validateTaxonomy.yml

on: 
  push:
    branches: [master, dev, esearch-input]
name: Validate taxonomy

jobs:
  build:
    runs-on: ${{ matrix.os }}
    strategy:
      matrix:
        os: ['ubuntu-20.04' ]
        perl: [ '5.32' ]
    name: Perl ${{ matrix.perl }} on ${{ matrix.os }}
    steps:
      - name: Set up perl
        uses: shogo82148/actions-setup-perl@v1
        with:
          perl-version: ${{ matrix.perl }}
          multi-thread: "true"
      - name: checkout my repo
        uses: actions/checkout@v4
        with:
          path: Kalamari
      - name: update PATH
        run:  |
          echo $GITHUB_WORKSPACE/Kalamari/bin >> $GITHUB_PATH
          echo $PATH
          echo ""
          cat $GITHUB_PATH
      - name: build taxonomy
        run:  |
          echo $PATH
          bash Kalamari/bin/buildTaxonomy.sh
          bash Kalamari/bin/filterTaxonomy.sh
          ls -lhR Kalamari/share/kalamari-*/taxonomy
      - name: validate taxonomy
        run:  |
          perl Kalamari/bin/validateTaxonomy.pl Kalamari/share/kalamari-*/taxonomy
      - name: validate filtered taxonomy
        run:  |
          perl Kalamari/bin/validateTaxonomy.pl Kalamari/share/kalamari-*/taxonomy/filtered
      - name: matching taxids
        run:  |
          export taxdir=$(\ls -d Kalamari/share/kalamari-*/taxonomy)
          echo "Making sure that all taxids in chromosomes.tsv and plasmids.tsv are present in nodes.tsv and names.tsv"
          tail -n +2 Kalamari/src/chromosomes.tsv Kalamari/src/plasmids.tsv -q | perl -F'\t' -lane 'BEGIN{@node=`cat $ENV{taxdir}/nodes.dmp`; for $n(@node){($taxid)=split(/\t/, $n); $taxid{$taxid}++; } } for my $t($F[2], $F[3]){ if(!$taxid{$t}){ print "Could not find $t taxid";} }'
          tail -n +2 Kalamari/src/chromosomes.tsv Kalamari/src/plasmids.tsv -q | perl -F'\t' -lane 'BEGIN{@name=`cat $ENV{taxdir}/names.dmp`; for $n(@name){($taxid)=split(/\t/, $n); $taxid{$taxid}++; } } for my $t($F[2], $F[3]){ if(!$taxid{$t}){ print "Could not find $t taxid";} }'