Detection Result
src.llm.tool_detection.detection_result.DetectionState
Bases: str
, Enum
Enumeration of possible tool call detection states.
Represents the different states a tool call detection can be in during the processing of streaming chunks.
Attributes:
Name | Type | Description |
---|---|---|
NO_MATCH |
No tool call pattern was detected in the current content. |
|
PARTIAL_MATCH |
A potential tool call was detected but is incomplete. |
|
COMPLETE_MATCH |
A complete and valid tool call was detected. |
Source code in src/llm/tool_detection/detection_result.py
11 12 13 14 15 16 17 18 19 20 21 22 23 24 |
|
src.llm.tool_detection.detection_result.DetectionResult
Bases: BaseModel
Container for tool call detection results.
Encapsulates the result of processing an SSE chunk for tool calls, including the detection state, any found tool calls, accumulated content, and the original chunk.
Attributes:
Name | Type | Description |
---|---|---|
state |
DetectionState
|
The current state of tool call detection. Indicates whether a tool call was found and if it's complete. |
tool_calls |
Optional[List[ToolCall]]
|
List of tool calls found in the content. Only present when state is COMPLETE_MATCH. |
content |
Optional[str]
|
Accumulated text content from the chunk, excluding any tool call syntax. |
sse_chunk |
Optional[SSEChunk]
|
The original SSE chunk that was processed, preserved for reference or further processing. |
Source code in src/llm/tool_detection/detection_result.py
27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 |
|