fix static notebook deploy passing python version string instead of interpreter path

[karawoo]

Intent

rsconnect deploy notebook --static runs nbconvert locally to render the notebook before bundling it. Since #648, that call was passed environment.python, which is the Python version string recorded for the manifest (e.g. 3.12.1), not an interpreter path. nbconvert was invoked as a subprocess named 3.12.1, so every static notebook deploy failed with Unable to include the file 3.12.1 in the bundle: No such file or directory.

Fixes #721

Approach

Use environment.python_interpreter (the resolved executable path, defaulting to sys.executable) instead of environment.python.

Automated Tests

Added a regression test that deploys a static notebook with --override-python-version set to a value that would visibly diverge from the real interpreter path, and asserts nbconvert is invoked with the real path. Confirmed it fails without the fix.

Directions for Reviewers

Checklist

• I have updated CHANGELOG.md to cover notable changes.
• I have updated all related GitHub issues to reflect their current state.
• I have run the rsconnect-python-tests-at-night workflow in Connect against this feature branch.

[github-actions]

PR Preview Action v1.8.1
Preview removed because the pull request was closed.
2026-07-14 19:07 UTC

[github-actions]

☂️ Python Coverage

current status: ✅

Overall Coverage

Lines	Covered	Coverage	Threshold	Status
7636	6352	83%	0%	🟢

New Files

No new covered files...

Modified Files

File	Coverage	Status
rsconnect/main.py	81%	🟢
TOTAL	81%	🟢

updated for commit: fe56dd7 by action🐍

[edavidaja]

I was having some difficulty discerning the intent of the tests relative to the fix, 🤖 suggested something like:

import sys
from os.path import join
from types import SimpleNamespace
from unittest import mock

from click.testing import CliRunner

from rsconnect.main import cli, make_notebook_html_bundle
from tests.utils import apply_common_args, get_dir


def test_deploy_notebook_static_passes_interpreter_not_version():
    # Regression test for #721. For a static notebook deploy, deploy_notebook must
    # hand make_notebook_html_bundle the local interpreter *path* (used to run
    # nbconvert), not environment.python, which is only a version string and would
    # make nbconvert fail with "No such file or directory".
    #
    # In practice these always diverge: environment.python comes from
    # sys.version_info ("3.12.1"), environment.python_interpreter from sys.executable.
    fake_env = SimpleNamespace(python="3.12.1", python_interpreter=sys.executable)

    with mock.patch(
        "rsconnect.main.Environment.create_python_environment", return_value=fake_env
    ), mock.patch("rsconnect.main.RSConnectExecutor") as MockExecutor:
        ce = MockExecutor.return_value

        runner = CliRunner()
        args = apply_common_args(
            ["deploy", "notebook", get_dir(join("pip1", "dummy.ipynb"))],
            server="http://fake_server",
            key="FAKE_API_KEY",
        )
        args += ["--static", "--no-verify"]
        result = runner.invoke(cli, args)

    assert result.exit_code == 0, result.output
    # make_bundle(make_notebook_html_bundle, file, <python>, hide_all, hide_tagged)
    bundle_args = ce.make_bundle.call_args.args
    assert bundle_args[0] is make_notebook_html_bundle
    assert bundle_args[2] == sys.executable