microsoft.opentelemetry.a365.core.models.messages module

OTEL gen-ai semantic convention message types.

Defines the structured message format for input/output message tracing, following the OpenTelemetry gen-ai semantic conventions:

• https://opentelemetry.io/docs/specs/semconv/gen-ai/gen-ai-input-messages.json

• https://opentelemetry.io/docs/specs/semconv/gen-ai/gen-ai-output-messages.json

class microsoft.opentelemetry.a365.core.models.messages.MessageRole(value)

Bases: Enum

Role of a message participant per OTEL gen-ai semantic conventions.

SYSTEM = 'system'

System / developer instructions that set up the conversation.

USER = 'user'

Input authored by the end user.

ASSISTANT = 'assistant'

Output produced by the model.

TOOL = 'tool'

Output of a tool returned back to the model.

class microsoft.opentelemetry.a365.core.models.messages.FinishReason(value)

Bases: Enum

Reason a model stopped generating per OTEL gen-ai semantic conventions.

STOP = 'stop'

The model emitted a natural stop point or a provided stop sequence.

LENGTH = 'length'

Generation stopped because the maximum token limit was reached.

CONTENT_FILTER = 'content_filter'

Generation stopped because content was filtered by a safety system.

TOOL_CALL = 'tool_call'

The model stopped to request a tool call.

ERROR = 'error'

Generation stopped because an error occurred.

class microsoft.opentelemetry.a365.core.models.messages.Modality(value)

Bases: Enum

Media modality for blob, file, and URI parts.

IMAGE = 'image'

Image content.

VIDEO = 'video'

Video content.

AUDIO = 'audio'

Audio content.

class microsoft.opentelemetry.a365.core.models.messages.TextPart(content)

Bases: object

Plain text content.

content: str

type: str = 'text'

class microsoft.opentelemetry.a365.core.models.messages.ToolCallRequestPart(name, id=None, arguments=None)

Bases: object

A tool call requested by the model.

name: str

id: str | None = None

arguments: dict[str, object] | list[object] | str | None = None

type: str = 'tool_call'

class microsoft.opentelemetry.a365.core.models.messages.ToolCallResponsePart(id=None, response=None)

Bases: object

Result of a tool call.

id: str | None = None

response: object | None = None

type: str = 'tool_call_response'

class microsoft.opentelemetry.a365.core.models.messages.ReasoningPart(content)

Bases: object

Model reasoning / chain-of-thought content.

content: str

type: str = 'reasoning'

class microsoft.opentelemetry.a365.core.models.messages.BlobPart(modality, content, mime_type=None)

Bases: object

Inline binary data (base64-encoded).

modality: Modality | str

content: str

mime_type: str | None = None

type: str = 'blob'

class microsoft.opentelemetry.a365.core.models.messages.FilePart(modality, file_id, mime_type=None)

Bases: object

Reference to a pre-uploaded file.

modality: Modality | str

file_id: str

mime_type: str | None = None

type: str = 'file'

class microsoft.opentelemetry.a365.core.models.messages.UriPart(modality, uri, mime_type=None)

Bases: object

External URI reference.

modality: Modality | str

uri: str

mime_type: str | None = None

type: str = 'uri'

class microsoft.opentelemetry.a365.core.models.messages.ServerToolCallPart(name, server_tool_call, id=None)

Bases: object

Server-side tool invocation.

name: str

server_tool_call: dict[str, object]

id: str | None = None

type: str = 'server_tool_call'

class microsoft.opentelemetry.a365.core.models.messages.ServerToolCallResponsePart(server_tool_call_response, id=None)

Bases: object

Server-side tool response.

server_tool_call_response: dict[str, object]

id: str | None = None

type: str = 'server_tool_call_response'

class microsoft.opentelemetry.a365.core.models.messages.GenericPart(type, data=<factory>)

Bases: object

Extensible part for custom / future types.

type: str

data: dict[str, object]

microsoft.opentelemetry.a365.core.models.messages.MessagePart

Union of all message part types per OTEL gen-ai semantic conventions.

alias of TextPart | ToolCallRequestPart | ToolCallResponsePart | ReasoningPart | BlobPart | FilePart | UriPart | ServerToolCallPart | ServerToolCallResponsePart | GenericPart

class microsoft.opentelemetry.a365.core.models.messages.ChatMessage(role, parts=<factory>, name=None)

Bases: object

An input message sent to a model (OTEL gen-ai semantic conventions).

role: MessageRole

parts: list[TextPart | ToolCallRequestPart | ToolCallResponsePart | ReasoningPart | BlobPart | FilePart | UriPart | ServerToolCallPart | ServerToolCallResponsePart | GenericPart]

name: str | None = None

class microsoft.opentelemetry.a365.core.models.messages.OutputMessage(role, parts=<factory>, name=None, finish_reason='stop')

Bases: ChatMessage

An output message produced by a model (OTEL gen-ai semantic conventions).

finish_reason defaults to "stop" per OTel spec when not provided.

finish_reason: str | None = 'stop'

class microsoft.opentelemetry.a365.core.models.messages.InputMessages(messages=<factory>)

Bases: object

Container for input messages. Serializes as a plain JSON array per OTel spec.

messages: list[ChatMessage]

class microsoft.opentelemetry.a365.core.models.messages.OutputMessages(messages=<factory>)

Bases: object

Container for output messages. Serializes as a plain JSON array per OTel spec.

messages: list[OutputMessage]

microsoft.opentelemetry.a365.core.models.messages.InputMessagesParam

Accepted input for record_input_messages.

Supports a single string, a list of strings (backward compat), or the structured container.

alias of str | list[str] | InputMessages

microsoft.opentelemetry.a365.core.models.messages.OutputMessagesParam

Accepted input for record_output_messages.

Supports a single string, a list of strings (backward compat), or the structured container.

alias of str | list[str] | OutputMessages