Mercurial > repos > fubar > data_manager_rnasta

changeset 4:102bdfdda10b draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author fubar
date Thu, 08 Jan 2015 17:53:14 -0500
parents 2468ef56337a
children 503097df1670
files data_manager/rnastar_index_builder.py data_manager/rnastar_index_builder.xml data_manager_conf.xml tool_data_table_conf.xml.sample
diffstat 4 files changed, 63 insertions(+), 45 deletions(-) [+]
line wrap: on
line diff
--- a/data_manager/rnastar_index_builder.py	Fri Oct 03 21:59:55 2014 -0400
+++ b/data_manager/rnastar_index_builder.py	Thu Jan 08 17:53:14 2015 -0500
@@ -11,7 +11,8 @@
 
 from json import loads, dumps
 
-DEFAULT_DATA_TABLE_NAME = "rnastar_indexes"
+DEFAULT_DATA_TABLE_NAME = "rnastar_index"
+CHUNK_SIZE = 2**20
 
 def get_id_name( params, dbkey, fasta_description=None):
     #TODO: ensure sequence_id is unique and does not already appear in location file
@@ -26,71 +27,85 @@
             sequence_name = dbkey
     return sequence_id, sequence_name
 
-def build_rnastar_index( data_manager_dict, fasta_filename, target_directory, dbkey, sequence_id, sequence_name, data_table_name,
+
+def _run_command( command, target_directory ):
+    tmp_stderr = tempfile.NamedTemporaryFile( prefix = "tmp-data-manager-rnastar_index_builder-stderr" )
+    return_code = subprocess.call( command, shell=False, stdout=sys.stdout, stderr=tmp_stderr.fileno() )
+    if return_code:
+        tmp_stderr.flush()
+        tmp_stderr.seek( 0 )
+        sys.stderr.write( "### star reports an error:\n" )
+        while True:
+            chunk = tmp_stderr.read( CHUNK_SIZE )
+            if not chunk:
+                break
+            sys.stderr.write( chunk )
+        sys.exit( return_code )
+    tmp_stderr.close()
+    return return_code
+
+
+def build_rnastar_index( data_manager_dict, fasta_filename,  target_directory, 
+    dbkey, sequence_id, sequence_name, data_table_name,
     sjdbOverhang,sjdbGTFfile, sjdbFileChrStartEnd,sjdbGTFtagExonParentTranscript,sjdbGTFfeatureExon,sjdbGTFchrPrefix,n_threads):
     #TODO: allow multiple FASTA input files
-    #tmp_dir = tempfile.mkdtemp( prefix='tmp-data-manager-bwa-index-builder-' )
-    fasta_base_name = os.path.split( fasta_filename )[-1]
-    sym_linked_fasta_filename = os.path.join( target_directory, fasta_base_name )
-    os.symlink( fasta_filename, sym_linked_fasta_filename )
-    pdict={'target_directory':target_directory,'n_threads':n_threads, 'sjdbFileChrStartEnd':sjdbFileChrStartEnd,
-           'sjdbGTFtagExonParentTranscript':sjdbGTFtagExonParentTranscript, 'sjdbGTFfeatureExon':sjdbGTFfeatureExon,
-           'sjdbGTFchrPrefix':sjdbGTFchrPrefix,'sjdbOverhang':sjdbOverhang, 'sjdbGTFfile':sjdbGTFfile,
-           'sym_linked_fasta_filename':sym_linked_fasta_filename}
+    fasta_base_name = os.path.basename( fasta_filename )
+    #sym_linked_fasta_filename = os.path.join( tmp_dir, fasta_base_name )
+    #os.symlink( fasta_filename, sym_linked_fasta_filename )
+    #print >> sys.stdout,'made',sym_linked_fasta_filename
+    cl = ['STAR','--runMode', 'genomeGenerate', '--genomeFastaFiles', fasta_filename, '--genomeDir', target_directory, '--runThreadN',n_threads ]
     
