When the VariantModel has multiple lexicons, in the find_all_matches results only the first lexicon is shown in the results, even when the match is from the other lexicon.
I added a test with 2 non-overlapping lexicons.
Thanks! Looks good, I rebased things a bit (made the 2nd commit a fixup to the first one so the unnecessary file don't pollute the history) and merged the resulting commit.
When the VariantModel has multiple lexicons, in the find_all_matches results only the first lexicon is shown in the results, even when the match is from the other lexicon. I added a test with 2 non-overlapping lexicons.