| Previous changeset 3:3f05bf162005 (2021-04-19) Next changeset 5:27258eb73f0c (2021-06-14) |
|
Commit message:
"planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/metaphlan/ commit 2b87bc7417360e2b2c9ec0605d475909f6f0482f" |
|
modified:
macros.xml |
|
added:
formatoutput.py |
| b |
| diff -r 3f05bf162005 -r 2e5ae05160e0 formatoutput.py --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/formatoutput.py Mon May 17 20:09:54 2021 +0000 |
| [ |
| @@ -0,0 +1,144 @@ +#!/usr/bin/env python +# -*- coding: utf-8 -*- + +import argparse +import re +from pathlib import Path + +taxo_level = { + 'k': 'kingdom', + 'p': 'phylum', + 'c': 'class', + 'o': 'order', + 'f': 'family', + 'g': 'genus', + 's': 'species', + 't': 'strains'} + + +def split_levels(metaphlan_output_fp, out_dp, legacy_output): + ''' + Split default MetaPhlAn into a report for each taxonomic level + + :param metaphlan_output_fp: Path default MetaPhlAn output + :param out_dp: Path to output directory + :param legacy_output: Boolean for legacy output + ''' + # prepare output files + abund_f = { + 'k': open(out_dp / Path('kingdom'), 'w'), + 'p': open(out_dp / Path('phylum'), 'w'), + 'c': open(out_dp / Path('class'), 'w'), + 'o': open(out_dp / Path('order'), 'w'), + 'f': open(out_dp / Path('family'), 'w'), + 'g': open(out_dp / Path('genus'), 'w'), + 's': open(out_dp / Path('species'), 'w'), + 't': open(out_dp / Path('strains'), 'w') + } + for level in abund_f: + abund_f[level].write("%s\t" % taxo_level[level]) + if not legacy_output: + abund_f[level].write("%s_id\t" % taxo_level[level]) + abund_f[level].write("abundance\n") + + levels_number = len(taxo_level) + + with open(metaphlan_output_fp, 'r') as metaphlan_output_f: + with open(out_dp / Path('all'), 'w') as all_level_f: + # write header in all leve file + for level in ['k', 'p', 'c', 'o', 'f', 'g', 's', 't']: + all_level_f.write("%s\t" % taxo_level[level]) + if not legacy_output: + all_level_f.write("%s_id\t" % taxo_level[level]) + all_level_f.write("abundance\n") + + # parse metaphlan file + for line in metaphlan_output_f.readlines(): + # skip headers + if line.startswith("#"): + continue + + # spit lines + split_line = line[:-1].split('\t') + taxo_n = split_line[0].split('|') + if legacy_output: + abundance = split_line[1] + else: + taxo_id = split_line[1].split('|') + abundance = split_line[2] + + # get taxon name and ids + for i in range(len(taxo_n)): + taxo = taxo_n[i].split('__')[1] + taxo = taxo.replace("_", " ") + all_level_f.write("%s\t" % taxo) + if not legacy_output: + all_level_f.write("%s\t" % taxo_id[i]) + + # if not all taxon levels + for i in range(len(taxo_n), levels_number): + all_level_f.write('\t') + + all_level_f.write("%s\n" % abundance) + + # write + last_taxo_level = taxo_n[-1].split('__') + taxo = last_taxo_level[1].replace("_", " ") + level = last_taxo_level[0] + abund_f[level].write("%s\t" % taxo) + if not legacy_output: + abund_f[level].write("%s\t" % taxo_id[-1]) + abund_f[level].write("%s\n" % abundance) + + # close files + for taxo_level_f in abund_f: + abund_f[taxo_level_f].close() + + +def format_for_krona(metaphlan_output_fp, krona_out_fp): + ''' + Split default MetaPhlAn into a report for each taxonomic levKRONAel + + :param metaphlan_output_fp: Path default MetaPhlAn output + :param krona_out: Path to output file for Krona + ''' + re_replace = re.compile(r"\w__") + re_bar = re.compile(r"\|") + re_underscore = re.compile(r"_") + + with open(metaphlan_output_fp, 'r') as metaphlan_output_f: + with open(krona_out_fp, 'w') as krona_out_f: + for line in metaphlan_output_f.readlines(): + if "s__" in line: + x = line.rstrip().split('\t') + lineage = re.sub(re_bar, '', x[0]) + lineage = re.sub(re_replace, '\t', lineage) + lineage = re.sub(re_underscore, ' ', lineage) + krona_out_f.write("%s\t%s\n" % (x[-1], lineage)) + + +if __name__ == '__main__': + parser = argparse.ArgumentParser(description='Format MetaPhlAn output') + subparsers = parser.add_subparsers(dest='function') + # split_levels + split_levels_parser = subparsers.add_parser('split_levels', help='Split default MetaPhlAn into a report for each taxonomic level') + split_levels_parser.add_argument('--metaphlan_output', help="Path to default MetaPhlAn output") + split_levels_parser.add_argument('--outdir', help="Path to output directory") + split_levels_parser.add_argument('--legacy-output', dest='legacy_output', action='store_true', help="Old MetaPhlAn2 two columns output") + split_levels_parser.set_defaults(legacy_output=False) + # format_for_krona + format_for_krona_parser = subparsers.add_parser('format_for_krona', help='Split default MetaPhlAn into a report for each taxonomic level') + format_for_krona_parser.add_argument('--metaphlan_output', help="Path to default MetaPhlAn output") + format_for_krona_parser.add_argument('--krona_output', help="Path to Krona output directory") + + args = parser.parse_args() + + if args.function == 'split_levels': + split_levels( + Path(args.metaphlan_output), + Path(args.outdir), + args.legacy_output) + elif args.function == 'format_for_krona': + format_for_krona( + Path(args.metaphlan_output), + Path(args.krona_output)) |
| b |
| diff -r 3f05bf162005 -r 2e5ae05160e0 macros.xml --- a/macros.xml Mon Apr 19 20:54:50 2021 +0000 +++ b/macros.xml Mon May 17 20:09:54 2021 +0000 |
| b |
| @@ -1,6 +1,6 @@ <?xml version="1.0"?> <macros> - <token name="@TOOL_VERSION@">3.0.7</token> + <token name="@TOOL_VERSION@">3.0.8</token> <token name="@VERSION_SUFFIX@">0</token> <token name="@PROFILE@">20.01</token> <xml name="edam_ontology"> @@ -24,17 +24,4 @@ <citation type="doi">1101/2020.11.19.388223</citation> </citations> </xml> - <token name="@FILE_FORMATS@">fastq,fastq.gz,fastq.bz2,fasta,fasta.gz,fasta.bz2</token> - <xml name="tax_lev"> - <param argument="--tax_lev" type="select" label="Taxonomic level for the relative abundance output"> - <option value="a" selected="true">All taxonomic levels</option> - <option value="k">Kingdoms only</option> - <option value="p">Phyla only</option> - <option value="c">Classes only</option> - <option value="o">Orders only</option> - <option value="f">Families only</option> - <option value="g">Genera only</option> - <option value="s">Species only</option> - </param> - </xml> </macros> |