wei.experiment_client

Contains the Experiment class that manages WEI flows and helps annotate the experiment run

Classes

ExperimentClient([server_host, server_port, ...])

Methods for the running and logging of a WEI Experiment including running WEI workflows

Methods for the running and logging of a WEI Experiment including running WEI workflows

__del__(): Logs the end of the experiment when cleaning up the experiment, if log_experiment_end_on_exit is True

__enter__(): Creates the experiment application context

__exit__(exc_type, exc_val, exc_tb): Logs the end of the experiment when exiting the experiment application context, if log_experiment_end_on_exit is True

Initializes an Experiment, and creates its log files

Parameters:

server_host (str) – address for WEI server
server_port (str) – port for WEI server
experiment (Optional[Union[ExperimentDesign, Experiment, str]]) – A new experiment design, or an existing experiment or experiment_id to continue
campaign (Optional[Union[CampaignDesign, Campaign, str]]) – A new campaign design, or an existing campaign or campaign_id to associate with the experiment
working_dir (Optional[Union[str, Path]]) – The directory to resolve relative paths from. Defaults to the current working directory.
log_experiment_end_on_exit (bool) – Whether to log the end of the experiment when cleaning up the experiment

await_runs(run_ids: List[str | ULID]) → Dict[str, WorkflowRun]: Waits for all provided runs to complete, then returns a dictionary of their results

check_in(): Checks in with the server to let it know the experiment is still running

create_datapoint(datapoint: DataPoint): Creates a new datapoint attached to the experiment

experiment_started = False: Lifecycle Methods

get_datapoint_info(datapoint_id: str) → DataPoint

Returns the metadata for the datapoint for the given id

Parameters:: datapoint_id (str) – The id of the datapoint to get
Returns:: response (DataPoint) – The metadata for the requested datapoint

get_datapoint_value(datapoint_id: str) → Dict[str, Any] | bytes

Returns the value of the datapoint for the given id

Parameters:: datapoint_id (str) – The id of the datapoint to get
Returns:: response (Dict[str, Any] | bytes) – Either a json object (for Value Datapoints) or bytes object (for File Datapoints) containing the value of the requested datapoint

get_experiment_datapoints() → Dict[str, DataPoint]: returns a dictionary of the datapoints for this experiment.

get_run(run_id: str | ULID) → WorkflowRun

Checks on a workflow run using the id given

Parameters:: run_id (str) – The id returned by the run_job function for this run
Returns:: response (Dict) – The JSON portion of the response from the server

get_run_log(run_id: str) → Dict[Any, Any]

Returns the log for this experiment as a string

Parameters:: None
Returns:: response (Dict) – The JSON portion of the response from the server with the experiment log

get_wf_result_file(run_id: str, filename: str, output_filepath: str) → Any: Returns a file from the WEI experiment result directory

get_workcell_state() → dict

Fetches the current state of the workcell from the server.

Returns:: dict – A dictionary containing the current state of the workcell
Raises:: requests.exceptions.RequestException – If there’s an error communicating with the server

list_wf_files(run_id: str) → Any: Returns a list of files from the WEI experiment run directory

log_comment(comment: str) → Event

Logs a comment in the experiment log

Parameters:: comment (str) – The comment to log
Returns:: Event – The event that was logged

log_decision(decision_name: str, decision_value: bool) → Event

Logs a decision in the experiment log

Parameters:

decision_name (str) – The name of the decision
decision_value (bool) – The value of the decision

Returns:

Event – The event that was logged

log_event(event: Event) → Event

Logs an event to the WEI event log

Parameters:: event (Event) – The event to log
Returns:: Event – The event that was logged

log_experiment_end() → Event

Logs the end of the experiment in the experiment log

Parameters:: None
Returns:: Event – The event that was logged

log_gladier_flow(flow_name: str, flow_id: Any) → Event: Logs a Gladier flow in the experiment log

log_globus_compute(function_name: str, args: List[Any] | None = None, kwargs: Dict[str, Any] | None = None, result: Any | None = None) → Event: Logs a Globus computation in the experiment log

log_local_compute(function_name: str, args: List[Any] | None = None, kwargs: Dict[str, Any] | None = None, result: Any | None = None) → Event

Logs a local computation in the experiment log

Parameters:

function_name (str) – The name of the function called
args (List[Any]) – The positional arguments passed to the function
kwargs (Dict[str, Any]) – The keyword arguments passed to the function

Returns:

Event – The event that was logged

log_loop_start(loop_name: str) → Event: Logs the start of a loop in the experiment log

query_run(run_id: str | ULID) → WorkflowRun

Checks on a workflow run using the id given

Parameters:: run_id (str) – The id returned by the run_job function for this run
Returns:: response (Dict) – The JSON portion of the response from the server

save_datapoint_value(datapoint_id: str, output_filepath: str) → None

Saves the datapoint for the given id to the specified file

Parameters:

datapoint_id (str) – The id of the datapoint to save
output_filepath (str) – The path to save the datapoint to

Returns:

None

Creates a new experiment, or continues an existing experiment

Parameters:

experiment (Union[ExperimentDesign, Experiment, str]) – A new experiment design, or an existing experiment or experiment_id to continue
campaign (Optional[Union[CampaignDesign, Campaign, str]]) – A new campaign design, or an existing campaign or campaign_id to associate with the experiment

Returns:

None

start_run(workflow: Workflow | str | Path, payload: Dict[str, Any] | None = None, parameters: Dict[str, Any] | None = {}, simulate: bool = False, blocking: bool = True, validate_only: bool = False, raise_on_failed: bool = True, raise_on_cancelled: bool = True) → WorkflowRun

Submits a workflow file to the server to be executed, and logs it in the overall event log.

Parameters:

workflow (str) – A workflow definition, or path to a workflow definition file
payload (Optional[Dict[str, Any]]) – Arguments to the workflow
simulate (bool = False) – Whether or not to run the workflow purely simulated
blocking (bool = True) – Whether to wait for the workflow to complete before returning
validate_only (bool = False) – Whether to only validate the workflow without actually running it
raise_on_failed (bool = True) – Whether to raise an exception if the workflow fails.
raise_on_cancelled (bool = True) – Whether to raise an exception if the workflow is cancelled.

Returns:

WorkflowRun – Information about the run that was started

validate_workflow(workflow: Workflow | str | Path, payload: Dict[str, Any] | None = None)

Submits a workflow file to the server to be validated

Parameters:

workflow (Union[Workflow, PathLike]) – A workflow definition, or path to a workflow definition file
payload (Optional[Dict[str, Any]]) – Arguments to the workflow

wait_for_workflow(workflow: WorkflowRun | str | ULID, polling_interval: float = 1.0) → WorkflowRun

Waits for a workflow to reach a terminal state (COMPLETED, FAILED, or CANCELLED).

Parameters:

workflow (Union[WorkflowRun, str, ULID]) – The workflow run object, run_id, or ULID to wait for.
polling_interval (float, optional) – The time in seconds to wait between status checks. Defaults to 1.0 second.

Returns:

WorkflowRun – The final state of the workflow run.

Raises:

ValueError – If the provided workflow is invalid.

wei.experiment_client

Lifecycle Methods