LCORE-1350: Increased type-safety of tool utilities by asimurka · Pull Request #1229 · lightspeed-core/lightspeed-stack

asimurka · 2026-02-26T13:45:18Z

Description

This PR refactors:

utility functions for getting available tools for request to use LLS API model types instead of plain dicts to make them more type-safe.
extends ResponsesApiParams model to contain all currently available attributes supported by LLS
makes query requests explicitly drop attributes that are not set (None)
introduces optional attributes notation

Type of change

Tools used to create PR

Identify any AI code assistants used in this PR (for transparency and review context)

Assisted-by: Cursor

Related Tickets & Documents

Related Issue #
Closes # LCORE-1350

Checklist before requesting a review

I have performed a self-review of my code.
PR has passed all pre-merge test jobs.
If it is a core feature, I have added thorough tests.

Testing

Please provide detailed steps to perform tests related to this code change.
How were the fix/results from this change verified? Please provide relevant screenshots or results.

Summary by CodeRabbit

New Features
- Expanded response configuration with new parameters: conversation history, response inclusion options, inference iterations, tool call limits, metadata, parallel tool execution, previous response references, custom prompts, temperature control, and tool choice strategies.
- Enhanced input flexibility supporting both text and rich structured response content.
Bug Fixes
- Improved payload cleanliness by excluding null fields from API responses.
- Refined input type handling with explicit string casting in moderation and persistence workflows.
Refactor
- Standardized tool definitions to use structured types instead of generic dictionaries for better type safety and consistency.

coderabbitai · 2026-02-26T13:45:50Z

Walkthrough

This PR refactors the type system for response handling and tools management. It introduces structured type definitions (InputTool, InputToolFileSearch, InputToolMCP) to replace dict-based tool representations, expands ResponsesApiParams with new fields for enhanced response composition, standardizes Optional type annotations throughout responses utilities, and adds explicit string casting and None-field filtering in endpoint response handlers.

Changes

Cohort / File(s)	Summary
Query Endpoints `src/app/endpoints/query.py`, `src/app/endpoints/streaming_query.py`	Added explicit str casting for input to moderation checks and conversation persistence; switched to `model_dump(exclude_none=True)` to filter None fields from response payloads.
Type and Responses Utilities `src/utils/responses.py`, `src/utils/types.py`, `src/utils/query.py`	Introduced public types InputTool, InputToolFileSearch, InputToolMCP; replaced dict-based tool representations with structured objects throughout; expanded ResponsesApiParams with conversation, include, max_infer_iters, max_tool_calls, metadata, parallel_tool_calls, previous_response_id, prompt, stream, temperature, text, tool_choice fields; standardized Optional typing in ~15+ function signatures; made TranscriptMetadata fields optional.
Tests `tests/unit/utils/test_responses.py`	Updated test assertions to use new InputToolFileSearch and InputToolMCP types with attribute access (e.g., `tool.type`, `tool.headers`) instead of dict-style access (e.g., `tool["type"]`, `tool["headers"]`).

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~25 minutes

Possibly related PRs

LCORE-1350: Refactored utilities for more generic use #1198: Modifies the same modules (responses.py, types.py, query endpoints) for Responses API changes with direct code-level overlap.
LCORE-1051: final updates #985: Updates the same endpoint functions (retrieve_response) with standardized Optional typing and function signatures.
LCORE-1264 + LCORE-1267: Created and documented LCORE OpenResponses specification #1211: Introduces typed InputTool and Response types that parallel this PR's tool type refactoring.

Suggested reviewers

are-ces
tisnik
jrobertboos

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title accurately summarizes the main focus of the PR, which is refactoring tool utilities to use structured API types instead of dicts and improving type safety throughout the codebase.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.

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

✨ Finishing Touches

🧪 Generate unit tests (beta)

Create PR with unit tests
Post copyable unit tests in a comment

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

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

asimurka · 2026-02-26T13:47:50Z

src/app/endpoints/query.py

    try:
        moderation_result = await run_shield_moderation(
-            client, responses_params.input, shield_ids
+            client, cast(str, responses_params.input), shield_ids


Query endpoints always use string-like inputs so it is safe to cast.

asimurka · 2026-02-26T13:48:30Z

src/app/endpoints/query.py

-        response = await client.responses.create(**responses_params.model_dump())
+        response = await client.responses.create(
+            **responses_params.model_dump(exclude_none=True)
+        )


Explicitly drop unset attributes (queries use reduced set of responses attributes).

asimurka · 2026-02-26T13:50:13Z

src/utils/responses.py

-            return tool.get("vector_store_ids", [])
-    return []
+        if tool.type == "file_search":
+            vector_store_ids.extend(tool.vector_store_ids)


Minor functional change - solves the case when there are multiple file_search objects.

