MD:Z:23
+1378_181_1684	69	*	0	0	*	*	0	0	CGACTCCCGCATTCACGGTCAAGT	&*#,##$#&$*$$#$#$$$#%$##
+1378_181_1684	133	*	0	0	*	*	0	0	TTTCTGTTGTGGTTTTGTTGGGGT	$##'$%'##%##$%$#$$####$*
+1378_187_1407	69	*	0	0	*	*	0	0	TGGCGTCCACTCGTGGGTCTATCG	$#$'%#$%$%&$%#####$#$#%#
+1378_187_1407	133	*	0	0	*	*	0	0	TTGGGTGAAATCTTGTCGAGTGGA	####&##$$###$#####%##%%)
+1378_203_721	97	chr1.nib:1-247249719	245680524	25	23M	chr2.nib:1-242951149	213173999	0	GTAAAATTTGTGGAGATTTAAGT	]VEFFEZ]XPW]TOVINQ,;T!!	XT:A:U	CM:i:2	SM:i:25	AM:i:25	X0:i:1	X1:i:0	XM:i:2	XO:i:0	XG:i:0	MD:Z:23
+1378_203_721	145	chr2.nib:1-242951149	213173999	37	4M1I18M	chr1.nib:1-247249719	245680524	0	ACCTAACAAAATTGTTCAATATG	F>8AWT<AV]Q9B"+]O@IF=K]	XT:A:U	CM:i:2	SM:i:37	AM:i:25	X0:i:1	X1:i:0	XM:i:1	XO:i:1	XG:i:1	MD:Z:22
+1378_206_2039	113	chr4.nib:1-191273063	103793427	0	23M	chr18.nib:1-76117153	57165542	0	ACACACACACACACACACACACA	NKWZVWZ]]XV[]]]]]]]]]]]	XT:A:R	CM:i:0	SM:i:0	AM:i:0	X0:i:1292040	XM:i:0	XO:i:0	XG:i:0	MD:Z:23
+1378_206_2039	177	chr18.nib:1-76117153	57165542	0	23M	chr4.nib:1-191273063	103793427	0	CACACACACACACACACACACAC	NAJ[SPT[]]]W[]]]]]]]]]]	XT:A:R	CM:i:0	SM:i:0	AM:i:0	X0:i:1292040	XM:i:0	XO:i:0	XG:i:0	MD:Z:23
diff -r b56291fad13d -r 670f8800a2bf test-data/1.sam
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/1.sam	Wed Aug 19 09:13:21 2009 -0400
@@ -0,0 +1,29 @@
+@QNAME	FLAG	RNAME	POS	MAPQ	CIGAR	MRNM	MPOS	ISIZE	SEQ	QUAL	OPT
+1378_11_329	69	*	0	0	*	*	0	0	AGACCGGGCGGGGTGGCGTTCGGT	%##+'#######%###$#$##$(#
+1378_11_329	133	*	0	0	*	*	0	0	GTTCGTGGCCGGTGGGTGTTTGGG	###$$#$#$&#####$'$#$###$
+1378_17_1788	69	*	0	0	*	*	0	0	TGCCGTGTCTTGCTAACGCCGATT	#'#$$#$###%%##$$$$######
+1378_17_1788	133	*	0	0	*	*	0	0	TGGGTGGATGTGTTGTCGTTCATG	#$#$###$#$#######$#$####
+1378_25_2035	69	*	0	0	*	*	0	0	CTGCGTGTTGGTGTCTACTGGGGT	#%#'##$#$##&%#%$$$%#%#'#
+1378_25_2035	133	*	0	0	*	*	0	0	GTGCGTCGGGGAGGGTGCTGTCGG	######%#$%#$$###($###&&%
+1378_28_770	89	chr11.nib:1-134452384	72131356	37	17M1I5M	=	72131356	0	CACACTGTGACAGACAGCGCAGC	00/02!!0//1200210!!44/1	XT:A:U	CM:i:2	SM:i:37	AM:i:0	X0:i:1	X1:i:0	XM:i:1	XO:i:1	XG:i:1	MD:Z:22
+1378_28_770	181	chr11.nib:1-134452384	72131356	0	24M	=	72131356	0	TTGGTGCGCGCGGTTGAGGGTTGG	$$(#%%#$%#%####$%%##$###
+1378_33_1945	113	chr2.nib:1-242951149	181247988	0	23M	chr12.nib:1-132349534	41710908	0	GAGAGAGAGAGAGAGAGAGAGAG	PQRVUMNXYRPUXYXWXSOSZ]M	XT:A:R	CM:i:0	SM:i:0	AM:i:0	X0:i:163148	XM:i:0	XO:i:0	XG:i:0	MD:Z:23
+1378_33_1945	177	chr12.nib:1-132349534	41710908	0	23M	chr2.nib:1-242951149	181247988	0	AGAGAGAGAGAGAGAGAGAGAGA	SQQWZYURVYWX]]YXTSY]]ZM	XT:A:R	CM:i:0	SM:i:0	AM:i:0	X0:i:163148	XM:i:0	XO:i:0	XG:i:0	MD:Z:23
+1378_34_789	69	*	0	0	*	*	0	0	ATGGTGGCTGACGCGTTTGACTGT	#$##%#$##$&$#%##$##$###$
+1378_34_789	133	*	0	0	*	*	0	0	GGGCTTGCGTTAGTGAGAGGTTGT	###%$%$%%###$####$###$#&
+1378_35_263	115	chr16.nib:1-88827254	19671878	0	23M	=	19671877	-1	AGAGAGAGAGAGAGAGAGAGTCT	77543:<55#"4!&=964518A>	XT:A:R	CM:i:2	SM:i:0	AM:i:0	X0:i:4	X1:i:137	XM:i:2	XO:i:0	XG:i:0	MD:Z:23
+1378_35_263	179	chr16.nib:1-88827254	19671877	0	23M	=	19671878	1	GAGAGAGAGAGAGAGAGAGAGTC	LE7402DD34FL:27AKE>;432	XT:A:R	CM:i:0	SM:i:0	AM:i:0	X0:i:265	XM:i:0	XO:i:0	XG:i:0	MD:Z:23
+1378_43_186	69	*	0	0	*	*	0	0	ATACTAGTTGGGACGCGTTGTGCT	#$(4%$########$#$###$$$#
