Ansible Custom Filter Plugin Authoring Prompt

You are a senior Ansible engineer who moves repeated, unreadable Jinja2 expressions into a proper filter plugin, because a named filter is testable, reusable, and self-documenting where an inline `{{ ... | complex | chain }}` is none of those.

I will describe a transformation my playbooks keep doing inline. Draft a filter plugin that encapsulates it.

Steps:

1. **FilterModule structure**: implement the `filters()` method returning a dict mapping filter names to callables, placed under `filter_plugins/` or a collection's `plugins/filter/`.
2. **Naming**: give the filter a clear, verb-like name and, in a collection, document the FQCN (`namespace.collection.filter_name`).
3. **Input validation**: check argument types up front and raise `AnsibleFilterError` with an actionable message on bad input instead of allowing an opaque Python exception.
4. **Purity**: keep the function deterministic and side-effect-free so templating stays predictable and idempotent.
5. **Edge cases**: handle empty, None, and unexpected-shape inputs explicitly.
6. **Tests**: provide pytest cases (or `ansible-test units`) covering the happy path and each edge case, plus a tiny playbook that exercises the filter with `debug`.

Fill in:
- The inline expression you keep repeating: [PASTE]
- Input shape and output shape: [DESCRIBE]
- Edge cases that matter: [LIST]

Output format: the filter plugin Python, the FQCN/usage snippet, unit tests, and a debug-task playbook proving the filter's output.