Mercurial > repos > dfornika > data_manager_build_kraken2_database
changeset 5:1f6fe9dc5490 draft
planemo upload for repository https://github.com/dfornika/galaxytools/tree/master/data_managers/data_manager_build_kraken2_database/ commit 47bcc8467200adee842a5574b67bc6ef46c0be71-dirty
| author | dfornika |
|---|---|
| date | Mon, 04 Mar 2019 18:59:44 -0500 |
| parents | 988c436ad798 |
| children | ce101123d015 |
| files | data_manager/kraken2_build_custom.py data_manager/kraken2_build_custom.xml data_manager/kraken2_build_standard.py data_manager/kraken2_custom_database_builder.xml data_manager/kraken2_standard_database_builder.xml data_manager/make_json.py data_manager_conf.xml tool_data_table_conf.xml.sample |
| diffstat | 8 files changed, 56 insertions(+), 188 deletions(-) [+] |
line wrap: on
line diff
--- a/data_manager/kraken2_build_custom.py Fri Mar 01 14:01:57 2019 -0500 +++ b/data_manager/kraken2_build_custom.py Mon Mar 04 18:59:44 2019 -0500 @@ -23,36 +23,43 @@ sys.exit( return_code ) -def kraken2_build(data_manager_dict, kraken2_args, params, target_directory, data_table_name=DATA_TABLE_NAME): +def kraken2_build(data_manager_dict, kraken2_args, database_name, params, target_directory, data_table_name=DATA_TABLE_NAME): today = datetime.date.today().isoformat() - database_name = "_".join([today, "custom"]) args = [ - 'kraken2-build', '--threads', str(kraken2_args["threads"]), '--download-taxonomy', '--db', database_name ] - # run(args, target_directory) + run(['kraken2-build'] + args, target_directory) args = [ - 'kraken2-build', '--threads', str(kraken2_args["threads"]), '--add-to-library', kraken2_args["fasta"], '--db', database_name ] - run(args, target_directory) - + run(['kraken2-build'] + args, target_directory) + args = [ - 'kraken2-build', + '--threads', str(kraken2_args["threads"]), + '--build', + '--kmer-len', str(kraken2_args["kmer_len"]), + '--minimizer-len', str(kraken2_args["minimizer_len"]), + '--minimizer-spaces', str(kraken2_args["minimizer_spaces"]), + '--db', database_name + ] + + run(['kraken2-build'] + args, target_directory) + + args = [ '--threads', str(kraken2_args["threads"]), '--clean', '--db', database_name ] - run(args, target_directory) + run(['kraken2-build'] + args, target_directory) data_table_entry = { "value": database_name, @@ -73,6 +80,7 @@ def main(): parser = argparse.ArgumentParser() parser.add_argument('params') + parser.add_argument( '-d', '--db', dest='database_name', help='database name' ) parser.add_argument( '-k', '--kmer-len', dest='kmer_len', type=int, default=35, help='kmer length' ) parser.add_argument( '-m', '--minimizer-len', dest='minimizer_len', type=int, default=31, help='minimizer length' ) parser.add_argument( '-s', '--minimizer-spaces', dest='minimizer_spaces', default=6, help='minimizer spaces' )
--- a/data_manager/kraken2_build_custom.xml Fri Mar 01 14:01:57 2019 -0500 +++ b/data_manager/kraken2_build_custom.xml Mon Mar 04 18:59:44 2019 -0500 @@ -10,6 +10,7 @@ python '$__tool_directory__/kraken2_build_custom.py' '${out_file}' --threads \${GALAXY_SLOTS:-1} + --db ${database_name} --kmer-len ${kmer_len} --minimizer-len ${minimizer_len} --minimizer-spaces ${minimizer_spaces} @@ -18,6 +19,7 @@ </command> <inputs> <param name="input_fasta" type="data" format="fasta" multiple="False" optional="true" label="Select history item" /> + <param name="database_name" type="text" label="Name for this database" /> <param name="kmer_len" type="integer" value="35" label="K-mer length in BP" /> <param name="minimizer_len" type="integer" value="31" label="Minimizer length" /> <param name="minimizer_spaces" type="integer" value="6" label="Minimizer spaces" />
