# This file was auto-generated by Fern from our API Definition.
import typing
from ..core.client_wrapper import AsyncClientWrapper, SyncClientWrapper
from ..core.request_options import RequestOptions
from ..types.ml_backend import MlBackend
from .raw_client import AsyncRawMlClient, RawMlClient
from .types.create_ml_request_auth_method import CreateMlRequestAuthMethod
from .types.list_model_versions_ml_response import ListModelVersionsMlResponse
from .types.update_ml_request_auth_method import UpdateMlRequestAuthMethod
# this is used as the default value for optional parameters
OMIT = typing.cast(typing.Any, ...)
class MlClient:
def __init__(self, *, client_wrapper: SyncClientWrapper):
self._raw_client = RawMlClient(client_wrapper=client_wrapper)
@property
def with_raw_response(self) -> RawMlClient:
"""
Retrieves a raw implementation of this client that returns raw responses.
Returns
-------
RawMlClient
"""
return self._raw_client
def list(
self, *, project: typing.Optional[int] = None, request_options: typing.Optional[RequestOptions] = None
) -> typing.List[MlBackend]:
"""
List all configured ML backends for a specific project by ID.
Use the following cURL command:
```bash
curl http://localhost:8000/api/ml?project={project_id} -H 'Authorization: Token abc123'
Parameters
----------
project : typing.Optional[int]
Project ID
request_options : typing.Optional[RequestOptions]
Request-specific configuration.
Returns
-------
typing.List[MlBackend]
Examples
--------
from label_studio_sdk import LabelStudio
client = LabelStudio(
api_key="YOUR_API_KEY",
)
client.ml.list()
"""
_response = self._raw_client.list(project=project, request_options=request_options)
return _response.data
def create(
self,
*,
auth_method: typing.Optional[CreateMlRequestAuthMethod] = OMIT,
basic_auth_pass: typing.Optional[str] = OMIT,
basic_auth_user: typing.Optional[str] = OMIT,
description: typing.Optional[str] = OMIT,
extra_params: typing.Optional[typing.Dict[str, typing.Any]] = OMIT,
is_interactive: typing.Optional[bool] = OMIT,
project: typing.Optional[int] = OMIT,
timeout: typing.Optional[int] = OMIT,
title: typing.Optional[str] = OMIT,
url: typing.Optional[str] = OMIT,
request_options: typing.Optional[RequestOptions] = None,
) -> MlBackend:
"""
Add an ML backend to a project using the Label Studio UI or by sending a POST request using the following cURL
command:
```bash
curl -X POST -H 'Content-type: application/json' http://localhost:8000/api/ml -H 'Authorization: Token abc123'\\
--data '{"url": "http://localhost:9090", "project": {project_id}}'
Parameters
----------
auth_method : typing.Optional[CreateMlRequestAuthMethod]
Auth method
basic_auth_pass : typing.Optional[str]
Basic auth password
basic_auth_user : typing.Optional[str]
Basic auth user
description : typing.Optional[str]
Description
extra_params : typing.Optional[typing.Dict[str, typing.Any]]
Extra parameters
is_interactive : typing.Optional[bool]
Is interactive
project : typing.Optional[int]
Project ID
timeout : typing.Optional[int]
Response model timeout
title : typing.Optional[str]
Title
url : typing.Optional[str]
ML backend URL
request_options : typing.Optional[RequestOptions]
Request-specific configuration.
Returns
-------
MlBackend
Examples
--------
from label_studio_sdk import LabelStudio
client = LabelStudio(
api_key="YOUR_API_KEY",
)
client.ml.create()
"""
_response = self._raw_client.create(
auth_method=auth_method,
basic_auth_pass=basic_auth_pass,
basic_auth_user=basic_auth_user,
description=description,
extra_params=extra_params,
is_interactive=is_interactive,
project=project,
timeout=timeout,
title=title,
url=url,
request_options=request_options,
)
return _response.data
def get(self, id: int, *, request_options: typing.Optional[RequestOptions] = None) -> MlBackend:
"""
Get details about a specific ML backend connection by ID. For example, make a GET request using the
following cURL command:
```bash
curl http://localhost:8000/api/ml/{ml_backend_ID} -H 'Authorization: Token abc123'
Parameters
----------
id : int
request_options : typing.Optional[RequestOptions]
Request-specific configuration.
