add a scriptlet to put cuopt_grpc_server on the default path

[tmckayus]

This changes adds a scriptlet for cuopt_grpc_server so that after installation of the cuopt-server-cuXX packages the binary is on the default path, similar to cuopt_cli.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: 696db095-75cf-458f-a897-233b0926a213

📥 Commits

Reviewing files that changed from the base of the PR and between 714fa48 and 1d3fa00.

📒 Files selected for processing (2)

• python/cuopt_server/cuopt_server/_grpc_server_wrapper.py
• python/cuopt_server/cuopt_server/tests/test_grpc_server_entry_point.py

🚧 Files skipped from review as they are similar to previous changes (2)

• python/cuopt_server/cuopt_server/_grpc_server_wrapper.py
• python/cuopt_server/cuopt_server/tests/test_grpc_server_entry_point.py

---

📝 Walkthrough

Walkthrough

Adds a console script entry cuopt_grpc_server, a Python wrapper main() that launches the packaged bin/cuopt_grpc_server binary forwarding args, and corresponding conda recipe test and unit tests validating the installed executable and its --help output.

Changes

Cohort / File(s)	Summary
Console script configuration python/cuopt_server/pyproject.toml	Added [project.scripts] entry registering cuopt_grpc_server = "cuopt_server._grpc_server_wrapper:main".
Python wrapper module python/cuopt_server/cuopt_server/_grpc_server_wrapper.py	New module defining main() that imports libcuopt, builds the path to bin/cuopt_grpc_server in the library install location, invokes it with subprocess.call forwarding sys.argv[1:], and exits with the child return code.
Unit tests (Python) python/cuopt_server/cuopt_server/tests/test_grpc_server_entry_point.py	New tests: one checks shutil.which("cuopt_grpc_server") is non-empty; another runs cuopt_grpc_server --help (10s timeout) and asserts return code 0 and stdout contains cuopt_grpc_server.
Conda recipe tests conda/recipes/cuopt-server/recipe.yaml	Added tests entry using package_contents to assert bin/cuopt_grpc_server exists in the built package.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 33.33% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title 'add a scriptlet to put cuopt_grpc_server on the default path' accurately describes the main objective of the pull request—adding a mechanism to make the cuopt_grpc_server binary accessible on the default PATH after package installation.
Description check	✅ Passed	The description directly relates to the changeset, explaining that the pull request adds a scriptlet for cuopt_grpc_server to place the binary on the default path following package installation, mirroring the existing behavior for cuopt_cli.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[rgsl888prabhu]

@tmckayus Would we need a test for this ?

[tmckayus]

added, and realize I left out the wrapper

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@python/cuopt_server/cuopt_server/tests/test_grpc_server_entry_point.py`:
- Around line 14-26: The test_cuopt_grpc_server_help currently calls
subprocess.run on the bare "cuopt_grpc_server" and only checks stdout, which can
raise FileNotFoundError and miss help output printed to stderr; update the test
to first locate the executable with shutil.which("cuopt_grpc_server") and call
pytest.skip if not found, or alternatively invoke the module via sys.executable
-m if an importable entrypoint exists, then call subprocess.run with
stderr=subprocess.STDOUT (or combine stdout/stderr) and a controlled env (e.g.,
env=os.environ.copy() with GPU vars like CUDA_VISIBLE_DEVICES cleared) so the
call is stream-agnostic and environment-isolated, and assert on the combined
output and returncode.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: 9c80b148-7cb0-405c-b790-842b0fdb0941

📥 Commits

Reviewing files that changed from the base of the PR and between 2d48c97 and 714fa48.

📒 Files selected for processing (4)

• conda/recipes/cuopt-server/recipe.yaml
• python/cuopt_server/cuopt_server/_grpc_server_wrapper.py
• python/cuopt_server/cuopt_server/tests/test_grpc_server_entry_point.py
• python/cuopt_server/pyproject.toml

✅ Files skipped from review due to trivial changes (1)

• python/cuopt_server/pyproject.toml

[coderabbitai]

⚠️ Potential issue | 🟡 Minor

Make test_cuopt_grpc_server_help self-contained and stream-agnostic.

On Line 16, invoking the bare command can raise FileNotFoundError when this test runs independently in a misconfigured environment, and Line 24 only inspects stdout even though many CLIs print help to stderr.

Proposed test hardening

 def test_cuopt_grpc_server_help():
+    exe = shutil.which("cuopt_grpc_server")
+    assert exe is not None, "cuopt_grpc_server should be on PATH before help test"
+
     result = subprocess.run(
-        ["cuopt_grpc_server", "--help"],
+        [exe, "--help"],
         capture_output=True,
         text=True,
         timeout=10,
     )
     assert result.returncode == 0, (
         f"cuopt_grpc_server --help failed (rc={result.returncode}): {result.stderr}"
     )
-    assert "cuopt_grpc_server" in result.stdout, (
-        f"Expected 'cuopt_grpc_server' in --help output, got: {result.stdout}"
+    output = f"{result.stdout}\n{result.stderr}"
+    assert "cuopt_grpc_server" in output, (
+        f"Expected 'cuopt_grpc_server' in --help output, got: {output}"
     )

As per coding guidelines, "Ensure test isolation: prevent GPU state, cached memory, and global variables from leaking between test cases; verify each test independently initializes its environment".

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@python/cuopt_server/cuopt_server/tests/test_grpc_server_entry_point.py`
around lines 14 - 26, The test_cuopt_grpc_server_help currently calls
subprocess.run on the bare "cuopt_grpc_server" and only checks stdout, which can
raise FileNotFoundError and miss help output printed to stderr; update the test
to first locate the executable with shutil.which("cuopt_grpc_server") and call
pytest.skip if not found, or alternatively invoke the module via sys.executable
-m if an importable entrypoint exists, then call subprocess.run with
stderr=subprocess.STDOUT (or combine stdout/stderr) and a controlled env (e.g.,
env=os.environ.copy() with GPU vars like CUDA_VISIBLE_DEVICES cleared) so the
call is stream-agnostic and environment-isolated, and assert on the combined
output and returncode.