[jayfeng7]

Closes #4402

Description

Two names advertised in submodule __all__ lists could not be resolved at import time:/n/n- camel.datagen.evol_instruct.__all__ advertised MathEvolInstructTemplates (defined in templates.py) without importing it — from camel.datagen.evol_instruct import * raised AttributeError, direct import raised ImportError.

• camel.utils.__all__ advertised generate_prompt_for_structured_output (defined in utils/commons.py) without including it in the from .commons import (...) block — direct import raised ImportError.

__all__ is a runtime contract (star imports resolve every entry with getattr), so both failures hit callers directly. Confirmed on camel-ai==0.2.90 from PyPI.

This PR adds the two missing imports and extends test/test_all_exports.py — which previously pinned only the root package — with star-import resolution checks for both submodules, so the same regression class cannot silently return.

Note: the phantom DescriptionVideoPromptTemplateDict entry in camel.prompts.__all__ is intentionally not touched here, as it is already covered by #3810.

What is the purpose of this pull request?

• Bug fix
• New Feature
• Documentation update
• Other

Verification

Red (before, camel-ai==0.2.90 PyPI wheel, Python 3.11.8):

>>> from camel.utils import generate_prompt_for_structured_output
ImportError: cannot import name 'generate_prompt_for_structured_output' from 'camel.utils'
>>> from camel.datagen.evol_instruct import *
AttributeError: module 'camel.datagen.evol_instruct' has no attribute 'MathEvolInstructTemplates'

Green (this branch, installed into the same environment):

>>> from camel.utils import generate_prompt_for_structured_output   # OK
>>> from camel.datagen.evol_instruct import *                        # OK, MathEvolInstructTemplates resolves
>>> pytest test/test_all_exports.py -v
4 passed

The two new tests cover both submodules; the two pre-existing tests in the file keep passing.

Checklist

• I have read and agree to the AI-Generated Code Policy (required)
• I have linked this PR to an issue (required)
• I have checked if any dependencies need to be added or updated in pyproject.toml and run uv lock — no dependency changes
• I have updated the tests accordingly (required for a bug fix or a new feature)
• I have updated the documentation if needed — no docs impact
• I have added examples if this is a new feature — not applicable

AI assistance disclosure

Tool(s) used: AI coding assistant (agentic coding session).

How you used it: The assistant located the two defects via a static __all__ audit, produced the two import additions and the regression tests, and ran the red/green verification shown above against the PyPI wheel. I have read and understand every line of the change, reproduced both failures independently, and take full responsibility for the diff.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

• Configuration used: Repository UI
• Review profile: CHILL
• Plan: Advanced
• Run ID: cb555d25-fa66-41c3-91ef-dc30396ebc3a

---

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

---

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

[jayfeng7]

Convenience note for whoever reviews this, because the check list reads worse than the diff.

All 6 red checks on this head are structural to fork PRs and cannot be turned green from the author side. I confirmed this from the job logs and annotations rather than guessing:

5 × pytest_* — fail during collection, before any test of this diff runs. Every one exits 2 with ValueError: Missing or empty required API keys in environment variables: OPENAI_API_KEY (and openai.OpenAIError: Missing credentials in the MCP server case). The five modules are test/agents/test_chat_agent.py, test/agents/test_knowledge_agent.py, test/agents/test_role_playing.py, test/services/test_agent_mcp_server.py, test/toolkits/test_function_tool_decorator.py. Repository secrets are not exposed to pull_request runs from a fork, so these jobs cannot pass on jayfeng7:... no matter what the diff contains — and none of those files is touched here (this PR only changes camel/utils/... and camel/datagen/... plus one test file).

1 × add-labels — Unhandled error: HttpError: Resource not accessible by integration. The workflow writes labels with the run's GITHUB_TOKEN, which is read-only on a fork head.

So the signal that actually describes this change is the 7 green checks; the 6 red ones are the environment. If there is a preferred way to get a real pytest signal on a fork head — a label that unlocks secrets, a re-run from the base repo, or a specific module you want exercised — point me at it and I will do that instead of asking you to read around the noise.

For the record, the diff is two import lines plus a test: camel/datagen/evol_instruct/__init__.py lists MathEvolInstructTemplates in __all__ without importing it, and camel/utils/__init__.py lists generate_prompt_for_structured_output without importing it. from camel.utils import * walks __all__ with getattr, so both names raised AttributeError on a star import while the package advertised them as public. The fix adds the two missing imports and a test that asserts every __all__ entry resolves under import *.