Feature: Use alias-backed Solr targets for ontology search indexing

[mdorf]

Summary

This PR updates ontologies_linked_data to use alias-backed logical Solr targets for term and property indexing.

The goal is to stop binding models directly to fixed physical Solr collection names and instead bind them to stable logical active aliases, while preserving bootstrap collections for initial setup and rebuild workflows.

Prerequisites

This PR depends on the Goo changes introduced in:

• Feature: Add alias-backed Solr support for search indexing goo#177

That PR adds the alias-backed search connection support used here, including logical search targets, alias repointing, bootstrap collections, and per-connection Solr topology settings.

What Changed

Alias-backed term and property search bindings

Updated the term and property indexing models to bind to logical active aliases instead of fixed physical collection names:

• :term_search_active
• :property_search_active

The previous physical collections are now used only as bootstrap collections:

• :term_search
• :property_search

This keeps model bindings stable while allowing the active alias to be repointed to rebuilt physical collections later.

Configurable Solr topology for search collections

Added config-driven shard and replication settings for the term and property search collections:

• term_search_num_shards
• term_search_replication_factor
• property_search_num_shards
• property_search_replication_factor

These values are passed through the existing search bindings so collection creation can be controlled from LinkedData config.

Naming cleanup

Removed the legacy *_core1 terminology from active runtime search bindings for term and property search.

This makes the physical collection names more consistent with the current SolrCloud collection model and with existing names such as ontology_metadata and ontology_data.

Why

Previously, term and property indexing still carried forward legacy physical collection names such as term_search_core1 and prop_search_core1.

With the alias-based search support added in goo, ontologies_linked_data can now bind models to stable logical targets and treat the physical collections as implementation details used for bootstrap or rebuild workflows.

This keeps production-facing search bindings stable while enabling safe rebuild-and-cutover flows in downstream tooling.