fix(agents-api): Fix tool conversion issues

[creatorrr]

Signed-off-by: Diwank Singh Tomer [email protected]

---

Important

Refactor agents-api to improve tool conversion and system call handling by centralizing handler logic and streamlining argument processing.

• Behavior:
  • Refactor execute_system in execute_system.py to use get_handler for handler retrieval and streamline argument transformation.
  • Update prompt_step in prompt_step.py to format tools using format_tool and handle tool types more effectively.
  • Modify tool_call_step in tool_call_step.py to construct tool calls using construct_tool_call and handle different tool types.
• Utilities:
  • Add get_handler function in utils.py to centralize handler retrieval logic based on SystemDef.
  • Introduce _create_search_request in execute_system.py to handle search request creation.
• Models:
  • Update ExecutionInput and StepContext in tasks.py to support CreateToolRequest alongside Tool.
• Workflows:
  • Adjust TaskExecutionWorkflow in task_execution/__init__.py to handle tool calls and system calls using the refactored logic.
• Misc:
  • Add langchain-core dependency in pyproject.toml.

^{This description was created by for 0af2684. It will automatically update as commits are pushed.}

[sweep-ai]

Hey @creatorrr, here is an example of how you can ask me to improve this pull request:

@Sweep Add unit tests for the new `get_handler` function in `utils.py` to verify it correctly maps different combinations of resource, subresource, and operation to the appropriate handler functions.

📖 For more information on how to use Sweep, please read our documentation.

[ellipsis-dev]

❌ Changes requested. Incremental review on 6e9e8aa in 43 seconds

More details

• Looked at 742 lines of code in 6 files
• Skipped 0 files when reviewing.
• Skipped posting 5 drafted comments based on config settings.

1. agents-api/agents_api/activities/execute_system.py:87

• Draft comment:
  Consider using a dictionary to map conditions to request types in _create_search_request to reduce repetitive code and improve maintainability.
• Reason this comment was not posted:
  Confidence changes required: 50%
  The refactoring in execute_system.py introduces a centralized handler retrieval using get_handler, which is a good practice for maintainability. However, the _create_search_request function could be improved for clarity and maintainability by using a dictionary to map conditions to request types, reducing repetitive code.

2. agents-api/agents_api/activities/task_steps/prompt_step.py:58

• Draft comment:
  Consider adding comments to explain the logic for handling system tools using get_handler and tool_decorator for better maintainability.
• Reason this comment was not posted:
  Confidence changes required: 33%
  In prompt_step.py, the format_tool function has a section for handling system tools that uses get_handler and tool_decorator. This is a good approach for dynamic tool handling, but the function could benefit from additional comments explaining the logic for future maintainability.

3. agents-api/agents_api/activities/utils.py:234

• Draft comment:
  Consider adding a default case in get_handler to handle unexpected SystemDef inputs more gracefully.
• Reason this comment was not posted:
  Confidence changes required: 33%
  In utils.py, the get_handler function uses a match-case structure to return handlers based on SystemDef. This is a clean approach, but consider adding a default case to handle unexpected inputs more gracefully.

4. agents-api/agents_api/common/protocol/tasks.py:151

• Draft comment:
  Ensure all necessary fields are correctly mapped when converting TaskToolDef to CreateToolRequest in the tools property to avoid runtime errors.
• Reason this comment was not posted:
  Confidence changes required: 50%
  In tasks.py, the tools property in StepContext converts TaskToolDef to CreateToolRequest. This conversion is necessary for the refactor, but ensure that all necessary fields are correctly mapped to avoid runtime errors.

5. agents-api/agents_api/workflows/task_execution/__init__.py:494

• Draft comment:
  Ensure all possible tool types are covered in ToolCallStep handling to prevent unhandled cases.
• Reason this comment was not posted:
  Confidence changes required: 50%
  In task_execution/__init__.py, the run method of TaskExecutionWorkflow handles various step types. The handling of ToolCallStep for integration and api_call types is well-structured, but ensure that all possible tool types are covered to prevent unhandled cases.

Workflow ID: wflow_RLgl60ZuIp6rSWti

---

Want Ellipsis to fix these issues? Tag @ellipsis-dev in a comment. You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.

[ellipsis-dev]

👍 Looks good to me! Incremental review on c281c8c in 36 seconds

More details

• Looked at 183 lines of code in 5 files
• Skipped 0 files when reviewing.
• Skipped posting 4 drafted comments based on config settings.

1. agents-api/agents_api/activities/execute_system.py:2

• Draft comment:
  Callable is used in the code but not imported. Re-add the import statement for Callable.
• Reason this comment was not posted:
  Decided after close inspection that this draft comment was likely wrong and/or not actionable:
  The comment suggests that 'Callable' is used in the code, but I don't see any evidence of 'Callable' being used in the provided file content. The removal of the import seems justified as there is no usage of 'Callable'.
  I might be missing some context or usage of 'Callable' that is not visible in the provided file content. However, based on the current information, there is no evidence of its usage.
  Given the file content provided, there is no indication that 'Callable' is used. Therefore, the comment seems incorrect.
  The comment should be deleted because there is no evidence that 'Callable' is used in the file.

