need to make sure forwarded IDs are not dropped from input trees

[mtholder]

No description provided.

[mtholder]

See handling of 4945810 which is now forwarded to 296662.
The leaf mapped to that in study pg_2460 is pruned (present in phylo_snapshot, absent in cleaned_phylo) when it should be retained.

[mtholder]

note: in some contexts, (e.g. apparently the conflict call) the ID is getting correctly forwarded, which can lead to the conflict API returning support statements that don't jive with the annotations on the synth tree (because the leaf set of the same tree is pruned differently in the differing cleaning steps).

[bredelings]

See screenshots and picture from @snacktavish on ot-private on Aug 10.

This edge in pg_2460 is orange -- conflicts with synth tree:

        Source: node912239
        Target: node912240
        Conflicts with mrcaott7686ott24142
        Edge length: 0.06818813934

However, when we view that edge in the synth tree viewer, it doesn't list pg_2460 as a conflicting study.