CompletenessScore filtering keeps genes from exluded HOGs

[nromashchenko]

This snippet:

filename = "tests/test-data/case_filtering.orthoxml"
otree = OrthoXMLTree.from_file(filename,
                               score_threshold=0.95, 
                               score_id="CompletenessScore", 
                               high_child_as_rhogs=True)
otree.to_orthoxml("out.orthoxml")

for the test orthoxml will produce valid groups, but in the species section it keeps the genes that are exluded:

<?xml version='1.0' encoding='utf-8'?>
<orthoXML xmlns="http://orthoXML.org/2011/" version="0.5" origin="orthoXML.org" originVersion="1.0">

...
<species name="Chelonia mydas" NCBITaxId="0">
    <database name="someDB" version="42">
      <genes>
        <gene id="cm1" protId="C00001"/>
        <gene id="cm2" protId="C00002"/>
      </genes>
    </database>
  </species>

  <species name="Arabidopsis thaliana" NCBITaxId="0">
    <database name="someDB" version="42">
      <genes>
        <gene id="at1" protId="A00001"/>
        <gene id="at2" protId="A00002"/>
      </genes>
    </database>
  </species>

...
  <groups>
    <orthologGroup id="HOG_Cryptodira">
      <score id="CompletenessScore" value="1.0"/>
      <geneRef id="cm1"/>
      <geneRef id="cm2"/>
    </orthologGroup>
  </groups>

</orthoXML>

e.g. Arabidopsis genes are not listed in any groups anymore, but they are described in the header.