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

[mdorf]

Summary

Renames Solr collection references and enables SolrCloud alias-backed re-indexing for term and property search models.

Prerequisites

This PR depends on the Goo changes introduced in:

• Feature: Add alias-backed Solr support for search indexing (ALT) goo#180

Changes

Models now reference alias names (:term_search, :prop_search) instead of hardcoded collection names (:term_search_core1, :prop_search_core1), and declare a bootstrap_collection: so Goo can set up the alias-to-collection mapping on first boot.

• Class and ProvisionalClass: enable_indexing(:term_search_core1) → enable_indexing(:term_search, :main, bootstrap_collection: :term_search_bootstrap)
• OntologyProperty, AnnotationProperty, DatatypeProperty, ObjectProperty: enable_indexing(:prop_search_core1, :property) → enable_indexing(:prop_search, :property, bootstrap_collection: :prop_search_bootstrap)
• Ontology and OntologySubmission: unchanged — these use enable_indexing(:ontology_metadata) with no alias, which runs in plain collection mode

How it works

On first boot, Goo creates the physical collection (e.g. term_search_bootstrap) and an alias (term_search) pointing to it. All reads and writes go through the alias. During a full re-index, a new collection is created and indexed into, then the alias is atomically swapped — zero downtime.

The _core1 suffix is dropped entirely since the old dual-core swap pattern no longer applies in SolrCloud.

Test plan

• Verify bundle exec rake test passes with a running Solr instance
• Confirm Solr shows aliases term_search → term_search_bootstrap and prop_search → prop_search_bootstrap after first boot
• Confirm existing search and indexing functionality is unaffected (alias is transparent to RSolr reads/writes)