Properly handle selenoproteins

[holtgrewe]

Is your feature request related to a problem? Please describe.
Certain transcripts have more than one stop codon. This leads to issues when trying to predict the effect using hgvs-rs (the issue already is in the Python-based hgvs). In this case, a prediction could still be made by ignoring the first stop codon. Of course, a note needs to be added.

Describe the solution you'd like
Implement appropriate behaviour to ignore the first stop codon where appropriate.

Describe alternatives you've considered
N/A

Additional context
Mehari log messages look like:

2023-10-21T07:20:59.106574Z DEBUG No consequences for VcfVariant { chromosome: "3", position: 49395540, reference: "G", alternative: "" } on NM_000581.4 (hgvs_p=p.?) - adding gene_variant;most likely the transcript has multiple stop codons and the variant lies behind the first.

[holtgrewe]

Most likely related to selenoproteins:

• Allow predictions even for transcripts that have more than one stop codon hgvs-rs#130
• Address stop codon recoding / stop codon readthrough (SCR) hgvs-rs#143