LLM inference in C/C++
C++ ggml
/**
* OpenAI-compatible tool call type.
*/
export enum ToolCallType {
FUNCTION = 'function'
}
/**
* Types of sections in agentic content display.
*/
export enum AgenticSectionType {
REASONING = 'reasoning',
REASONING_PENDING = 'reasoning_pending',
TEXT = 'text',
TOOL_CALL = 'tool_call',
TOOL_CALL_PENDING = 'tool_call_pending',
TOOL_CALL_STREAMING = 'tool_call_streaming'
}
/**
* How a Continue click on an assistant message resumes generation.
*/
export enum ContinueIntentKind {
APPEND_TEXT = 'append_text',
NEXT_TURN = 'next_turn',
RERUN_TURN = 'rerun_turn'
}
/**
* Renderer tier for a tool-result blob shown in the default tool-call block.
*/
export enum ToolResultKind {
JSON = 'json',
MARKDOWN = 'markdown',
TEXT = 'text'
}
/**
* Line classification for the unified-diff renderer of `edit_file` results.
*/
export enum DiffLineKind {
ADD = 'add',
CONTEXT = 'context',
REMOVE = 'remove'
}