Add ogc.osc.prov-processing-step: PROV profile for a single geospatial processing step

[nsnarayanam]

D104 (Aganitha), Activities 1 and 2. Supersedes #13 and #14, which were built from earlier branches.

A profile of ogc.ogc-utils.prov constrained to a single geospatial processing step, adding a required processType link to a registered process type. That is the link PROV-O and Processes Part 5 do not themselves provide, and where the profile-plus-register pattern closes.

Nine examples, all validating, profiled against a drought intelligence workflow we run over Telangana.

On the self-assembly point from the 12 Aug call: the examples form one connected provenance graph with ten resolving links. From the projection a consumer can walk back through inference to the fitted model and CDSI, then to VCI and SMDI, and finally to ingest.

Writing the chain out surfaced a real gap: CDSI consumed an SMDI entity no recorded activity produced. That step is now included, which is a small argument for chaining as a validation technique rather than only a demo.

Two examples cover machine learning: training generates a fitted model entity, and inference consumes that same entity.

Flagged as conflicting: the fork has diverged from master over three weeks. The conflicts are on scaffolding files, not content. The substantive addition is _sources/prov-processing-step/. Happy to rebase if you'd prefer it clean.

[nsnarayanam]

Also included: a one-word fix to _sources/ontology/openscience/bblock.json.

The master build is currently failing, locally and in CI, on:

Invalid reference to bblock ogc.contrib.stac.extensions.processes-ontology from ogc.osc.ontology.openscience

The STAC register publishes processing-ontology, not processes-ontology. Correcting the identifier makes the full build pass, all 21 blocks.

Happy to split this into its own PR if you'd rather keep it separate from the D104 block.