test: add unit tests for managers/utils.py and template_manager.py by Xio-Shark · Pull Request #21227 · sgl-project/sglang

Xio-Shark · 2026-03-23T20:40:37Z

Part of #20865: add CPU-only unit tests for managers/ module.

test_utils.py (13 tests):

validate_input_length: normal/boundary/reject/truncate
get_alloc_len_per_decode: no spec/basic spec/draft tokens/page_size error
GenerationBatchResult: default and custom fields

test_template_manager.py (12 tests):

_detect_reasoning_pattern: None/empty/no-think/with-think/partial
guess_chat_template_from_model_path: known/unknown model
_select_named_template: empty/single/preferred/not-found/multi
_load_explicit_chat_template: invalid raises/builtin sets

gemini-code-assist · 2026-03-23T20:40:51Z

Summary of Changes

Hello, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request significantly improves the test coverage and reliability of the sglang.srt.managers module by introducing dedicated CPU-only unit tests for utils.py and template_manager.py. These new tests validate critical functionalities related to input processing, speculative decoding resource allocation, and chat template handling, ensuring robust behavior and preventing regressions in these core components.

Highlights

Unit Tests for managers/utils.py: Comprehensive CPU-only unit tests were added for validate_input_length, get_alloc_len_per_decode, and the GenerationBatchResult dataclass, covering normal, boundary, and error conditions.
Unit Tests for managers/template_manager.py: Extensive CPU-only unit tests were introduced for TemplateManager, verifying its initialization, _detect_reasoning_pattern, guess_chat_template_from_model_path, _select_named_template, and _load_explicit_chat_template methods.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	`/gemini review`	Performs a code review for the current pull request in its current state.
Pull Request Summary	`/gemini summary`	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	`/gemini help`	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for GitHub and other Google products, sign up here.

Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

gemini-code-assist

Code Review

This pull request introduces a solid suite of unit tests for template_manager.py and utils.py, significantly improving test coverage for the managers module. The tests are well-structured and cover a good range of scenarios, including normal operation, boundary conditions, and error cases. I have a couple of minor suggestions to improve the clarity and maintainability of the new tests.

gemini-code-assist · 2026-03-23T20:43:00Z

test/registered/unit/managers/test_template_manager.py

+        mock_tm = MagicMock()
+        with self.assertRaises(ValueError):
+            self.tm._select_named_template({}, mock_tm)


The variable mock_tm is used to mock a TokenizerManager, but its name could be confused with self.tm, which is an instance of TemplateManager. For better readability, consider renaming mock_tm to mock_tokenizer_manager throughout the test classes that use it (e.g., TestSelectNamedTemplate and TestLoadExplicitChatTemplate).

Suggested change

mock_tm = MagicMock()

with self.assertRaises(ValueError):

self.tm._select_named_template({}, mock_tm)

mock_tokenizer_manager = MagicMock()

with self.assertRaises(ValueError):

self.tm._select_named_template({}, mock_tokenizer_manager)

gemini-code-assist · 2026-03-23T20:43:00Z

test/registered/unit/managers/test_utils.py

+        )
+        result = get_alloc_len_per_decode(args)
+        # page_size > 1 but topk == 1, so no error
+        self.assertEqual(result, max(3, 5))


It's a good practice in unit testing to assert against explicit, pre-calculated values rather than expressions that replicate the implementation's logic. This makes the test's intent clearer and decouples it from the implementation details.

Suggested change

self.assertEqual(result, max(3, 5))

self.assertEqual(result, 5)

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: b1da11d53c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

chatgpt-codex-connector · 2026-03-23T20:44:21Z

test/registered/unit/managers/test_utils.py

+
+from sglang.test.ci.ci_register import register_cpu_ci
+
+register_cpu_ci(est_time=5, suite="stage-a-cpu-only")


Register these unit tests in the CPU CI suite

The new manager tests are tagged with suite="stage-a-cpu-only", but the CPU PR workflow only runs python3 run_suite.py --hw cpu --suite stage-a-test-cpu (.github/workflows/pr-test.yml lines 929-935), and test/run_suite.py only lists stage-a-test-cpu as a per-commit CPU suite. As a result, neither this file nor the companion test_template_manager.py will execute in CI, so regressions in managers/utils.py and template_manager.py can still merge unnoticed.

Useful? React with 👍 / 👎.

