min_sample_depth uses exclusive threshhold, unlike other qiime2 plugins

[mestaki]

Hey @cameronmartino,
Trying the new and improved gemelli and loving all the upgrades! Kudos my friend.

One small convenience thing. From what I can tell, rpca's min-sample-count leads to an exclusive thresshold whereas other qiime2 plugins, like rarefy use inclusive thressholding.

My case example:

!qiime srs SRS \
  --i-table table-dada2-150-gg2.qza \
  --p-c-min 11661 \
  --o-normalized-table srs-11661-table-dada2-150-gg2.qza

The sample with 11661 is retained.

!qiime gemelli rpca \
    --i-table srs-11661-table-dada2-150-gg2.qza \
    --o-biplot gemelli/rpca-srs-11661-ordination.qza \
    --o-distance-matrix gemelli/rpca-srs-11661-distance.qza

Same number of samples as original table.

!qiime gemelli rpca \
    --i-table table-dada2-150-gg2.qza \
    --p-min-sample-count 11661 \
    --o-biplot gemelli/rpca-nonrarefy-ordination.qza \
    --o-distance-matrix gemelli/rpca-nonrarefy-distance.qza

Here, the sample with 11661 is dropped in the distance matrix.
Would be great if the behavior is consistent so one could just set a $min-sample-depth object and use that in both instead of doing an n-1.