feat(clients): add stats() to the Python and TypeScript clients

[cl0ud08]

Summary

Adds stats() to the Python and TypeScript clients, wrapping GET /api/v1/stats. Returns a new Stats model/interface exposing tenant_count, memory_count, agent_count, plus the raw payload on .raw.

Related Issue

Closes #552.

Type of Change

• Bug fix (non-breaking change that fixes an issue)
• New feature (non-breaking change that adds functionality)
• Breaking change (fix or feature that changes existing behavior)
• Documentation update
• Refactor / internal cleanup
• Other:

How Has This Been Tested?

• pytest in clients/python: 138 passed. The 7 failures in test_interviewer_installer.py / test_interviewer_platform.py are POSIX-only (fcntl, chmod semantics) and reproduce identically on unmodified main on Windows.
• npm test in clients/typescript: 17/17 pass; tsc compiles clean.
• New tests cover: correct path with empty query string, counter parsing, missing keys defaulting to 0, and a 503 raising CauraAPIError with the right status code.

Checklist

• I have read CONTRIBUTING.md
• I have added tests that cover my changes (or explained why none are needed)
• ruff check and ruff format --check pass — see tooling note below
• mypy passes — see tooling note below
• pytest passes locally
• I have updated relevant documentation (README, docs, etc.)
• I have updated CHANGELOG.md under the Unreleased section (if user-facing)

Additional Notes

The issue describes query params (scope, memory_type, status, fleet_id, include_deleted) that don't exist on the route — public_stats() in core-api/src/core_api/routes/stats.py has an empty signature. Implemented as a no-arg method matching the route as it stands; flagged this on the issue before starting.

Two deliberate choices, both easy to reverse:

• Counters are deployment-wide (count_all(tenant_id="")), not scoped to tenant_id like every other client method. Called out explicitly in both docstrings and both READMEs so nobody misreads memory_count as their own tenant's count.
• stats() calls _raise_for_status() although health() deliberately doesn't. A wrong base_url or a deployment predating this route returns non-2xx, and a bare .json() would surface a ValueError instead of a CauraAPIError.

Tests mock the exact response shape emitted by stats.py rather than an assumed one.

CHANGELOG: not edited — it's release-please-generated with no Unreleased section. The commit message is a conventional commit, so the entry should generate on release. Happy to add one manually if that's wrong.

Tooling notes, both pre-existing and unrelated to this change:

• ruff check passes clean. ruff format --check reports 11 files, including tests/test_client.py — all are already unformatted on main, verified by checking out upstream's copy and re-running. Not reformatted here to keep the diff scoped; happy to do it as a separate PR.
• mypy passes on every file this PR touches. It reports 3 errors in interviewer/cli.py (fcntl.flock / LOCK_EX / LOCK_NB), which are Windows-only — that file is untouched here and errors identically on upstream's version.

[Eldad-Caura]

Thank you for implementing stats() in both clients and adding focused tests. The Python and TypeScript suites pass, but DCO fails because commit 355f714 does not contain your Signed-off-by trailer. Please amend and sign off that commit yourself, rebase onto current main, and force-push the updated branch; we will rerun exact-head CI and review after that.