span_data

Classes

AdalFlowGeneratorSpanData([generator_id, ...])	Represents Generator execution spans in AdalFlow.
AdalFlowResponseSpanData([answer, ...])	Represents AdalFlow Response spans that track the final result.
AdalFlowRunnerSpanData([runner_id, ...])	Represents the top-level Runner execution span in AdalFlow.
AdalFlowStepSpanData([step_number, ...])	Represents a Step Span in AdalFlow workflow execution.
AdalFlowToolSpanData([tool_name, ...])	Represents Tool execution spans in AdalFlow.
AgentSpanData(name[, handoffs, tools, ...])	Represents an Agent Span in the trace.
CustomSpanData(name, data)	Represents a Custom Span in the trace.
FunctionSpanData(name, input, output[, mcp_data])	Represents a Function Span in the trace.
GenerationSpanData([input, output, model, ...])	Represents a Generation Span in the trace.
GuardrailSpanData(name[, triggered])	Represents a Guardrail Span in the trace.
HandoffSpanData(from_agent, to_agent)	Represents a Handoff Span in the trace.
MCPListToolsSpanData([server, result])	Represents an MCP List Tools Span in the trace.
ResponseSpanData([response, input])	Represents a Response Span in the trace.
SpanData()	Represents span data in the trace.
SpeechGroupSpanData([input])	Represents a Speech Group Span in the trace.
SpeechSpanData([input, output, ...])	Represents a Speech Span in the trace.
TranscriptionSpanData([input, input_format, ...])	Represents a Transcription Span in the trace.

class SpanData

Bases: ABC

Represents span data in the trace.

abstract export() → Dict[str, Any]

Export the span data as a dictionary.

abstract property type: str

Return the type of the span.

class AgentSpanData(name: str, handoffs: List[str] | None = None, tools: List[str] | None = None, output_type: str | None = None)

Bases: SpanData

Represents an Agent Span in the trace. Includes name, handoffs, tools, and output type.

name

handoffs: List[str] | None

tools: List[str] | None

output_type: str | None

property type: str

Return the type of the span.

export() → Dict[str, Any]

Export the span data as a dictionary.

class FunctionSpanData(name: str, input: str | None, output: Any | None, mcp_data: Dict[str, Any] | None = None)

Bases: SpanData

Represents a Function Span in the trace. Includes input, output and MCP data (if applicable).

name

input

output

mcp_data

property type: str

Return the type of the span.

export() → Dict[str, Any]

Export the span data as a dictionary.

class GenerationSpanData(input: Sequence[Mapping[str, Any]] | None = None, output: Sequence[Mapping[str, Any]] | None = None, model: str | None = None, model_config: Mapping[str, Any] | None = None, usage: Dict[str, Any] | None = None)

Bases: SpanData

Represents a Generation Span in the trace. Includes input, output, model, model configuration, and usage.

input

output

model

model_config

usage

property type: str

Return the type of the span.

export() → Dict[str, Any]

Export the span data as a dictionary.

class ResponseSpanData(response: Response | None = None, input: str | List[ResponseInputItemParam] | None = None)

Bases: SpanData

Represents a Response Span in the trace. Includes response and input.

response

input

property type: str

Return the type of the span.

export() → Dict[str, Any]

Export the span data as a dictionary.

class HandoffSpanData(from_agent: str | None, to_agent: str | None)

Bases: SpanData

Represents a Handoff Span in the trace. Includes source and destination agents.

from_agent

to_agent

property type: str

Return the type of the span.

export() → Dict[str, Any]

Export the span data as a dictionary.

class CustomSpanData(name: str, data: Dict[str, Any])

Bases: SpanData

Represents a Custom Span in the trace. Includes name and data property bag.

name

data

property type: str

Return the type of the span.

export() → Dict[str, Any]

Export the span data as a dictionary.

class GuardrailSpanData(name: str, triggered: bool = False)

Bases: SpanData

Represents a Guardrail Span in the trace. Includes name and triggered status.

name

triggered

property type: str

Return the type of the span.

export() → Dict[str, Any]

Export the span data as a dictionary.

class TranscriptionSpanData(input: str | None = None, input_format: str | None = 'pcm', output: str | None = None, model: str | None = None, model_config: Mapping[str, Any] | None = None)

Bases: SpanData

Represents a Transcription Span in the trace. Includes input, output, model, and model configuration.

input

input_format

output

model

model_config

property type: str

Return the type of the span.

export() → Dict[str, Any]

Export the span data as a dictionary.

class SpeechSpanData(input: str | None = None, output: str | None = None, output_format: str | None = 'pcm', model: str | None = None, model_config: Mapping[str, Any] | None = None, first_content_at: str | None = None)

Bases: SpanData

Represents a Speech Span in the trace. Includes input, output, model, model configuration, and first content timestamp.

input

output

output_format

model

model_config

first_content_at

property type: str

Return the type of the span.

export() → Dict[str, Any]

Export the span data as a dictionary.

class SpeechGroupSpanData(input: str | None = None)

Bases: SpanData

Represents a Speech Group Span in the trace.

