alab_management.builders.experimentbuilder module

alab_management.builders.experimentbuilder module#

class ExperimentBuilder(name, tags=None, **metadata)[source]#

Bases: object

It takes a list of samples and a list of tasks, and returns a dictionary that can be used to generate an input file for the experiment to run.

Parameters:

name (str) – The name of the experiment.

add_sample(name, tags=None, **metadata)[source]#

Add a sample to the batch. Each sample already has multiple tasks binded to it. Each batch is a directed graph of tasks.

Parameters:

  • name (str) – The name of the sample. This must be unique within this ExperimentBuilder.

  • tags (List[str]) – A list of tags to attach to the sample.

  • **metadata – Any additional keyword arguments will be attached to this sample as metadata.

Return type:

SampleBuilder

Returns:

A SampleBuilder object. This can be used to add tasks to the sample.

add_task(task_id, task_name, task_kwargs, samples)[source]#

This function adds a task to the sample. You should use this function only for special cases which are not handled by the add_sample function.

Parameters:

  • task_id (str) – The object id of the task in mongodb

  • task_name (str) – The name of the task.

  • task_kwargs (Dict[str, Any]) – Any additional keyword arguments will be attached to this sample as metadata.

  • samples (List[SampleBuilder]) – A list of samples to which this task is binded to.

Return type:

None

Returns:

None

generate_input_file(filename, fmt='json')[source]#

Genreate an input file for the experiment command.

Parameters:

  • filename (str) – The name of the file to be generated.

  • fmt (Literal["json", "yaml"]) – The format of the file to be generated.

Return type:

None

Returns:

None.

plot(ax=None)[source]#

Plot the directed graph of tasks.

Parameters:

ax (matplotlib.axes.Axes) – The axes on which to plot the graph.

Return type:

None

Returns:

None.

submit(address='http://localhost:8895', **kwargs)[source]#

Submit the experiment to server.

Parameters:

  • address (str) – The address of the server. It is defaulted to http://localhost:8895, which is the default address of the alabos server.

  • **kwargs – Additional keyword arguments to be passed to the requests.post function.

Return type:

ObjectId

Returns:

The object id of the experiment.

to_dict()[source]#

Return a dictionary that can be used to generate an input file for the experiment to run.

Return type:

dict[str, Any]

Returns:

A dictionary that can be used to generate an input file for the experiment to run.

get_experiment_result(exp_id, address='http://localhost:8895', **kwargs)[source]#

Get the result of the experiment.

Parameters:

  • exp_id (ObjectId | str) – The object id of the experiment.

  • address (str) – The address of the server. It is defaulted to http://localhost:8895, which is the default address of the alabos server.

  • **kwargs – Additional keyword arguments to be passed to the requests.get function.

Returns:

See the dashboard code for the response format. alab_management.dashboard.routes.experiment.query_experiment_results()

get_experiment_status(exp_id, address='http://localhost:8895', **kwargs)[source]#

Get the status of the experiment.

Parameters:

  • exp_id (ObjectId) – The object id of the experiment.

  • address (str) – The address of the server. It is defaulted to http://localhost:8895, which is the default address of the alabos server.

  • **kwargs – Additional keyword arguments to be passed to the requests.get function.

Returns:

  • The status of the experiment.

  • .. seealso:: – See the dashboard code for the response format. alab_management.dashboard.routes.experiment.query_experiment()

Contents