Termination positions for Frameshift/Nonesence SeqVars are not set properly

[gromdimon]

Describe the bug
In some decision blocks we rely on Termination position of the Variant if it's the frameshift/nonesence one. Currently we use pHGVS attribute for the determination of this, but due to many missing values of pHGVSs we have to find better approach to do this.

To Reproduce
Example:
NM_004360.3:c.2506G>T (p.Glu836Ter) (Gene: CDH1)
It gives the pHGVS result by Autopvs1 as NP_004351.1:p.Glu836Ter
compare to one from mehari: 'NM_004360.5:p.?'
this results in different results of termination position.

Expected behavior
The termination positions should be computed in the unified way.

Screenshots
N/A

Additional context
It seems, that mehari lacks some data of pHGVS. In this case, we have to find an alternative way of findng the termination position.

Affected method: _get_pHGVS_termination

[xiamaz]

This looks like a mehari bug. Could you create an issue there? I'll try to prioritize it and perform root-cause analysis.

[gromdimon]

Here: varfish-org/mehari#437

[gromdimon]

The issue is stilll on mehari site, but in auto-acmg it's outdated