-    cl = 'STAR --runMode genomeGenerate --genomeFastaFiles %(sym_linked_fasta_filename)s --genomeDir %(target_directory)s --runThreadN %(n_threads)s' % pdict
     if sjdbGTFfile:
-         cl += '''--sjdbGTFchrPrefix %(sjdbGTFchrPrefix)s --sjdbGTFfeatureExon %(sjdbGTFfeatureExon)s --sjdbOverhang %(sjdbOverhang)s
-   --sjdbGTFfile %(sjdbGTFfile)s --sjdbGTFtagExonParentTranscript %(sjdbGTFtagExonParentTranscript)s''' %  pdict
+         cl += [ '--sjdbGTFfeatureExon', sjdbGTFfeatureExon,'--sjdbGTFtagExonParentTranscript',sjdbGTFtagExonParentTranscript]
+         if (sjdbGTFchrPrefix > ''):
+             cl += ['--sjdbGTFchrPrefix', sjdbGTFchrPrefix]
+         cl += ['--sjdbOverhang', sjdbOverhang, '--sjdbGTFfile', sjdbGTFfile]
     elif sjdbFileChrStartEnd:
-        cl += '--sjdbFileChrStartEnd %(sjdbFileChrStartEnd)s --sjdbOverhang %(sjdbOverhangs)s' % pdict
-    args = cl.split(' ')
-    proc = subprocess.Popen( args=args, shell=False, cwd=target_directory )
-    return_code = proc.wait()
-    if return_code:
-        sys.exit( return_code )
+        cl += ['--sjdbFileChrStartEnd', sjdbFileChrStartEnd, '--sjdbOverhang', sjdbOverhang]
+    return_code = _run_command(command=cl,target_directory=target_directory)
     data_table_entry = dict( value=sequence_id, dbkey=dbkey, name=sequence_name, path=fasta_base_name )
-    _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry )
-
-def _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry ):
-    data_manager_dict['data_tables'] = data_manager_dict.get( 'data_tables', {} )
-    data_manager_dict['data_tables'][ data_table_name ] = data_manager_dict['data_tables'].get( data_table_name, [] )
+    data_manager_dict.setdefault('data_tables',{})
+    data_manager_dict['data_tables'].setdefault(data_table_name,[])
     data_manager_dict['data_tables'][ data_table_name ].append( data_table_entry )
     return data_manager_dict
 
+
 def main():
     #Parse Command Line
     parser = optparse.OptionParser()
-    parser.add_option( '-f', '--fasta_filename', dest='fasta_filename', action='store', type="string", default=None, help='fasta_filename' )
-    parser.add_option( '-d', '--fasta_dbkey', dest='fasta_dbkey', action='store', type="string", default=None, help='fasta_dbkey' )
-    parser.add_option( '-t', '--fasta_description', dest='fasta_description', action='store', type="string", default=None, help='fasta_description' )
-    parser.add_option( '-n', '--data_table_name', dest='data_table_name', action='store', type="string", default=None, help='data_table_name' )
-    parser.add_option( '--out_file', default=None)
-    parser.add_option( '--out_index_path', default=None)
+    parser.add_option( '--fasta_filename', dest='fasta_filename', action='store', type="string", default=None, help='fasta_filename' )
+    parser.add_option( '--fasta_dbkey', dest='fasta_dbkey', action='store', type="string", default=None, help='fasta_dbkey' )
+    parser.add_option( '--fasta_description', dest='fasta_description', action='store', type="string", default=None, help='fasta_description' )
+    parser.add_option( '--data_table_name', dest='data_table_name', action='store', type="string", default=None, help='data_table_name' )
     parser.add_option( '--sjdbGTFfile', type="string", default=None )
     parser.add_option( '--sjdbGTFchrPrefix', type="string", default=None )
     parser.add_option( '--sjdbGTFfeatureExon', type="string", default=None )
     parser.add_option( '--sjdbGTFtagExonParentTranscript', type="string", default=None )
     parser.add_option( '--sjdbFileChrStartEnd', type="string", default=None )
-    parser.add_option( '--sjdbOverhang', type="int", default=100 )
-    parser.add_option( '--runThreadN', type="int", default=4 )
+    parser.add_option( '--sjdbOverhang', type="string", default='100' )
+    parser.add_option( '--runThreadN', type="string", default='4' )
     (options, args) = parser.parse_args()
-    
-    filename = options.out_file
+    filename = args[0]    
+    # this is passed as an positional parameter in the original wrapper with the extra_files_path from the params
     params = loads( open( filename ).read() )
-    target_directory = options.out_index_path
-    os.mkdirs( target_directory )
+    target_directory = params[ 'output_data' ][0]['extra_files_path'].encode('ascii','replace')
+    try:
+         os.mkdir( target_directory )
+    except OSError:
+        pass
+
     data_manager_dict = {}
     
     dbkey = options.fasta_dbkey
     
     if dbkey in [ None, '', '?' ]:
-        raise Exception( '"%s" is not a valid dbkey. You must specify a valid dbkey.' % ( dbkey ) )
+        raise Exception( 'dbkey "%s" does not specify any locally known genome to be indexed.' % ( dbkey ) )
     
     sequence_id, sequence_name = get_id_name( params, dbkey=dbkey, fasta_description=options.fasta_description )
     
