promptflow.storage.run_records module

class promptflow.storage.run_records.LineRunRecord(line_number: int, run_info: dict, start_time: datetime, end_time: datetime, name: str, description: str, status: str, tags: str)

Bases: object

A dataclass for storing the run record of a single line execution on the flow.

Parameters:

• line_number (int) – The line number in the record

• run_info (dict) – The information about the line run

• start_time (datetime) – The time the line started executing

• end_time (datetime) – The time the line finished executing

• name (str) – The name of the line run

• description (str) – The description of the line run

• status (str) – The status of the line execution

• tags (str) – The tags associated with the line run

description: str

end_time: datetime

static from_run_info(run_info: FlowRunInfo) → LineRunRecord

Create a LineRunRecord from a FlowRunInfo object.

Parameters:

run_info (FlowRunInfo) – The run info to create the LineRunRecord from

Returns:

The created LineRunRecord

Return type:

LineRunRecord

line_number: int

name: str

run_info: dict

serialize() → str

Serialize the LineRunRecord for storage in a blob.

Returns:

The serialized result

Return type:

str

start_time: datetime

status: str

tags: str

class promptflow.storage.run_records.NodeRunRecord(node_name: str, line_number: int, run_info: dict, start_time: datetime, end_time: datetime, status: str)

Bases: object

Dataclass for storing the run record of each node during single line execution on the flow

Parameters:

• node_name (str) – The name of the node

• line_number (int) – The line number in the source file

• run_info (dict) – The information about the run

• start_time (datetime) – The time the node started running

• end_time (datetime) – The time the node finished running

• status (str) – The status of the node run

end_time: datetime

static from_run_info(run_info: RunInfo) → NodeRunRecord

Create a NodeRunRecord from a RunInfo object.

Parameters:

run_info (RunInfo) – The run info to create the NodeRunRecord from

Returns:

The created NodeRunRecord

Return type:

NodeRunRecord

line_number: int

node_name: str

run_info: dict

serialize() → str

Serialize the NodeRunRecord for storage in blob.

Returns:

The serialized result

Return type:

str

start_time: datetime

status: str