Mercurial > repos > dfornika > data_manager_build_mash_sketch_refseq
changeset 3:350da4807491 draft
planemo upload for repository https://github.com/dfornika/galaxytools/tree/master/data_managers/data_manager_build_mash_sketch/ commit 34a1400396f10e63e79511a4b28501fdaa03e77d-dirty
author | dfornika |
---|---|
date | Fri, 14 Jun 2019 19:08:05 -0400 |
parents | a6b8ec872e3c |
children | 77d6638116d0 |
files | data_manager/mash_build_sketch.py data_manager/mash_build_sketch.xml data_manager/mash_build_sketch_refseq.py data_manager/mash_build_sketch_refseq.xml |
diffstat | 4 files changed, 139 insertions(+), 130 deletions(-) [+] |
line wrap: on
line diff
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/data_manager/mash_build_sketch.py Fri Jun 14 19:08:05 2019 -0400 @@ -0,0 +1,109 @@ +#!/usr/bin/env python + +from __future__ import print_function + +import argparse +import datetime +import errno +import json +import os +import string +import subprocess +import sys + + +DATA_TABLE_NAME = "mash_sketches" + +def run(args, cwd): + proc = subprocess.Popen(args=args, shell=False, cwd=cwd) + return_code = proc.wait() + if return_code: + print("Error building sketch.", file=sys.stderr) + sys.exit( return_code ) + +def mash_build_sketch(data_manager_dict, mash_args, target_directory, data_table_name=DATA_TABLE_NAME): + + now = datetime.datetime.utcnow().strftime("%Y-%m-%dT%H%M%SZ") + + database_value = "_".join([ + now, + "refseq" + "." + + sketch_type + "." + + "k21.s1000.msh", + ]) + + database_name = " ".join([ + "refseq" + "." + + sketch_type + "." + + "k21.s1000.msh" + "(Created:", + now + ")" + ]) + + database_path = database_value + + args = [ + '-p', mash_args['threads'], + '-k', mash_args['kmer_size'], + '-s', mash_args['sketch_size'], + '-o', 'sketch' + ] + + subprocess.check_call(['mash', 'sketch'] + args, target_directory) + + data_table_entry = { + "value": database_value, + "name": database_name, + "path": database_path, + } + + _add_data_table_entry(data_manager_dict, data_table_entry) + + +def _add_data_table_entry(data_manager_dict, data_table_entry, data_table_name=DATA_TABLE_NAME): + 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(): + parser = argparse.ArgumentParser() + parser.add_argument('data_manager_json') + parser.add_argument('--kmer-size', dest='kmer_size', help='K-mer size' ) + parser.add_argument('--sketch-size', dest='sketch_size', help='Sketch size' ) + parser.add_argument( '--threads', dest='threads', default=1, help='threads' ) + + args = parser.parse_args() + + data_manager_input = json.loads(open(args.data_manager_json).read()) + + target_directory = data_manager_input['output_data'][0]['extra_files_path'] + + mash_args = { + 'kmer_size': args.kmer_size, + 'sketch_size': args.kmer_size, + 'threads': args.threads, + } + + try: + os.mkdir( target_directory ) + except OSError as exc: + if exc.errno == errno.EEXIST and os.path.isdir( target_directory ): + pass + else: + raise + + data_manager_output = {} + + mash_build_sketch( + data_manager_output, + mash_args, + target_directory, + ) + + open(args.data_manager_json, 'wb').write(json.dumps(data_manager_output)) + + +if __name__ == "__main__": + main()
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/data_manager/mash_build_sketch.xml Fri Jun 14 19:08:05 2019 -0400 @@ -0,0 +1,30 @@ +<?xml version="1.0"?> +<tool id="mash_build_sketch" name="Mash Sketch Builder" tool_type="manage_data" version="2.1"> + <description>mash sketch builder</description> + <requirements> + <requirement type="package" version="2.1">mash</requirement> + </requirements> + <version_command>mash --version</version_command> + <command detect_errors="exit_code"> + <![CDATA[ + python '$__tool_directory__/mash_build_sketch.py' + '${out_file}' + --threads \${GALAXY_SLOTS:-1} + --kmer-size ${kmer_size} + --sketch-size ${sketch_size} + ]]> + </command> + <inputs> + <param name="input" type="data" multiple="true" label="Sequences to sketch" /> + <param name="kmer_size" type="integer" default="21" label="K-mer size" /> + <param name="sketch_size" type="integer" default="1000" label="Sketch size" /> + </inputs> + <outputs> + <data name="out_file" format="data_manager_json" /> + </outputs> + <help> + </help> + <citations> + <citation type="doi">10.1186/gb-2014-15-3-r46</citation> + </citations> +</tool>