     #build the index
-    build_rnastar_index( data_manager_dict, options.fasta_filename, target_directory, dbkey, sequence_id, sequence_name, data_table_name=options.data_table_name,
+    data_manager_dict = build_rnastar_index( data_manager_dict = data_manager_dict, fasta_filename = options.fasta_filename, target_directory = target_directory,
+      dbkey = dbkey, sequence_id = sequence_id, sequence_name = sequence_name, data_table_name=options.data_table_name,
       sjdbOverhang=options.sjdbOverhang,sjdbGTFfile=options.sjdbGTFfile,
       sjdbFileChrStartEnd=options.sjdbFileChrStartEnd,sjdbGTFtagExonParentTranscript=options.sjdbGTFtagExonParentTranscript,
       sjdbGTFfeatureExon=options.sjdbGTFfeatureExon,sjdbGTFchrPrefix=options.sjdbGTFchrPrefix,
--- a/data_manager/rnastar_index_builder.xml	Fri Oct 03 21:59:55 2014 -0400
+++ b/data_manager/rnastar_index_builder.xml	Thu Jan 08 17:53:14 2015 -0500
@@ -4,10 +4,11 @@
         <requirement type="package" version="2.4.0d">rnastar</requirement>
     </requirements>
     <command interpreter="python">
-rnastar_index_builder.py --out_file "${out_file}" --fasta_filename "${all_fasta_source.fields.path}" 
+rnastar_index_builder.py "${out_file}" --fasta_filename "${all_fasta_source.fields.path}" 
 --fasta_dbkey "${all_fasta_source.fields.dbkey}" --fasta_description "${all_fasta_source.fields.name}" 
---data_table_name "rnastar_indexes" --out_index_path "${out_file.extra_files_path}" --runThreadN 1
+--runThreadN 1
 #if $genemodel.modelformat=="gff3":
+ --sjdbGTFchrPrefix "${genemodel.sjdbGTFchrPrefix}"
  --sjdbOverhang "${genemodel.sjdbOverhang}"
  --sjdbGTFfile "${genemodel.sjdbGTFfile}"
  --sjdbGTFtagExonParentTranscript "${genemodel.sjdbGTFtagExonParentTranscript}"
@@ -20,7 +21,9 @@
 #if $genemodel.modelformat=="None":
  --sjdbOverhang 0
 #end if
+--data_table_name "rnastar_index" 
 </command>
+
     <stdio>
         <regex match=".*" source="both" level="warning" description="stdout/err chatter:"/>
     </stdio>
@@ -50,7 +53,7 @@
                         <valid initial="string.printable"/>
                      </sanitizer>
                 </param>
-                <param type="text" name="sjdbGTFfeatureExon" value="exon_id" label="GTF feature to use as exon marker"
+                <param type="text" name="sjdbGTFfeatureExon" value="exon" label="GTF feature to use as exon marker"
                    help="GTF feature type in GTF file to be used as exons for building transcripts - use what's in your GTF">
                     <sanitizer invalid_char="">
                         <valid initial="string.printable"/>
@@ -58,7 +61,7 @@
                 </param>
                    
                 <param type="text" name="sjdbGTFtagExonParentTranscript" value="transcript_id" label="GTF feature to define for each exon's parents"
-                   help="GTF tag name to be used as exons' parents for building transcripts - use what's in your GTF">
+                   help="GTF tag name to be used as exons' parents for building transcripts - use what's in your gene model file eg parent for gff3">
                     <sanitizer invalid_char="">
                         <valid initial="string.printable"/>
                      </sanitizer>
--- a/data_manager_conf.xml	Fri Oct 03 21:59:55 2014 -0400
+++ b/data_manager_conf.xml	Thu Jan 08 17:53:14 2015 -0500
@@ -2,7 +2,7 @@
 <data_managers>
     
     <data_manager tool_file="data_manager/rnastar_index_builder.xml" id="rnastar_index_builder" version="0.0.1">
-        <data_table name="rnastar_indexes">
+        <data_table name="rnastar_index">
             <output>
                 <column name="value" />
                 <column name="dbkey" />
--- a/tool_data_table_conf.xml.sample	Fri Oct 03 21:59:55 2014 -0400
+++ b/tool_data_table_conf.xml.sample	Thu Jan 08 17:53:14 2015 -0500
@@ -5,8 +5,8 @@
         <file path="tool-data/all_fasta.loc" />
     </table>
     <!-- Locations of indexes in the BWA mapper format -->
-    <table name="rnastar_indexes" comment_char="#">
+    <table name="rnastar_index" comment_char="#">
         <columns>value, dbkey, name, path</columns>
-        <file path="tool-data/rnastar_indexes.loc" />
+        <file path="tool-data/rnastar_index.loc" />
     </table>
 </tables>