docs: publish v2.2 as latest (#2111)

docs(datasets): a time_format mismatch fails the dataset, and iso8601 on a timestamp or date column is ignored with a warning (#2325) * docs(datasets): a time_format mismatch fails the dataset, and iso8601 on a timestamp or date column is ignored with a warning (spiceai/spiceai#14777) * docs(datasets): a string time_column needs time_format iso8601 (spiceai/spiceai#14777)

docs: publish v2.2 as latest (#2111)

docs: publish v2.2 as latest (#2111)

docs: publish v2.2 as latest (#2111)

docs(snapshots): snapshots.location must be a URI; a bare filesystem path is rejected

docs: publish v2.2 as latest (#2111)

docs(models): paged attention is requested automatically, not unconditionally enabled The local-model paged-attention note collapsed three branches into one absolute sentence. `MistralLlama::paged_attention_config` returns `None` when the device is CPU or when the build is not `all(feature = "cuda", target_family = "unix")`, and the engine downgrades to dense attention again for architectures with no paged kernel (the Multi-head Latent Attention GGUFs) — so a CUDA+Unix build is necessary, not sufficient. From v2.2.0 (spiceai/spiceai#11990) the note is also wrong to say no configuration is needed: `FileModelParams::paged_attention` accepts `auto` (default) or `disabled`, and `disabled` forces dense attention. It is a `from: file:` parameter only — `HuggingFaceModelParams` has no such field and `from_hf` calls `paged_attention_config` unconditionally. Scoped accordingly: the branch distinction lands in v2.0.x-v2.3.x and vNext; the escape hatch only in v2.2.x, v2.3.x and vNext.

docs(runtime): runtime.auth.api-key.keys is required

docs: publish v2.2 as latest (#2111)

docs(cayenne): refresh_append_overlap is supported from v2.2.1 spiceai/spiceai#13574 removed the load-time rejection and made Cayenne honor the overlap window; git tag --contains puts it in v2.2.1 only. #2186 dropped the limitation bullet from the vNext Cayenne page but left the same claim standing on reference/spicepod/datasets.md and views.md, which restate per-engine capability claims and are the usual miss for component-page-only fixes. - vNext datasets.md/views.md: claim removed - version-2.2.x: split by patch (v2.2.0 rejects, v2.2.1 supports) on the Cayenne limitations bullet and both reference pages - version-2.1.x and older left alone: the rejection is present at v2.1.5, v2.0.1, v1.11.6, v1.10.x and v1.9.x

docs: publish v2.2 as latest (#2111)