view format_taxon_list.py @ 1:5ee29434330f draft

planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/checkm commit a96defec82bfa9628fa821ffa6df5c68314a41d8
author iuc
date Mon, 08 Aug 2022 20:00:57 +0000
parents 7ccea589e8b7
children
line wrap: on
line source

from pathlib import Path

from checkm.taxonParser import TaxonParser
from checkm.util.taxonomyUtils import taxonomicRanks

if __name__ == '__main__':
    tool_data_dp = Path('tool-data')

    # get all available marker sets
    taxonParser = TaxonParser()
    taxonMarkerSets = taxonParser.readMarkerSets()

    # create a table per rank
    for rank in taxonomicRanks:
        rank_fp = tool_data_dp / Path("%s.loc.sample" % rank)
        with rank_fp.open('w') as rank_f:
            rank_f.write('# File generated by format_taxon_list.py script\n')
            rank_f.write('# taxon\tdescription\n')
            for taxon in sorted(taxonMarkerSets[rank]):
                markerSet = taxonMarkerSets[rank][taxon]
                numMarkers, numMarkerSets = markerSet.size()
                rank_f.write('{taxon}\t{taxon} ({markerSet.numGenomes} genomes, {numMarkers} marker genes, {numMarkerSets} marker sets)\n')