Skip to content

Class ZCL_LLM_TOOL_ECHO

AI Generated documentation.

Overview

The zcl_llm_tool_echo class implements a simple echo tool that mirrors input data. It adheres to the zif_llm_tool interface, providing standardized access to tool metadata and execution results.

Key Public Methods:

  • constructor: Initializes the tool with metadata (tool_details).
  • execute: Stores input data and tool_call_id, returning them as part of the result.
  • get_tool_details: Returns metadata (e.g., name, purpose) configured during initialization.
  • get_result: Provides access to the stored response data and execution context.

Dependencies

  • Implements interface zif_llm_tool (mandatory methods and type definitions).
  • Relies on the structure zif_llm_tool=>tool_details for metadata management.

Details

Data Flow

sequenceDiagram  
    User->>+zcl_llm_tool_echo: execute(data, tool_call_id)  
    zcl_llm_tool_echo->>zcl_llm_tool_echo: Store data in response_data  
    zcl_llm_tool_echo->>zcl_llm_tool_echo: Update tool_call_id  
    zcl_llm_tool_echo-->>-User: Return result with mirrored data

Key Implementation Notes

  1. Generic Data Handling:
  2. Uses TYPE REF TO data for response_data to support arbitrary input structures.

  3. No data transformation occurs – input is stored and returned as-is.

  4. Execution Context:

  5. The tool_call_id links execution requests to results (critical for asynchronous operations or audit trails).

  6. Metadata Propagation:

  7. The tool_details (provided during construction) are reused in get_result and get_tool_details, ensuring consistency.

  8. Stateless Design:

  9. Each execute call overwrites previous response_data and tool_call_id, making the instance unsuitable for concurrent use.