tq42 package

Subpackages

• tq42.exceptions package
  • Submodules
  • tq42.exceptions.authentication_error module
    • AuthenticationError
  • tq42.exceptions.experiment_run_cancel_error module
    • ExperimentRunCancelError
  • tq42.exceptions.invalid_argument_error module
    • InvalidArgumentError
  • tq42.exceptions.no_default_error module
    • NoDefaultError
  • tq42.exceptions.permission_denied_error module
    • PermissionDeniedError
  • tq42.exceptions.retries_exceeded_error module
    • ExceedRetriesError
  • tq42.exceptions.tq42_api_error module
    • TQ42APIError
  • tq42.exceptions.unauthenticated_error module
    • UnauthenticatedError
  • Module contents
• tq42.internal package
  • Submodules
  • tq42.internal.functionality module
    • Functionality
      • Functionality.check()
  • Module contents

Submodules

tq42.channel module

class tq42.channel.Channel(client: TQ42Client, id: str)

Bases: object

Reference an existing channel with its id. A channel can be created with create().

Parameters:

• client – a client instance

• id – existing channel id

async connect(callback: Callable[[Ask], Awaitable[Tell]], finish_callback: Callable, max_duration_in_sec: int | None = None, message_timeout_in_sec: int | None = None) → None

Connects to the stream and handles every message with the provided callback to create an answer. ASK gets into the callback and then we expect a TELL answer

Parameters:

• callback – Async callback that handles an ASK message and returns a TELL message

• finish_callback – Callback that is called when channel is completed

• max_duration_in_sec (int) – Timeout for whole connection in seconds. None -> no timeout for overall flow

• message_timeout_in_sec (int) – Timeout between messages in seconds. None -> no timeout between messages

static create(client: TQ42Client) → Channel

Create a new channel

Parameters:

client – a client instance

Returns:

a new channel

id: str

ID of the channel

tq42.client module

class tq42.client.TQ42Client

Bases: object

Create a new instance of the TQ42Client to pass to any resource

Example:

>>> from tq42.experiment import list_all
...
... with TQ42Client() as client:
...     print(list_all(client=client, project_id="some-project-id"))

login()

Trigger authentication flow. This opens a new browser window to authenticate the sdk.

If the environment variables TQ42_AUTH_CLIENT_ID and TQ42_AUTH_CLIENT_SECRET are set the flow is performed without user interaction.

tq42.dataset module

class tq42.dataset.Dataset(client: TQ42Client, id: str, data: StorageProto | None = None)

Bases: object

Reference an existing dataset.

Parameters:

• client – a client instance

• id – the id of the existing dataset

• data – only used internally

static create(client: ~tq42.client.TQ42Client, project_id: str, name: str, description: str, sensitivity: <google.protobuf.internal.enum_type_wrapper.EnumTypeWrapper object at 0x7168c8f7a710>, file: str = None, url: str = None) → Dataset

Create a dataset for a project.

Params client:

a client instance

Parameters:

• project_id – the id of the project where the dataset should be created in

• name – name for the dataset

• description – description for the dataset

• sensitivity – sensitivity of the dataset (e.g. DatasetSensitivityProto.SENSITIVE for a sensitive dataset)

• file – path to local file that should be uploaded to the dataset

• url – url to remote file that should be uploaded to the dataset

Returns:

the created dataset

Only one of url or file can be specified.

data: StorageProto

Object containing all attributes of the dataset

delete()

Delete this dataset.

export(directory_path: str = '.') → List[str]

Export all files within a dataset to a local path

Parameters:

directory_path – local path where all files should be exported to (must exist and be a directory)

Returns:

a list of exported file paths

id: str

ID of the dataset

tq42.dataset.list_all(client: TQ42Client, project_id: str) → List[Dataset]

List all datasets in a project.

Parameters:

• client – a client instance

• project_id – the id of a project

Returns:

a list of datasets

tq42.experiment module

class tq42.experiment.Experiment(client: TQ42Client, id: str, data: ExperimentProto | None = None)

Bases: object

Reference an existing experiment.

Parameters:

• client – a client instance

• id – the id of the existing experiment

• data – only used internally

data: ExperimentProto

Object containing all attributes of the experiment

id: str

ID of the experiment

set_friendly_name(friendly_name: str) → Experiment

Set the friendly name of the experiment

Parameters:

friendly_name – new friendly name for the experiment

Returns:

the updated experiment

update(name: str) → Experiment

Update the name of the experiment

Parameters:

name – new name for the experiment

Returns:

the updated experiment

tq42.experiment.list_all(client: TQ42Client, project_id: str | None = None) → List[Experiment]

List all the experiments you have permission to view within a specific project. If no project_id is specified the currently set project id will be used for this.

Parameters:

• client – a client instance

• project_id – the id of the project to list experiments for (defaults to the currently set project)

Returns:

a list of all experiments

tq42.experiment_run module

class tq42.experiment_run.ExperimentRun(client: TQ42Client, id: str, data: ExperimentRunProto | None = None)

Bases: object

Reference an existing experiment run.