…ger.py (sgl-project#20865) Add CPU-only unit tests for two modules under sglang.srt.managers/: test_utils.py (13 tests): - validate_input_length: normal pass, boundary (==max), reject, truncate, truncate order preservation, empty input - get_alloc_len_per_decode: no speculation, basic spec, draft>steps, topk==1 with large page_size, page_size>1 && topk>1 raises - GenerationBatchResult: default fields, custom fields test_template_manager.py (12 tests): - __init__: default values - _detect_reasoning_pattern: None, empty, no-think, with-think, partial - guess_chat_template_from_model_path: known model, unknown model - _select_named_template: empty dict, single, preferred, not-found, first-of-multi - _load_explicit_chat_template: invalid raises, builtin sets name

- Remove unused imports: dataclasses, patch - Break long lines to comply with line-length limit

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 5328540b33

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

chatgpt-codex-connector · 2026-03-23T20:53:22Z

test/registered/unit/managers/test_template_manager.py

+    def test_template_with_think_tag_returns_true(self):
+        template = r"<|im_start|>assistant\n<think>\n"
+        self.assertTrue(self.tm._detect_reasoning_pattern(template))


Exercise force-reasoning detection with real newlines

load_chat_template() feeds tokenizer.chat_template straight into _detect_reasoning_pattern(), and both _load_jinja_template() and the built-in chat templates in python/sglang/lang/chat_template.py produce strings with real newline characters (for example "<|im_start|>assistant\n"), not the literal \\n sequences used in this test. In that scenario the current regex does not match, so this test would stay green while force-reasoning detection is broken for the templates users actually load.

Useful? React with 👍 / 👎.

chatgpt-codex-connector · 2026-03-23T20:53:22Z

test/registered/unit/managers/test_utils.py

+    def test_exactly_at_limit_returns_error(self):
+        """Input length == max is treated as exceeding (>= check)."""
+        req = self._make_req(200)
+        result = validate_input_length(
+            req, max_req_input_len=200, allow_auto_truncate=False


Stop treating max-length prompts as out of bounds

The helper documents max_req_input_len as the maximum allowed input length, and --allow-auto-truncate is described as applying only when a request exceeds that limit. tp_worker.py also already reserves five tokens of headroom before computing max_req_input_len, so asserting that len == max_req_input_len must fail bakes the current off-by-one behavior into CI and will block a future fix that accepts prompts exactly at the published limit.

Useful? React with 👍 / 👎.

Xio-Shark · 2026-03-23T21:02:09Z

@ispobock Could you please add the run-ci label to trigger CI? This PR adds 25 CPU-only unit tests for managers/utils.py and managers/template_manager.py. Thanks!

chatgpt-codex-connector · 2026-03-24T05:47:26Z

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

gemini-code-assist bot reviewed Mar 23, 2026

View reviewed changes

chatgpt-codex-connector bot reviewed Mar 23, 2026

View reviewed changes

Xio-Shark added 2 commits March 24, 2026 04:48

style: fix ruff lint errors in test_utils.py

5328540

- Remove unused imports: dataclasses, patch - Break long lines to comply with line-length limit

Xio-Shark force-pushed the unit-test/managers-utils-template-manager branch from 0403f16 to 5328540 Compare March 23, 2026 20:48

Xio-Shark mentioned this pull request Mar 23, 2026

[Feature] Improve Unit Test Coverage #20865

Open

13 tasks

chatgpt-codex-connector bot reviewed Mar 23, 2026

View reviewed changes

Merge branch 'main' into unit-test/managers-utils-template-manager

853cc82

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

test: add unit tests for managers/utils.py and template_manager.py#21227

test: add unit tests for managers/utils.py and template_manager.py#21227
Xio-Shark wants to merge 3 commits intosgl-project:mainfrom
Xio-Shark:unit-test/managers-utils-template-manager

Xio-Shark commented Mar 23, 2026

Uh oh!

gemini-code-assist bot commented Mar 23, 2026

Uh oh!

gemini-code-assist bot left a comment

Uh oh!

gemini-code-assist bot Mar 23, 2026

Uh oh!

gemini-code-assist bot Mar 23, 2026

Uh oh!

chatgpt-codex-connector bot left a comment

Uh oh!

chatgpt-codex-connector bot Mar 23, 2026

Uh oh!

chatgpt-codex-connector bot left a comment

Uh oh!

chatgpt-codex-connector bot Mar 23, 2026

Uh oh!

chatgpt-codex-connector bot Mar 23, 2026

Uh oh!

Xio-Shark commented Mar 23, 2026

Uh oh!

chatgpt-codex-connector bot commented Mar 24, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

	self.assertEqual(result, max(3, 5))
	self.assertEqual(result, 5)


		from sglang.test.ci.ci_register import register_cpu_ci

		register_cpu_ci(est_time=5, suite="stage-a-cpu-only")

Conversation

Xio-Shark commented Mar 23, 2026

Uh oh!

gemini-code-assist bot commented Mar 23, 2026

Summary of Changes

Highlights

Footnotes

Uh oh!

gemini-code-assist bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

gemini-code-assist bot Mar 23, 2026

Choose a reason for hiding this comment

Uh oh!

gemini-code-assist bot Mar 23, 2026

Choose a reason for hiding this comment

Uh oh!

chatgpt-codex-connector bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector bot Mar 23, 2026

Choose a reason for hiding this comment

Uh oh!

chatgpt-codex-connector bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector bot Mar 23, 2026

Choose a reason for hiding this comment

Uh oh!

chatgpt-codex-connector bot Mar 23, 2026

Choose a reason for hiding this comment

Uh oh!

Xio-Shark commented Mar 23, 2026

Uh oh!

chatgpt-codex-connector bot commented Mar 24, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant