enh: sandbox client

[echarles]

No description provided.

[Copilot]

Pull request overview

This PR enhances the sandbox “client” surface by standardizing exposure of the underlying kernel client across sandbox variants and aligning Jupyter/Docker sandboxes with the jupyter_kernel_client.JupyterKernelClient API.

Changes:

• Switch Jupyter/Docker sandboxes from KernelClient to JupyterKernelClient.
• Introduce ISandboxClient/IJupyterKernelClient typing protocols and add a kernel_client property to kernel-backed sandboxes (and a default None in Sandbox).
• Export the new interfaces from code_sandboxes and bump the package version.

Reviewed changes

Copilot reviewed 9 out of 9 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
tests/test_jupyter.py	Updates test stubs/expectations to patch JupyterKernelClient.
code_sandboxes/interfaces.py	Adds typing protocols for sandbox kernel clients with optional-dependency fallback.
code_sandboxes/base.py	Adds a default kernel_client property returning None for non-kernel sandboxes.
code_sandboxes/jupyter_sandbox.py	Uses JupyterKernelClient, types _client, and exposes it via kernel_client.
code_sandboxes/docker_sandbox.py	Uses JupyterKernelClient and exposes it via kernel_client.
code_sandboxes/colab_sandbox.py	Exposes the underlying Colab client via kernel_client.
code_sandboxes/kaggle_sandbox.py	Exposes the underlying Kaggle client via kernel_client.
code_sandboxes/init.py	Re-exports IJupyterKernelClient and ISandboxClient.
code_sandboxes/version.py	Updates the package version to 0.16.0.

Comments suppressed due to low confidence (2)

code_sandboxes/jupyter_sandbox.py:348

• JupyterSandbox.start() now imports JupyterKernelClient directly, but pyproject.toml does not pin a minimum jupyter-kernel-client version. If a user has an older version installed that only exports KernelClient, this will raise ImportError even though the package is present. Consider a backward-compatible import fallback (or pin a minimum version).

        try:
            from jupyter_kernel_client import JupyterKernelClient
        except ImportError as exc:
            raise SandboxConfigurationError(
                "jupyter-kernel-client is required for JupyterSandbox. "
                "Install it with: pip install code-sandboxes[test]"
            ) from exc

code_sandboxes/docker_sandbox.py:133

• DockerSandbox.start() now imports JupyterKernelClient directly, but dependencies are not version-pinned. To avoid breaking environments that have an older jupyter-kernel-client (exporting KernelClient but not JupyterKernelClient), add an import fallback (or pin a minimum version).

        self._ensure_docker()
        try:
            from jupyter_kernel_client import JupyterKernelClient
        except ImportError as exc:  # pragma: no cover - optional dependency
            raise SandboxConfigurationError(
                "jupyter-kernel-client is required for DockerSandbox. "
                "Install it with: pip install code-sandboxes"
            ) from exc

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.