# HG changeset patch
# User devteam
# Date 1426774596 14400
# Node ID eec64de5f7fbf1b148d70d566a3165a69e895031
Uploaded
diff -r 000000000000 -r eec64de5f7fb .shed.yml
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/.shed.yml Thu Mar 19 10:16:36 2015 -0400
@@ -0,0 +1,4 @@
+# repository published to https://toolshed.g2.bx.psu.edu/repos/devteam/data_manager_bowtie2_index_builder
+owner: devteam
+name: data_manager_bowtie2_index_builder
+
diff -r 000000000000 -r eec64de5f7fb data_manager/bowtie2_index_builder.py
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/data_manager/bowtie2_index_builder.py Thu Mar 19 10:16:36 2015 -0400
@@ -0,0 +1,76 @@
+#!/usr/bin/env python
+#Dan Blankenberg
+
+import sys
+import os
+import optparse
+import subprocess
+
+from json import loads, dumps
+
+
+DEFAULT_DATA_TABLE_NAME = "bowtie2_indexes"
+
+def get_id_name( params, dbkey, fasta_description=None):
+ #TODO: ensure sequence_id is unique and does not already appear in location file
+ sequence_id = params['param_dict']['sequence_id']
+ if not sequence_id:
+ sequence_id = dbkey
+
+ sequence_name = params['param_dict']['sequence_name']
+ if not sequence_name:
+ sequence_name = fasta_description
+ if not sequence_name:
+ sequence_name = dbkey
+ return sequence_id, sequence_name
+
+def build_bowtie2_index( data_manager_dict, fasta_filename, params, target_directory, dbkey, sequence_id, sequence_name, data_table_name=DEFAULT_DATA_TABLE_NAME ):
+ #TODO: allow multiple FASTA input files
+ 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 )
+ args = [ 'bowtie2-build', sym_linked_fasta_filename, sequence_id ]
+ proc = subprocess.Popen( args=args, shell=False, cwd=target_directory )
+ return_code = proc.wait()
+ if return_code:
+ print >> sys.stderr, "Error building index."
+ sys.exit( return_code )
+ 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['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' )
+ (options, args) = parser.parse_args()
+
+ filename = args[0]
+
+ params = loads( open( filename ).read() )
+ target_directory = params[ 'output_data' ][0]['extra_files_path']
+ os.mkdir( target_directory )
+ 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 ) )
+
+ sequence_id, sequence_name = get_id_name( params, dbkey=dbkey, fasta_description=options.fasta_description )
+
+ #build the index
+ build_bowtie2_index( data_manager_dict, options.fasta_filename, params, target_directory, dbkey, sequence_id, sequence_name, data_table_name=options.data_table_name or DEFAULT_DATA_TABLE_NAME )
+
+ #save info to json file
+ open( filename, 'wb' ).write( dumps( data_manager_dict ) )
+
+if __name__ == "__main__": main()
diff -r 000000000000 -r eec64de5f7fb data_manager/bowtie2_index_builder.xml
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/data_manager/bowtie2_index_builder.xml Thu Mar 19 10:16:36 2015 -0400
@@ -0,0 +1,29 @@
+
+ builder
+
+ bowtie2
+
+ bowtie2_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 "bowtie2_indexes"
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+.. class:: infomark
+
+**Notice:** If you leave name, description, or id blank, it will be generated automatically.
+
+
+
diff -r 000000000000 -r eec64de5f7fb data_manager_conf.xml
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/data_manager_conf.xml Thu Mar 19 10:16:36 2015 -0400
@@ -0,0 +1,22 @@
+
+
+
+
+
+
+
+
+
+
diff -r 000000000000 -r eec64de5f7fb tool-data/bowtie2_indices.loc.sample
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/tool-data/bowtie2_indices.loc.sample Thu Mar 19 10:16:36 2015 -0400
@@ -0,0 +1,37 @@
+# bowtie2_indices.loc.sample
+# This is a *.loc.sample file distributed with Galaxy that enables tools
+# to use a directory of indexed data files. This one is for Bowtie2 and Tophat2.
+# See the wiki: http://wiki.galaxyproject.org/Admin/NGS%20Local%20Setup
+# First create these data files and save them in your own data directory structure.
+# Then, create a bowtie_indices.loc file to use those indexes with tools.
+# Copy this file, save it with the same name (minus the .sample),
+# follow the format examples, and store the result in this directory.
+# The file should include an one line entry for each index set.
+# The path points to the "basename" for the set, not a specific file.
+# It has four text columns seperated by TABS.
+#
+#
+#
+# So, for example, if you had hg18 indexes stored in:
+#
+# /depot/data2/galaxy/hg19/bowtie2/
+#
+# containing hg19 genome and hg19.*.bt2 files, such as:
+# -rw-rw-r-- 1 james james 914M Feb 10 18:56 hg19canon.fa
+# -rw-rw-r-- 1 james james 914M Feb 10 18:56 hg19canon.1.bt2
+# -rw-rw-r-- 1 james james 683M Feb 10 18:56 hg19canon.2.bt2
+# -rw-rw-r-- 1 james james 3.3K Feb 10 16:54 hg19canon.3.bt2
+# -rw-rw-r-- 1 james james 683M Feb 10 16:54 hg19canon.4.bt2
+# -rw-rw-r-- 1 james james 914M Feb 10 20:45 hg19canon.rev.1.bt2
+# -rw-rw-r-- 1 james james 683M Feb 10 20:45 hg19canon.rev.2.bt2
+#
+# then the bowtie2_indices.loc entry could look like this:
+#
+#hg19 hg19 Human (hg19) /depot/data2/galaxy/hg19/bowtie2/hg19canon
+#
+#More examples:
+#
+#mm10 mm10 Mouse (mm10) /depot/data2/galaxy/mm10/bowtie2/mm10
+#dm3 dm3 D. melanogaster (dm3) /depot/data2/galaxy/mm10/bowtie2/dm3
+#
+#
diff -r 000000000000 -r eec64de5f7fb tool_data_table_conf.xml.sample
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/tool_data_table_conf.xml.sample Thu Mar 19 10:16:36 2015 -0400
@@ -0,0 +1,8 @@
+
+
+
+
+ value, dbkey, name, path
+
+
+
diff -r 000000000000 -r eec64de5f7fb tool_dependencies.xml
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/tool_dependencies.xml Thu Mar 19 10:16:36 2015 -0400
@@ -0,0 +1,6 @@
+
+
+
+
+
+