--- a/data_manager/mash_build_sketch_refseq.py Wed Mar 06 19:52:04 2019 -0500 +++ /dev/null Thu Jan 01 00:00:00 1970 +0000 @@ -1,99 +0,0 @@ -#!/usr/bin/env python - -from __future__ import print_function - -import argparse -import datetime -import errno -import json -import os -import string -import subprocess -import sys - - -DATA_TABLE_NAME = "mash_sketches" - -def run(args, cwd): - proc = subprocess.Popen(args=args, shell=False, cwd=cwd) - return_code = proc.wait() - if return_code: - print("Error building sketch.", file=sys.stderr) - sys.exit( return_code ) - -def mash_build_sketch_refseq(data_manager_dict, sketch_type, target_directory, data_table_name=DATA_TABLE_NAME): - - now = datetime.datetime.utcnow().strftime("%Y-%m-%dT%H%M%SZ") - - database_value = "_".join([ - now, - "refseq" + "." + - sketch_type + "." + - "k21.s1000.msh", - ]) - - database_name = " ".join([ - "refseq" + "." + - sketch_type + "." + - "k21.s1000.msh" - "(Created:", - now + ")" - ]) - - database_path = database_value - - args = [ - 'https://gembox.cbcb.umd.edu/mash/refseq.' + sketch_type + '.k21s1000.msh' - ] - - run(['wget'] + args, target_directory) - - data_table_entry = { - "value": database_value, - "name": database_name, - "path": database_path, - } - - _add_data_table_entry(data_manager_dict, data_table_entry) - - -def _add_data_table_entry(data_manager_dict, data_table_entry, data_table_name=DATA_TABLE_NAME): - 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(): - parser = argparse.ArgumentParser() - parser.add_argument('data_manager_json') - parser.add_argument('--sketch-type', dest='sketch_type', help='Sketch Type (genome or plasmid or genome+plasmid)' ) - parser.add_argument( '-t', '--threads', dest='threads', default=1, help='threads' ) - - args = parser.parse_args() - - data_manager_input = json.loads(open(args.data_manager_json).read()) - - target_directory = data_manager_input['output_data'][0]['extra_files_path'] - - try: - os.mkdir( target_directory ) - except OSError as exc: - if exc.errno == errno.EEXIST and os.path.isdir( target_directory ): - pass - else: - raise - - data_manager_output = {} - - mash_build_sketch_refseq( - data_manager_output, - args.sketch_type, - target_directory, - ) - - open(args.data_manager_json, 'wb').write(json.dumps(data_manager_output)) - - -if __name__ == "__main__": - main()
--- a/data_manager/mash_build_sketch_refseq.xml Wed Mar 06 19:52:04 2019 -0500 +++ /dev/null Thu Jan 01 00:00:00 1970 +0000 @@ -1,31 +0,0 @@ -<?xml version="1.0"?> -<tool id="mash_build_sketch_refseq" name="Mash RefSeq Sketch Builder" tool_type="manage_data" version="2.0"> - <description>database builder</description> - <requirements> - <requirement type="package" version="2.1">mash</requirement> - </requirements> - <version_command>mash --version</version_command> - <command detect_errors="exit_code"> - <![CDATA[ - python '$__tool_directory__/mash_build_sketch_refseq.py' - '${out_file}' - --threads \${GALAXY_SLOTS:-1} - --sketch-type ${sketch_type} - ]]> - </command> - <inputs> - <param name="sketch_type" type="select" multiple="false" label="Select MiniKraken2 database version to download"> - <option value="genome">genome</option> - <option value="plasmid">plasmid</option> - <option value="genome+plasmid">genome+plasmid</option> - </param> - </inputs> - <outputs> - <data name="out_file" format="data_manager_json" /> - </outputs> - <help> - </help> - <citations> - <citation type="doi">10.1186/gb-2014-15-3-r46</citation> - </citations> -</tool>