2. agents-api/agents_api/activities/task_steps/prompt_step.py:31

• Draft comment:
  Use an if statement to check for the existence of attributes before accessing them, instead of using the and operator. This applies to lines 31, 32, 33, 34, 35, 36, 48, and 49.
• Reason this comment was not posted:
  Decided after close inspection that this draft comment was likely wrong and/or not actionable:
  The use of 'and' for conditional attribute access is a common and concise pattern in Python. The suggestion to use 'if' statements would make the code more verbose without a clear benefit. The current implementation is clear and idiomatic.
  I might be overlooking potential readability improvements or specific coding standards that prefer 'if' statements over 'and'.
  The current use of 'and' is a widely accepted practice in Python for such cases, and changing it to 'if' statements would unnecessarily complicate the code.
  The comment should be deleted as the current use of 'and' is idiomatic and does not require a change.

3. agents-api/agents_api/activities/task_steps/prompt_step.py:221

• Draft comment:
  Consider refactoring the unwrap logic into a separate function to avoid code duplication and improve maintainability.
• Reason this comment was not posted:
  Confidence changes required: 50%
  In prompt_step.py, the unwrap logic is duplicated. It would be better to refactor this logic into a separate function to avoid code duplication and improve maintainability.

4. agents-api/agents_api/activities/task_steps/prompt_step.py:247

• Draft comment:
  Consider including more context in the ApplicationError message to aid in debugging. For example, include the tool name and expected tool types.
• Reason this comment was not posted:
  Confidence changes required: 50%
  In prompt_step.py, the ApplicationError is raised with a message that might not be informative enough for debugging. It would be helpful to include more context in the error message.

Workflow ID: wflow_jmKi7nGkEsacQ2SL

---

You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.

[ellipsis-dev]

👍 Looks good to me! Incremental review on 5953d72 in 12 seconds

More details

• Looked at 12 lines of code in 1 files
• Skipped 1 files when reviewing.
• Skipped posting 1 drafted comments based on config settings.

1. agents-api/pyproject.toml:53

• Draft comment:
  Ensure 'langchain-core' is compatible with existing dependencies to avoid potential conflicts.
• Reason this comment was not posted:
  Confidence changes required: 50%
  The addition of the 'langchain-core' dependency should be verified for compatibility with existing dependencies.

Workflow ID: wflow_Ihv02pfK4yNdugGf

---

You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.

[ellipsis-dev]

👍 Looks good to me! Incremental review on 4d0ee63 in 14 seconds

More details

• Looked at 13 lines of code in 1 files
• Skipped 0 files when reviewing.
• Skipped posting 1 drafted comments based on config settings.

1. agents-api/agents_api/activities/task_steps/prompt_step.py:122

• Draft comment:
  Use a single get call with a default value for better readability and efficiency.

fmt_tool.get("name", fmt_tool.get("function", {}).get("name")) : orig_tool

• Reason this comment was not posted:
  Confidence changes required: 50%
  The code uses fmt_tool.get("name") or fmt_tool.get("function", {}).get("name") to retrieve the tool name. This is a good approach to handle cases where the tool name might be nested under a "function" key. However, it would be more efficient to use a single call to get with a default value, which is a more Pythonic way to handle this.

Workflow ID: wflow_rV5J2ZGkecS8muJT

---

You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.

[ellipsis-dev]

👍 Looks good to me! Incremental review on 2168fe5 in 13 seconds

More details

• Looked at 28 lines of code in 1 files
• Skipped 0 files when reviewing.
• Skipped posting 1 drafted comments based on config settings.

1. agents-api/agents_api/common/protocol/tasks.py:238

• Draft comment:
  Consider using the @beartype decorator for all functions in this file for consistency, such as spec_to_task_data and spec_to_task.
• Reason this comment was not posted:
  Confidence changes required: 50%
  The use of beartype decorator is inconsistent across the file. It is used for task_to_spec but not for other functions like spec_to_task_data and spec_to_task. Consistency in using type-checking decorators is important for maintainability and readability.

Workflow ID: wflow_HeU1VhuwYYN5DDlF

---

You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.

[ellipsis-dev]

❌ Changes requested. Incremental review on 0af2684 in 19 seconds

More details

• Looked at 19 lines of code in 1 files
• Skipped 0 files when reviewing.
• Skipped posting 0 drafted comments based on config settings.

Workflow ID: wflow_L8BHgXEDsdv5kEO1

---

Want Ellipsis to fix these issues? Tag @ellipsis-dev in a comment. You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.