IMP: add skbio.DistanceMatrix validation for LSMatFormat

[VinzentRisch]

Description

• Expands the validation of the LSMatFormat to the skbio.DistanceMatrix restrictions.
• Matrix has to be square, numerical, with unique ids, symmetrical and 0 on the diagonal.
• The LSMatFormat was always designed to be a skbio.DistanceMatrix as seen in the transformers but the validation did not enforce the symmetry and the 0 on the diagonal until now.
• I added this better validation because I want to use the format for the PLSDA implementation.
• Added transformers for LSMatFormat <-> pd.DataFrame

AI Disclosure

• NO AI USED.
• AI USED.

AI Usage Details

Codex

[lizgehret]

hey @VinzentRisch overall these changes look good! just few comments inline.

[lizgehret]

Hey @VinzentRisch these changes look great, thanks! I pushed some small lint/format updates (since you're ooo) but once CI passes this will be good to merge 🙂