chore: bring the test suite up to ruff's standards

Mostly automatic: empty f-strings, unused imports, formatting. The manual half
was duplicated import blocks stranded mid-file and `== None` assertions, which
become `is None` here because they compare plain attributes -- unlike the
identical rule in filters/book.py, where they build SQL.

Two unused variables are the tests never checking the disk in either
delete_files direction. Left in place under a noqa so the gap stays visible.

TODO.md

@@ -114,6 +114,24 @@ CMD ["litestar", "--app-dir", "chitai", "run", "--host", "0.0.0.0", "--port", "8
 `litestar run` is the CLI development runner. Production should invoke uvicorn or granian
 directly, with a worker count.
 
+### The delete_files flag is untested in both directions
+
+`backend/tests/integration/test_book.py` — `test_remove_file_with_delete_files_false_keeps_filesystem_file`
+and `test_remove_file_with_delete_files_true_removes_filesystem_file`
+
+Both tests capture the file's path and then assert only `response.status_code == 204`.
+Neither looks at the disk. So the flag that decides whether removing a file from a book
+also **erases it from the filesystem** is covered in name only, in both directions.
+
+Ruff surfaced this as two `F841` unused variables; the variables carry a `# noqa: F841`
+and a comment rather than being deleted, so the gap stays visible. Remove the noqa when
+the assertions land.
+
+The reason it is not a two-line fix: `FileMetadata.path` is stored relative to `book.path`,
+so the test has to resolve it against the library root to know what to stat. That
+resolution is the same thing `BookService.get_files` is recorded as getting wrong (see
+`backend/AGENTS.md`), so it is worth settling once and using in both places.
+
 ### No type checker on the backend
 
 Formatting, linting and tests now run in CI (`.gitea/workflows/ci.yml`) and block, and the