Thanks, it might be relevant in the future! 👍

asimurka · 2026-02-26T13:50:46Z

src/utils/responses.py

-        tools.append(tool_def)
+        authorization = headers.pop("Authorization", None)
+        tools.append(
+            InputToolMCP(


Creates the whole object at once

asimurka · 2026-02-26T13:51:16Z

src/utils/types.py

    tool_results: list[dict[str, Any]] = Field(default_factory=list)
-
-
-ResponseItem: TypeAlias = (


Moved at the top of the module

asimurka · 2026-02-26T13:51:40Z

src/utils/types.py

+ResponseInput: TypeAlias = str | list[ResponseItem]
+

 class ResponsesApiParams(BaseModel):


Extended model with all currently available attributes in LLS.

coderabbitai

Actionable comments posted: 3

🤖 Prompt for all review comments with AI agents

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

Inline comments:
In `@src/app/endpoints/query.py`:
- Around line 296-305: Replace the unsafe typing cast usage by converting
structured input to a string before passing it to moderation and conversation
functions: call content_to_str(responses_params.input) and use that returned
string in run_shield_moderation(...) and append_turn_to_conversation(...).
Update the calls where cast(str, responses_params.input) appears (refer to
responses_params.input, run_shield_moderation, and append_turn_to_conversation)
so they pass the content_to_str(...) result instead, ensuring you import
content_to_str from src.utils.types if not already imported.

In `@src/app/endpoints/streaming_query.py`:
- Around line 287-299: Replace the unsafe type-only cast of
responses_params.input with a runtime coercion using content_to_str from
utils.types: import content_to_str and use
content_to_str(responses_params.input) when calling run_shield_moderation and
append_turn_to_conversation (locations involving run_shield_moderation,
moderation_result, append_turn_to_conversation and responses_params.input) so
both string and list[ResponseItem] inputs are handled correctly at runtime;
mirror the same change in the similar patterns found in
src/app/endpoints/query.py.

In `@src/utils/responses.py`:
- Around line 298-299: The docstring for the parameter "tools" still mentions
"typed InputTool or dict" which is stale; update the docstring to state that
"tools" is a sequence/list of typed InputTool instances (or the exact type used
in this module) and remove any mention of dict so it reflects the typed-only API
(referencing the "tools" parameter and the InputTool type in the docstring).

ℹ️ Review info

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 9b21300 and 0965026.

📒 Files selected for processing (6)

src/app/endpoints/query.py
src/app/endpoints/streaming_query.py
src/utils/query.py
src/utils/responses.py
src/utils/types.py
tests/unit/utils/test_responses.py

src/app/endpoints/query.py

src/app/endpoints/streaming_query.py

src/utils/responses.py

tisnik

LGTM, indeed, now it's more readable

max-svistunov · 2026-02-26T14:42:21Z

src/utils/responses.py

+        tool.server_url: tool.headers
        for tool in tools
-        if tool.get("type") == "mcp" and tool.get("headers") and tool.get("server_url")
+        if tool.type == "mcp" and tool.headers


Do we have assurance that server_url won't be e.g. "" or None (I think InputToolMCP(server_label="x", server_url="") would be valid but we don't want this)? If not, better bring back and tool.server_url.

max-svistunov · 2026-02-26T14:46:01Z

src/utils/responses.py


    Parameters:
-        tools: The prepared tools list from ResponsesApiParams.
+        tools: The prepared tools list (typed InputTool or dict).


Given the change you've made, this should be more accurate?
The prepared tools list of InputTool objects.

Increased type-safety of utilities

0965026

asimurka commented Feb 26, 2026

View reviewed changes

coderabbitai bot reviewed Feb 26, 2026

View reviewed changes

src/app/endpoints/query.py Show resolved Hide resolved

src/app/endpoints/streaming_query.py Show resolved Hide resolved

src/utils/responses.py Show resolved Hide resolved

asimurka requested review from max-svistunov and tisnik February 26, 2026 13:57

tisnik approved these changes Feb 26, 2026

View reviewed changes

tisnik merged commit ca21850 into lightspeed-core:main Feb 26, 2026
23 of 24 checks passed

max-svistunov reviewed Feb 26, 2026

View reviewed changes

coderabbitai bot mentioned this pull request Feb 27, 2026

LCORE-86: Prioritize BYOK content over built-in content #1208

Open

19 tasks

		tool_results: list[dict[str, Any]] = Field(default_factory=list)


		ResponseItem: TypeAlias = (

		ResponseInput: TypeAlias = str \| list[ResponseItem]


		class ResponsesApiParams(BaseModel):

Conversation

asimurka commented Feb 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Type of change

Tools used to create PR

Related Tickets & Documents

Checklist before requesting a review

Testing

Summary by CodeRabbit

Uh oh!

coderabbitai bot commented Feb 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Walkthrough

Changes

Estimated code review effort

Possibly related PRs

Suggested reviewers

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

coderabbitai bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

tisnik left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

asimurka commented Feb 26, 2026 •

edited

Loading

coderabbitai bot commented Feb 26, 2026 •

edited

Loading