--- a/data_manager/kraken2_build_standard.py Fri Mar 01 14:01:57 2019 -0500 +++ b/data_manager/kraken2_build_standard.py Mon Mar 04 18:59:44 2019 -0500 @@ -11,15 +11,26 @@ import subprocess import sys -from pprint import pprint DATA_TABLE_NAME = "kraken2_databases" -def kraken2_build_standard(data_manager_dict, kraken2_args, params, target_directory, data_table_name=DATA_TABLE_NAME): +def run(args, cwd): + proc = subprocess.Popen(args=args, shell=False, cwd=cwd) + return_code = proc.wait() + if return_code: + print("Error building database.", file=sys.stderr) + sys.exit( return_code ) + +def kraken2_build_standard(data_manager_dict, kraken2_args, target_directory, data_table_name=DATA_TABLE_NAME): today = datetime.date.today().isoformat() - database_name = "_".join([today, "standard"]) + database_name = "_".join([ + today, + "standard", + "kmer-len=" + str(kraken2_args["kmer_len"]), + "minimizer-len=" + str(kraken2_args["minimizer_len"]), + "minimizer-spaces=" + str(kraken2_args["minimizer_spaces"]), + ]) args = [ - 'kraken2-build', '--threads', str(kraken2_args["threads"]), '--standard', '--kmer-len', str(kraken2_args["kmer_len"]), @@ -27,41 +38,37 @@ '--minimizer-spaces', str(kraken2_args["minimizer_spaces"]), '--db', database_name ] - proc = subprocess.Popen(args=args, shell=False, cwd=target_directory) - return_code = proc.wait() - if return_code: - print("Error building database.", file=sys.stderr) - sys.exit( return_code ) + + run(['kraken2-build'] + args, target_directory) + args = [ 'kraken2-build', '--threads', str(kraken2_args["threads"]), '--clean', '--db', database_name ] - proc = subprocess.Popen(args=args, shell=False, cwd=target_directory) - return_code = proc.wait() - if return_code: - print("Error building database.", file=sys.stderr) - sys.exit( return_code ) + + run(['kraken2-build'] + args, target_directory) + data_table_entry = { "value": database_name, "name": database_name, "path": database_name } - _add_data_table_entry(data_manager_dict, data_table_name, data_table_entry) + _add_data_table_entry(data_manager_dict, data_table_entry) -def _add_data_table_entry(data_manager_dict, data_table_name, 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 ) + 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('params') + parser.add_argument('data_manager_json') parser.add_argument( '-k', '--kmer-len', dest='kmer_len', type=int, default=35, help='kmer length' ) parser.add_argument( '-m', '--minimizer-len', dest='minimizer_len', type=int, default=31, help='minimizer length' ) parser.add_argument( '-s', '--minimizer-spaces', dest='minimizer_spaces', default=6, help='minimizer spaces' ) @@ -75,9 +82,9 @@ "threads": args.threads, } - params = json.loads(open(args.params).read()) - pprint(params) - target_directory = params['output_data'][0]['extra_files_path'] + 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 ) @@ -87,18 +94,13 @@ else: raise - data_manager_dict = {} - - # build the index kraken2_build_standard( - data_manager_dict, + data_manager_input, kraken2_args, - params, - target_directory + target_directory, ) - # save info to json file - open(args.params, 'wb').write(json.dumps(data_manager_dict)) + open(args.data_manager_json, 'wb').write(json.dumps(data_manager_input)) if __name__ == "__main__":
--- a/data_manager/kraken2_custom_database_builder.xml Fri Mar 01 14:01:57 2019 -0500 +++ /dev/null Thu Jan 01 00:00:00 1970 +0000 @@ -1,50 +0,0 @@ -<?xml version="1.0"?> -<tool id="kraken2_custom_database_builder" name="Kraken2 Custom Database Builder" tool_type="manage_data" version="2.0.7_beta"> - <description>database builder</description> - <requirements> - <requirement type="package" version="2.0.7_beta">kraken2</requirement> - </requirements> - <version_command>kraken2 -version | head -n 1 | awk '{print $NF}'</version_command> - <command detect_errors="exit_code"> - <![CDATA[ - #set $library_list = str($libraries).split(',') - mkdir kraken2-database && cd kraken2-database && - kraken2-build --threads \${GALAXY_SLOTS:-1} --download-taxonomy --db '${database_name}' && - #for $library in $library_list: - #if str($library) != 'None': - kraken2-build --threads \${GALAXY_SLOTS:-1} --download-library $library --db '${database_name}' && - #end if - #end for - #for $fasta in $input_fasta: - #if str($fasta) != 'None': - kraken2-build --threads \${GALAXY_SLOTS:-1} --db '${database_name}' --add-to-library '$fasta' && - #end if - #end for - kraken2-build --threads \${GALAXY_SLOTS:-1} --db '${database_name}' --build --kmer-len ${kmer_len} --minimizer-len ${minimizer_len} && - cd .. && - python '$__tool_directory__/make_json.py' --db '${database_name}' --out '${out_file}' - ]]> - </command> - <inputs> - <param name="database_name" type="text" label="Name for this database" /> - <param name="libraries" type="select" multiple="true" label="Select partial library to download"> - <option value="archaea">Archaea</option> - <option value="bacteria">Bacteria</option> - <option value="plasmid">Plasmid</option> - <option value="viral">Viral</option> - <option value="human">Human</option> - </param> - <param name="input_fasta" type="data" format="fasta" multiple="True" optional="true" label="Select history item" /> - <param name="kmer_len" type="integer" value="35" label="K-mer length in BP" /> - <param name="minimizer_len" type="integer" value="31" label="Minimizer length" /> - <param name="minimizer_spaces" type="integer" value="6" label="Minimizer spaces" /> - </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/kraken2_standard_database_builder.xml Fri Mar 01 14:01:57 2019 -0500 +++ /dev/null Thu Jan 01 00:00:00 1970 +0000 @@ -1,37 +0,0 @@ -<?xml version="1.0"?> -<tool id="kraken2_standard_database_builder" name="Kraken2 Standard Database Builder" tool_type="manage_data" version="2.0.7_beta"> - <description>database builder</description> - <requirements> - <requirement type="package" version="2.0.7_beta">kraken2</requirement> - </requirements> - <version_command>kraken2 -version | head -n 1 | awk '{print $NF}'</version_command> - <command detect_errors="exit_code"> - <![CDATA[ - #import datetime - #set kraken2_db_name = datetime.date.today().strftime('%Y-%m-%d') + "_standard" - mkdir kraken2-database && cd kraken2-database && - kraken2-build - --standard - --threads \${GALAXY_SLOTS:-1} - --db '${kraken2_db_name}' - --kmer-len ${kmer_len} - --minimizer-len ${minimizer_len} - --minimizer-spaces ${minimizer_spaces} && - cd .. && - python '$__tool_directory__/make_json.py' --db '${kraken2_db_name}' --out '${out_file}' - ]]> - </command> - <inputs> - <param name="kmer_len" type="integer" value="35" label="K-mer length in BP" /> - <param name="minimizer_len" type="integer" value="31" label="Minimizer length" /> - <param name="minimizer_spaces" type="integer" value="6" label="Minimizer spaces" /> - </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/make_json.py Fri Mar 01 14:01:57 2019 -0500 +++ /dev/null Thu Jan 01 00:00:00 1970 +0000 @@ -1,27 +0,0 @@ -import argparse -import json -import os -import shutil - - -def main(args): - data_manager_entry = {} - data_manager_entry['value'] = args.database.lower() - data_manager_entry['name'] = args.database - data_manager_entry['path'] = '.' - data_manager_json = dict(data_tables=dict(kraken2_databases=data_manager_entry)) - params = json.loads(open(args.output).read()) - target_directory = params['output_data'][0]['extra_files_path'] - os.mkdir(target_directory) - output_path = os.path.join(os.getcwd(), 'kraken2-database') - for filename in os.listdir(output_path): - shutil.move(os.path.join(output_path, filename), target_directory) - file(args.output, 'w').write(json.dumps(data_manager_json)) - - -if __name__ == '__main__': - parser = argparse.ArgumentParser(description='Create data manager json.') - parser.add_argument('--db', dest='database', action='store', help='Database name') - parser.add_argument('--out', dest='output', action='store', help='JSON filename') - args = parser.parse_args() - main(args)
--- a/data_manager_conf.xml Fri Mar 01 14:01:57 2019 -0500 +++ b/data_manager_conf.xml Mon Mar 04 18:59:44 2019 -0500 @@ -1,34 +1,4 @@ <data_managers> - <data_manager tool_file="data_manager/kraken2_custom_database_builder.xml" id="kraken2_custom_database_builder" version="1.0"> - <data_table name="kraken2_databases"> - <output> - <column name="value"/> - <column name="name"/> - <column name="path" output_ref="out_file"> - <move type="directory" relativize_symlinks="True"> - <target base="${GALAXY_DATA_MANAGER_DATA_PATH}">kraken2_database/${value}</target> - </move> - <value_translation>${GALAXY_DATA_MANAGER_DATA_PATH}/kraken2_database/${value}</value_translation> - <value_translation type="function">abspath</value_translation> - </column> - </output> - </data_table> - </data_manager> - <data_manager tool_file="data_manager/kraken2_standard_database_builder.xml" id="kraken2_standard_database_builder" version="1.0"> - <data_table name="kraken2_databases"> - <output> - <column name="value"/> - <column name="name"/> - <column name="path" output_ref="out_file"> - <move type="directory" relativize_symlinks="True"> - <target base="${GALAXY_DATA_MANAGER_DATA_PATH}">kraken2_database/${value}</target> - </move> - <value_translation>${GALAXY_DATA_MANAGER_DATA_PATH}/kraken2_database/${value}</value_translation> - <value_translation type="function">abspath</value_translation> - </column> - </output> - </data_table> - </data_manager> <data_manager tool_file="data_manager/kraken2_build_standard.xml" id="kraken2_build_standard" version="1.0"> <data_table name="kraken2_databases"> <output> @@ -36,9 +6,9 @@ <column name="name"/> <column name="path" output_ref="out_file"> <move type="directory" relativize_symlinks="True"> - <target base="${GALAXY_DATA_MANAGER_DATA_PATH}">kraken2_database/${value}</target> + <target base="${GALAXY_DATA_MANAGER_DATA_PATH}">kraken2_databases</target> </move> - <value_translation>${GALAXY_DATA_MANAGER_DATA_PATH}/kraken2_database/${value}</value_translation> + <value_translation>${GALAXY_DATA_MANAGER_DATA_PATH}/kraken2_databases</value_translation> <value_translation type="function">abspath</value_translation> </column> </output> @@ -51,9 +21,9 @@ <column name="name"/> <column name="path" output_ref="out_file"> <move type="directory" relativize_symlinks="True"> - <target base="${GALAXY_DATA_MANAGER_DATA_PATH}">kraken2_database/${value}</target> + <target base="${GALAXY_DATA_MANAGER_DATA_PATH}">kraken2_databases</target> </move> - <value_translation>${GALAXY_DATA_MANAGER_DATA_PATH}/kraken2_database/${value}</value_translation> + <value_translation>${GALAXY_DATA_MANAGER_DATA_PATH}/kraken2_databases</value_translation> <value_translation type="function">abspath</value_translation> </column> </output>
--- a/tool_data_table_conf.xml.sample Fri Mar 01 14:01:57 2019 -0500 +++ b/tool_data_table_conf.xml.sample Mon Mar 04 18:59:44 2019 -0500 @@ -1,6 +1,6 @@ <?xml version="1.0"?> <tables> - <!-- Locations of Kraken database in the required format --> + <!-- Locations of Kraken2 database in the required format --> <table name="kraken2_databases" comment_char="#"> <columns>value, name, path</columns> <file path="tool-data/kraken2_databases.loc" />