input

property type: str

Return the type of the span.

export() → Dict[str, Any]

Export the span data as a dictionary.

class MCPListToolsSpanData(server: str | None = None, result: List[str] | None = None)

Bases: SpanData

Represents an MCP List Tools Span in the trace. Includes server and result.

server

result

property type: str

Return the type of the span.

export() → Dict[str, Any]

Export the span data as a dictionary.

class AdalFlowRunnerSpanData(runner_id: str | None = None, max_steps: int | None = None, steps_executed: int | None = None, final_answer: str | None = None, workflow_status: str | None = None, execution_time: float | None = None)

Bases: CustomSpanData

Represents the top-level Runner execution span in AdalFlow. Tracks complete lifecycle of multi-step agent workflows. Maps to OpenAI’s AgentSpanData format.

The caller should update span attributes by using update_attributes() method.

data

runner_id

max_steps

steps_executed

final_answer

workflow_status

execution_time

export() → Dict[str, Any]

Export the span data as a dictionary.

update_attributes(attributes: Dict[str, Any]) → None

Update span attributes directly for MLflow compatibility. Note: MLflow’s OpenAI Trace processor reads attributes directly from the data attribute instead of using export(). This is a known limitation of the current MLflow implementation.

class AdalFlowGeneratorSpanData(generator_id: str | None = None, model_kwargs: Dict[str, Any] | None = None, generator_state_logger: Any | None = None, prompt_kwargs: Dict[str, Any] | None = None, prompt_template_with_keywords: str | None = None, raw_response: str | None = None, api_response: Any | None = None, generation_time_in_seconds: float | None = None, token_usage: Dict[str, int] | None = None, final_response: Any | None = None, api_kwargs: Dict[str, Any] | None = None)

Bases: CustomSpanData

Represents Generator execution spans in AdalFlow. Tracks LLM generation, processing, and output formatting using generator_state_logger. Maps to OpenAI’s CustomSpanData format.

The caller should update span attributes by using update_attributes() method.

data

generator_id

model_kwargs

generator_state_logger

prompt_kwargs

prompt_template_with_keywords

raw_response

api_response

generation_time_in_seconds

token_usage

final_response

api_kwargs

update_attributes(attributes: Dict[str, Any]) → None

Update span attributes directly for MLflow compatibility. Note: MLflow’s OpenAI Trace processor reads attributes directly from the data attribute instead of using export(). This is a known limitation of the current MLflow implementation.

export() → Dict[str, Any]

Export the span data as a dictionary.

class AdalFlowResponseSpanData(answer: Any | None = None, result_type: str | None = None, execution_metadata: Dict[str, Any] | None = None, response: Any | None = None)

Bases: CustomSpanData

Represents AdalFlow Response spans that track the final result. Based on CustomSpanData but specialized for AdalFlow final outputs.

The caller should update span attributes by using update_attributes() method.

data

answer

result_type

execution_metadata

response

update_attributes(attributes: Dict[str, Any]) → None

Update span attributes directly for MLflow compatibility. Note: MLflow’s OpenAI Trace processor reads attributes directly from the data attribute instead of using export(). This is a known limitation of the current MLflow implementation.

export() → Dict[str, Any]

Export the span data as a dictionary.

class AdalFlowStepSpanData(step_number: int | None = None, action_type: str | None = None, observation: Any | None = None, is_final: bool = False, tool_name: str | None = None, tool_output: Any | None = None, execution_time: float | None = None, error_info: Dict[str, Any] | None = None)

Bases: CustomSpanData

Represents a Step Span in AdalFlow workflow execution. Tracks individual steps within a multi-step agent workflow.

The caller should update span attributes by using update_attributes() method.

data

step_number

action_type

observation

is_final

tool_name

tool_output

execution_time

error_info

update_attributes(attributes: Dict[str, Any]) → None

Update span attributes directly for MLflow compatibility. Note: MLflow’s OpenAI Trace processor reads attributes directly from the data attribute instead of using export(). This is a known limitation of the current MLflow implementation.

export() → Dict[str, Any]

Export the span data as a dictionary.

class AdalFlowToolSpanData(tool_name: str | None = None, function_name: str | None = None, input_params: Dict[str, Any] | None = None, function_args: Dict[str, Any] | None = None, function_kwargs: Dict[str, Any] | None = None, output_result: Any | None = None, execution_time: float | None = None, error_info: Dict[str, Any] | None = None)

Bases: CustomSpanData

Represents Tool execution spans in AdalFlow. Tracks function calls, tool usage, and external integrations. Inherits OpenAI’s CustomSpanData format.

The caller should update span attributes by using update_attributes() method.

data

tool_name

function_name

input_params

function_args

function_kwargs

output_result

execution_time

error_info

update_attributes(attributes: Dict[str, Any]) → None

Update span attributes directly for MLflow compatibility. Note: MLflow’s OpenAI Trace processor reads attributes directly from the data attribute instead of using export(). This is a known limitation of the current MLflow implementation.

export() → Dict[str, Any]

Export the span data as a dictionary.