test: add unit tests for Bedrock provider

[mvincentbb]

• inference-config: verify bedrock returns correct config and default model
• onboard: verify getSandboxInferenceConfig maps bedrock with supportsStore:false
• onboard-selection: verify curated model list, region check, and endpoint probe

Summary

Add unit tests for the Amazon Bedrock inference provider introduced in #963. Covers provider config, sandbox inference mapping,
and interactive selection UX.

Related Issue

Relates to #963

Changes

• test/inference-config.test.js: verify getProviderSelectionConfig("bedrock") returns correct OpenAI-compatible config,
  BEDROCK_API_KEY credential env, and default model (nvidia.nemotron-super-3-120b)
• test/onboard.test.js: verify getSandboxInferenceConfig maps bedrock to routed inference provider with supportsStore: false
• test/onboard-selection.test.js: verify interactive setup shows curated Bedrock models, enforces BEDROCK_REGION check, and
  probes endpoint via Chat Completions API

Type of Change

• Code change for a new feature, bug fix, or refactor.
• Code change with doc updates.
• Doc only. Prose changes without code sample modifications.
• Doc only. Includes code sample changes.

Testing

• npx prek run --all-files passes (or equivalently make check).
• npm test passes.
• make docs builds without warnings. (for doc-only changes)

Checklist

General

• I have read and followed the contributing guide.
• I have read and followed the style guide. (for doc-only changes)

Code Changes

• Formatters applied — npx prek run --all-files auto-fixes formatting (or make format for targeted runs).
• Tests added or updated for new or changed behavior.
• No secrets, API keys, or credentials committed.
• Doc pages updated for any user-facing behavior changes (new commands, changed defaults, new features, bug fixes that
  contradict existing docs).

Doc Changes

N/A — no doc changes in this PR.

Signed-off-by: BAKPATINA Vincent de Paul 47268786+mvincentbb@users.noreply.github.com

Summary by CodeRabbit

• Tests
  • Updated Bedrock model selection test to reflect a revised default model choice.
  • Added an onboarding test that simulates selecting Bedrock and verifies user-facing confirmation messages and availability of the Chat Completions API.
  • Added a sandbox inference test to confirm Bedrock is routed to the expected inference path and that API/compatibility flags are set.

[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

Run ID: 0dbcdc3c-6f29-4451-95db-21d0ad3d79b0

📥 Commits

Reviewing files that changed from the base of the PR and between 0fe0d30 and 9f2ac6d.

📒 Files selected for processing (1)

• test/onboard.test.js

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

• test/onboard.test.js

---

📝 Walkthrough

Walkthrough

Updated and added tests to validate Amazon Bedrock model selection (model id changed to nvidia.nemotron-super-3-120b), simulate interactive onboarding selecting Bedrock and Chat Completions, and verify sandbox inference mapping routes Bedrock models through the inference provider.

Changes

Cohort / File(s)	Summary
Bedrock provider config tests test/inference-config.test.js	Updated assertions to expect custom.model = nvidia.nemotron-super-3-120b for bedrock and adjusted the fallback getProviderSelectionConfig("bedrock").model.
Onboard selection flow test test/onboard-selection.test.js	Added a subprocess test that sets BEDROCK_API_KEY/BEDROCK_REGION, stubs interactive prompts to pick Bedrock and nvidia.nemotron-super-3-120b, injects a fake curl, captures logs, and asserts provider/model selection plus Chat Completions API availability.
Sandbox inference mapping test test/onboard.test.js	Added a unit test asserting getSandboxInferenceConfig("nvidia.nemotron-super-3-120b", "bedrock") maps to routed inference: providerKey: "inference", primaryModelRef prefixed with inference/, inferenceBaseUrl = https://inference.local/v1, inferenceApi = openai-completions, and inferenceCompat.supportsStore = false.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~15 minutes

Poem

🐰 I hopped through tests with nimble feet,
Bedrock picked, a model tidy and neat,
A fake curl chirps, prompts answered with glee,
Sandbox routes whisper "inference" to me,
Hooray — the burrow's checks all pass! 🥕

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately and specifically describes the main change: adding unit tests for the Bedrock provider, which is reflected across all three modified test files.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ 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.}

[coderabbitai]

🧹 Nitpick comments (1)

test/onboard-selection.test.js (1)

2375-2465: Test title overstates BEDROCK_REGION validation coverage.

Line 2375 says this validates the region check, but the case only tests the success path with BEDROCK_REGION already set (Line 2430). Add a companion negative-path test (missing BEDROCK_REGION) to ensure the guard is actually enforced.

✅ Suggested follow-up test

+  it("requires BEDROCK_REGION before accepting Bedrock selection", () => {
+    // same setup as the Bedrock success test, but omit BEDROCK_REGION
+    // assert that setup does not complete successfully and prompts/logs region guidance
+  });

🤖 Prompt for AI Agents

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

In `@test/onboard-selection.test.js` around lines 2375 - 2465, The test title is
misleading because it only covers the success path with BEDROCK_REGION set; add
a negative-path test that mirrors the existing "shows curated Bedrock models..."
test but does not set process.env.BEDROCK_REGION (remove or unset BEDROCK_REGION
in the inline script before calling setupNim) and assert that setupNim fails
(non-zero process status) and emits an error or message indicating a
missing/invalid BEDROCK_REGION; reference the same symbols used in the existing
test (setupNim, BEDROCK_API_KEY, BEDROCK_REGION, and the spawnSync-based test
harness) so the new test reuses the fake curl and prompt runner scaffolding but
expects failure when BEDROCK_REGION is absent.

🤖 Prompt for all review comments with AI agents

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

Nitpick comments:
In `@test/onboard-selection.test.js`:
- Around line 2375-2465: The test title is misleading because it only covers the
success path with BEDROCK_REGION set; add a negative-path test that mirrors the
existing "shows curated Bedrock models..." test but does not set
process.env.BEDROCK_REGION (remove or unset BEDROCK_REGION in the inline script
before calling setupNim) and assert that setupNim fails (non-zero process
status) and emits an error or message indicating a missing/invalid
BEDROCK_REGION; reference the same symbols used in the existing test (setupNim,
BEDROCK_API_KEY, BEDROCK_REGION, and the spawnSync-based test harness) so the
new test reuses the fake curl and prompt runner scaffolding but expects failure
when BEDROCK_REGION is absent.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 78a413ad-4ebe-41aa-9a1d-6dbe6f59f834

📥 Commits

Reviewing files that changed from the base of the PR and between 82789a6 and 05aec40.

📒 Files selected for processing (3)

• test/inference-config.test.js
• test/onboard-selection.test.js
• test/onboard.test.js

[cv]

@mvincentbb can you make sure your commits are signed and DCO tag has been added to the PR description, please?

[cv]

@mvincentbb Thanks for the work on the Bedrock tests and the original provider PR (#963) — the testing was thorough and the implementation was clean.

After reviewing the direction this takes us, we've decided to revert Bedrock as a first-class provider and close this PR. The core issue: adding a dedicated provider key for each CSP (Bedrock, Vertex, Azure OpenAI, etc.) doesn't scale. Each one requires config in onboard.js, a case in inference-config.js, UX menu entries, curated model lists, validation logic, tests, and docs. Bedrock's OpenAI-compatible endpoint already works through the existing compatible-endpoint option, which is the path all CSP wrappers should use.

We've added a guard test in #1212 that enforces this — it'll fail if anyone tries to add a new provider key, pointing them to the compatible-endpoint options instead.

No issue with the quality of the contribution; this is a design direction decision. If you run into friction using compatible-endpoint for Bedrock, let us know and we can improve the generic path.