Add ox_prune --queue, to match ox_health --queue

[burakeyler]

Closes #37

What

ox_prune gains --queue, spelled and defaulted like ox_health --queue (default=None, "all queues"). When given, it adds queue_name=<queue> to the task queryset before the batched delete loop, so it only narrows which rows are eligible — --batch-size behaves exactly as before. The output label names the queue (Deleted 3 SUCCESSFUL/DISCARDED (queue emails) task row(s) ...); without --queue output is unchanged.

Schedule ticks: OxScheduleTick has no queue, so tick pruning is left as it is with or without --queue (anchors still kept). I documented that in the flag's row. If you'd rather --queue skip tick pruning entirely, that's a one-line change — happy to switch.

Out of scope, per the issue: several queues in one invocation.

Tests

Three new tests in tests/test_prune.py, each failing without the change:

• only the named queue's old rows are deleted;
• --dry-run counts only that queue;
• --queue with --batch-size 2 still deletes every eligible row across batches.

Gates (run locally, Windows, SQLite)

• pytest -q tests/test_prune.py: 29 passed. Full suite: 1010 passed, 35 skipped, 3 failed — the 3 are in tests/test_platform_signals.py and fail identically on unmodified main on Windows.
• ruff check ., ruff format --check ., mypy --strict src/, tools/check_release.py: pass.
• Not run: mkdocs build --strict, PostgreSQL/MySQL legs (this doesn't touch the claim path).

Docs: new row in the ox_prune table in docs/configuration.md; CHANGELOG entry under Unreleased → Added. I didn't touch the generated docs/llms-full.txt.

🤖 Generated with Claude Code

[burakeyler]

Same issue as #41: docs/llms-full.txt was stale after this PR's docs update, which failed test_llms_full_is_current on every matrix job. Pushed a follow-up commit regenerating it with python tools/build_llms_full.py.

Verified locally:

• pytest tests/test_llms_full.py — 3 passed
• ruff check . — clean
• mypy --strict src/ — clean

[PhiLily]

Thanks @burakeyler , merged. I pushed a small commit on top: the README's ox_prune table now has the --queue row, the docs say plainly that a --queue run still prunes every schedule's old ticks, and there's a test for that.

A few things that'll save a round next time:

• When you add or change a flag, update every table that lists it. For ox_prune that's the README and docs/configuration.md.
• After any docs edit, run python tools/build_llms_full.py and commit the result.
• If the docs describe a behaviour, add a test that pins it.

and nice work! looking forward seeing you here again:)