Migrate dense_slide_extraction off hs2p's removed TilingConfig.tissue_threshold (#97)

[clemsgrs]

Context

hs2p 4.2.0 (clemsgrs/hs2p#158) collapsed the derived TilingConfig.tissue_threshold scalar into the resolved min_coverage map — min_coverage["tissue"] is now the single source of truth for the tissue threshold. soma's slide-manifest dense path constructed hs2p.TilingConfig(tissue_threshold=...) directly, which now raises TypeError.

soma was in fact already broken on a fresh install: the existing slide2vec[fm]>=4.8.0 floor resolves to slide2vec 5.0.0, which requires hs2p>=4.2.0 — dragging in the field removal regardless of soma's own hs2p pin.

Changes

• dense_slide_extraction._build_tiling_config: feed min_coverage={"tissue": ...} instead of the removed scalar. The per-class sampling map still flows separately through _resolve_sampling_spec_from_masks(masks, …); this tiling-level entry only feeds the result's binary min_tissue_fraction provenance, so ROI coordinates — and the dense cache key, which never referenced TilingConfig — are unchanged.
• pyproject.toml: bump hs2p >=4.1.1 → >=4.2.0 and slide2vec[fm] >=4.8.0 → >=5.0.0. The bumps are coupled: slide2vec must be >=5.0.0 to run on hs2p 4.2.0 (4.8.0 still reads the removed field). soma does not touch the one breaking public API in slide2vec 5.0.0 (Model.embed_slides nested dict, slide2vec#178) — it drives the lower-level runtime, whose 4.8.0→5.0.0 delta is small and additive.

Testing

Full suite green against hs2p 4.2.0 + slide2vec 5.0.0: 1106 passed, 3 skipped, 1 xfailed. test_sphinx_docs_build deselected for a missing docutils env dep (unrelated).

Acceptance criteria

• No bare tissue_threshold= scalar into hs2p.TilingConfig
• hs2p>=4.2.0 pin (+ coupled slide2vec>=5.0.0)
• No on-disk cache invalidation (dense key independent of TilingConfig)
• Full soma suite green

Closes #97