* chore: move v3 RFC boundary number to settings

* feat: button for downloading unprepped RFCXML

* chore: ruff

* chore: refactor test to avoid setup overhead

* feat: serve notprepped bytes from blobdb

* fix: typo

* chore; improve test class name

* feat: wrapper explaining notprepped xml

* refactor: DRY

* chore: typesense docker container (commented out)

* feat: batched RFC search index import

* feat: rebuild_searchindex task

* feat: logging / error reporting

* refactor: _task suffix for task name

* test: tests for searchindex utils + tasks

* fix: only create collection if dropped

* fix: typing / lint

* refactor: decode_document_content() utility method

* fix: fall back to latin-1 in retrieve_str()

* refactor: match structure with retrieve_bytes()

* refactor: separate tests_text.py module

* test: test_decode_document_content + ruff

* fix: revert misguided refactor

* test: assert to guarantee test is valid

* test: enforce queryset order assumed by test

* test: match html escaping in test

* test: search more specifically for tokens to avoid mis-reading them when they occur in faker data

These are not useful any more, blobdb is fast