Parameters:

• client – a client instance

• id – the id of the existing experiment run

• data – only used internally

cancel() → ExperimentRun

Cancel a run that is QUEUED, PENDING, or RUNNING.

Returns:

the cancelled experiment run

Raises:

ExperimentRunCancelError if the experiment run is not queued, pending or running

check() → ExperimentRun

Update the state of the experiment run

Returns:

the updated experiment run

property completed: bool

Check if the experiment run is completed

Returns:

true if the experiment run is in the state COMPLETED

static create(client: ~tq42.client.TQ42Client, algorithm: str, version: str, experiment_id: str, compute: <google.protobuf.internal.enum_type_wrapper.EnumTypeWrapper object at 0x7168c8ea6e90>, parameters: ~typing.Mapping[str, ~typing.Any]) → ExperimentRun

Start a new experiment run in an experiment

Parameters:

• client – a client instance

• algorithm – name of the algorithm (e.g. ‘TOY’)

• version – version of the algorithm in the format x.y.z

• experiment_id – id of the experiment in which the run should be started

• compute – the hardware specification on which the run should be started (e.g. HardwareProto.SMALL)

• parameters – dict with parameters for the algorithm

Returns:

the created experiment run

data: ExperimentRunProto

Object containing all attributes of the experiment run

id: str

ID of the experiment run

property outputs: dict[str, Any] | None

Get the outputs of the experiment run if the run is completed.

Returns:

a dict with the outputs of the experiment run. If the run is not completed yet, returns None.

poll(tries=1000, initial_delay=1.0, delay=1.0, backoff=1.0) → ExperimentRun

Monitor an experiment run until it completes, then automatically display the results (if there are no errors).

Parameters:

• tries – how many retries until the poll loop is cancelled (default: 1000)

• initial_delay – initial delay before starting poll loop (default: 1 second)

• delay – initial delay between retries (default: 1 second)

• backoff – backoff factor between retries (default: 1)

Returns:

the finished experiment run

Raises:

ExceedRetriesError if tries are exceeded

property result: dict[str, Any] | None

Get the result of the experiment run if the run is completed.

If the result contains a results_string or if the result is a string, it will be parsed and returned.

Returns:

a dict with the result of the experiment run. If the run is not completed yet, returns None.

tq42.experiment_run.list_all(client: TQ42Client, experiment_id: str) → List[ExperimentRun]

List all the runs within an experiment you have permission to view.

Parameters:

• client – a client instance

• experiment_id – id of the experiment

Returns:

a list of experiment runs

tq42.model module

class tq42.model.Model(client: TQ42Client, id: str, data: StorageProto | None = None)

Bases: object

Reference an existing model.

Parameters:

• client – a client instance

• id – the id of the existing model

• data – only used internally

data: StorageProto

Object containing all attributes of the model

id: str

ID of the model

tq42.model.list_all(client: TQ42Client, project_id: str) → List[Model]

List all models for a project.

Parameters:

• client – a client instance

• project_id – the id of a project

tq42.organization module

class tq42.organization.Organization(client: TQ42Client, id: str, data: OrganizationProto | None = None)

Bases: object

Reference an existing organization.

Parameters:

• client – a client instance

• id – the id of the existing organization

• data – only used internally

data: OrganizationProto

Object containing all attributes of the organization

static get_default_org(client: TQ42Client) → Organization | None

Gets the default organization for this user based on the default_org field

Returns:

the default organization if one is set as a default

id: str

ID of the organization

set() → Organization

Sets the current organization as the default organization.

Returns:

organization instance

tq42.organization.list_all(client: TQ42Client) → List[Organization]

List all the organizations you have permission to view.

Parameters:

client – a client instance

Returns:

a list of all organizations

tq42.project module

class tq42.project.Project(client: TQ42Client, id: str, data: ProjectProto | None = None)

Bases: object

Reference an existing project.

Parameters:

• client – a client instance

• id – the id of the existing project

• data – only used internally

data: ProjectProto

Object containing all attributes of the project

static get_default(client: TQ42Client, organization_id: str) → Project | None

Gets the default project in an organization

Parameters:

• client – a client instance

• organization_id – the id of the organization

Returns:

the default project if there is at least one project in the organization

id: str

ID of the dataset

set() → Project

Set this project as the default project

Returns:

the project

set_friendly_name(friendly_name: str) → Project

Set a friendly name for a project.

Parameters:

friendly_name – new friendly name for the project

Returns:

the updated project

static show(client: TQ42Client) → Project

Returns the current default project.

Parameters:

client – a client instance

Raises:

NoDefaultError if no default project is set

Returns:

the current default project

update(name: str) → Project

Update the name of the project

Parameters:

name – new name for the project

Returns:

the updated project

tq42.project.list_all(client: TQ42Client, organization_id: str | None = None) → List[Project]

List all the projects you have permission to view within the organization.

Parameters:

• client – a client instance

• organization_id – optionally an id of an organization, defaults to the default organization

Returns:

a list of projects

Module contents