Support indexing any content entity type

Problem / Motivation

Currently, AI Content Chat only supports indexing node content types. Many Drupal sites store important content across multiple entity types beyond nodes:

• Taxonomy terms – vocabularies with descriptions (e.g., glossaries, FAQs, product categories)
• Block content (custom blocks) – reusable content blocks with important information (e.g., "About us", contact details, policies)
• Media – media entities with descriptive metadata
• Commerce products – product descriptions for e-commerce sites
• Any custom entity type – contributed or custom modules that define their own content entities

By only supporting nodes, the chatbot misses a significant portion of the site's content, leading to incomplete or inaccurate answers.

Proposed resolution

Make the indexer fully entity-type agnostic:

1. Settings form: Replace the hardcoded node bundle list with a dynamic list of all available content entity types and their bundles. The admin selects which entity types, bundles, and fields to index.
2. ContentIndexer service: Replace all node-specific code (NodeInterface, getStorage('node'), getTitle()) with generic entity handling using ContentEntityInterface, EntityInterface::label(), and dynamic bundle key lookups.
3. Entity hooks: Update hook_entity_insert, hook_entity_update, and hook_entity_delete to react to any configured content entity type, not just nodes.
4. Config schema: Extend content_sources to include an entity_type key alongside bundle and fields.
5. Database: No schema changes needed – the entity_type column already exists in the index table.
6. Backward compatibility: An update hook migrates existing content_type config entries to the new entity_type + bundle format automatically.

Acceptance criteria

• Admin can select any content entity type (node, taxonomy_term, block_content, media, etc.) on the settings form
• Text fields for each bundle are dynamically discovered
• Indexing, re-indexing, and automatic index updates work for all configured entity types
• The chatbot uses content from all indexed entity types when answering questions
• Existing node-only configurations continue to work after running drush updb
• Internal/non-useful entity types (user, file, path_alias, etc.) are filtered from the settings form