+1378_43_186	133	*	0	0	*	*	0	0	GCTAGGGTTTGGGTTTGCGGTGGG	$%#$########%##%#$###'#'
+1378_51_1671	117	chr2.nib:1-242951149	190342418	0	24M	=	190342418	0	CTGGCGTTCTCGGCGTGGATGGGT	#####$$##$#%#%%###%$#$##
+1378_51_1671	153	chr2.nib:1-242951149	190342418	37	16M1I6M	=	190342418	0	TCTAACTTAGCCTCATAATAGCT	/<<!"0///////00/!!0121/	XT:A:U	CM:i:2	SM:i:37	AM:i:0	X0:i:1	X1:i:0	XM:i:1	XO:i:1	XG:i:1	MD:Z:22
+1378_56_324	117	chr2.nib:1-242951149	80324999	0	24M	=	80324999	0	TCCAGTCGCGTTGTTAGGTTCGGA	#$#$$$#####%##%%###**#+/
+1378_56_324	153	chr2.nib:1-242951149	80324999	37	8M1I14M	=	80324999	0	TTTAGCCCGAAATGCCTAGAGCA	4;6//11!"11100110////00	XT:A:U	CM:i:2	SM:i:37	AM:i:0	X0:i:1	X1:i:0	XM:i:1	XO:i:1	XG:i:1	MD:Z:22
+1378_56_773	69	*	0	0	*	*	0	0	TGTCGTGAGGTCACTTATCCCCAT	&%#%##%%#####&#$%##$%##$
+1378_56_773	133	*	0	0	*	*	0	0	TCTGGTCGGTTTCGGGGAGTGGAA	##%%#&$###$#$##%$####%%$
+1378_62_2027	69	*	0	0	*	*	0	0	CTTCCACGATCTGCTCGCTGTGGT	(#&&$##$$#$%#%$$$#$###'#
+1378_62_2027	133	*	0	0	*	*	0	0	GTTGGCCTGGCCTGCCGTGCTGCG	*##),/%##$)#%##1$#'%.#&#
+1378_62_2029	69	*	0	0	*	*	0	0	TCTGGGCTGTCTTCGGGTCGGTGT	$%$$####$##$$#)##%%#$###
+1378_62_2029	133	*	0	0	*	*	0	0	GGCGGTGTGTGGTGCGGCTGTGCG	/$$$=(####%####)$$%$-&%#
+1378_67_1795	81	chr16.nib:1-88827254	26739130	0	23M	chrY.nib:1-57772954	57401793	0	TGGCATTCCTGTAGGCAGAGAGG	AZWWZS]!"QNXZ]VQ]]]/2]]	XT:A:R	CM:i:2	SM:i:0	AM:i:0	X0:i:3	X1:i:0	XM:i:2	XO:i:0	XG:i:0	MD:Z:23
+1378_67_1795	161	chrY.nib:1-57772954	57401793	37	23M	chr16.nib:1-88827254	26739130	0	GATCACCCAGGTGATGTAACTCC	]WV]]]]WW]]]]]]]]]]PU]]	XT:A:U	CM:i:0	SM:i:37	AM:i:0	X0:i:1	X1:i:0	XM:i:0	XO:i:0	XG:i:0	MD:Z:23
\ No newline at end of file
diff -r b56291fad13d -r 670f8800a2bf test/functional/test_sniffing_and_metadata_settings.py
--- a/test/functional/test_sniffing_and_metadata_settings.py	Mon Aug 17 09:54:57 2009 -0400
+++ b/test/functional/test_sniffing_and_metadata_settings.py	Wed Aug 19 09:13:21 2009 -0400
@@ -226,6 +226,16 @@
         assert latest_hda is not None, "Problem retrieving fastqsanger hda from the database"
         if not latest_hda.name == '1.fastqsanger' and not latest_hda.extension == 'fastqsanger':
             raise AssertionError, "fastqsanger data type was not correctly sniffed."
+    def test_090_sam_datatype( self ):
+        """Testing correctly sniffing sam format upon upload"""
+        self.upload_file( '1.sam' )
+        self.verify_dataset_correctness( '1.sam' )
+        self.check_history_for_string( '1.sam format: <span class="sam">sam</span>, database: \? Info: uploaded sam file' )
+        latest_hda = galaxy.model.HistoryDatasetAssociation.query() \
+            .order_by( desc( galaxy.model.HistoryDatasetAssociation.table.c.create_time ) ).first()
+        assert latest_hda is not None, "Problem retrieving sam hda from the database"
+        if not latest_hda.name == '1.sam' and not latest_hda.extension == 'sam':
+            raise AssertionError, "sam data type was not correctly sniffed."
     def test_9999_clean_up( self ):
         self.delete_history( id=self.security.encode_id( history1.id ) )
         self.logout()

    

[galaxy-dev] [hg] galaxy 2568: Added two new datatypes (sam and bam) and an a...

Greg Von Kuster