Returns
-------
MlBackend
Examples
--------
from label_studio_sdk import LabelStudio
client = LabelStudio(
api_key="YOUR_API_KEY",
)
client.ml.get(
id=1,
)
"""
_response = self._raw_client.get(id, request_options=request_options)
return _response.data
def delete(self, id: int, *, request_options: typing.Optional[RequestOptions] = None) -> None:
"""
Remove an existing ML backend connection by ID. For example, use the
following cURL command:
```bash
curl -X DELETE http://localhost:8000/api/ml/{ml_backend_ID} -H 'Authorization: Token abc123'
Parameters
----------
id : int
request_options : typing.Optional[RequestOptions]
Request-specific configuration.
Returns
-------
None
Examples
--------
from label_studio_sdk import LabelStudio
client = LabelStudio(
api_key="YOUR_API_KEY",
)
client.ml.delete(
id=1,
)
"""
_response = self._raw_client.delete(id, request_options=request_options)
return _response.data
def update(
self,
id: int,
*,
auth_method: typing.Optional[UpdateMlRequestAuthMethod] = OMIT,
basic_auth_pass: typing.Optional[str] = OMIT,
basic_auth_user: typing.Optional[str] = OMIT,
description: typing.Optional[str] = OMIT,
extra_params: typing.Optional[typing.Dict[str, typing.Any]] = OMIT,
is_interactive: typing.Optional[bool] = OMIT,
project: typing.Optional[int] = OMIT,
timeout: typing.Optional[int] = OMIT,
title: typing.Optional[str] = OMIT,
url: typing.Optional[str] = OMIT,
request_options: typing.Optional[RequestOptions] = None,
) -> MlBackend:
"""
Update ML backend parameters using the Label Studio UI or by sending a PATCH request using the following cURL command:
```bash
curl -X PATCH -H 'Content-type: application/json' http://localhost:8000/api/ml/{ml_backend_ID} -H 'Authorization: Token abc123'\\
--data '{"url": "http://localhost:9091"}'
Parameters
----------
id : int
auth_method : typing.Optional[UpdateMlRequestAuthMethod]
Auth method
basic_auth_pass : typing.Optional[str]
Basic auth password
basic_auth_user : typing.Optional[str]
Basic auth user
description : typing.Optional[str]
Description
extra_params : typing.Optional[typing.Dict[str, typing.Any]]
Extra parameters
is_interactive : typing.Optional[bool]
Is interactive
project : typing.Optional[int]
Project ID
timeout : typing.Optional[int]
Response model timeout
title : typing.Optional[str]
Title
url : typing.Optional[str]
ML backend URL
request_options : typing.Optional[RequestOptions]
Request-specific configuration.
Returns
-------
MlBackend
Examples
--------
from label_studio_sdk import LabelStudio
client = LabelStudio(
api_key="YOUR_API_KEY",
)
client.ml.update(
id=1,
)
"""
_response = self._raw_client.update(
id,
auth_method=auth_method,
basic_auth_pass=basic_auth_pass,
basic_auth_user=basic_auth_user,
description=description,
extra_params=extra_params,
is_interactive=is_interactive,
project=project,
timeout=timeout,
title=title,
url=url,
request_options=request_options,
)
return _response.data
def predict_interactive(
self,
id: int,
*,
task: int,
context: typing.Optional[typing.Any] = OMIT,
request_options: typing.Optional[RequestOptions] = None,
) -> None:
"""
Send a request to the machine learning backend set up to be used for interactive preannotations to retrieve a
predicted region based on annotator input.
See [set up machine learning](https://labelstud.io/guide/ml.html#Get-interactive-preannotations) for more.
Parameters
----------
id : int
A unique integer value identifying this ML backend.
task : int
ID of task to annotate
context : typing.Optional[typing.Any]
Context for ML model
request_options : typing.Optional[RequestOptions]
Request-specific configuration.
Returns
-------
None
Examples
--------
from label_studio_sdk import LabelStudio
client = LabelStudio(
api_key="YOUR_API_KEY",
)
client.ml.predict_interactive(
id=1,
task=1,
)
"""
_response = self._raw_client.predict_interactive(
id, task=task, context=context, request_options=request_options
)
return _response.data
def predict_all_tasks(
self,
id: int,
*,
batch_size: typing.Optional[int] = None,
request_options: typing.Optional[RequestOptions] = None,
) -> None:
"""
This endpoint is not available in Label Studio Community Edition. [Learn more about Label Studio Enterprise](https://humansignal.com/goenterprise)
Create predictions for all tasks using a specific ML backend so that you can set up an active learning strategy based on the confidence or uncertainty scores associated with the predictions. Creating predictions requires a Label Studio ML backend set up and configured for your project.
See [Set up machine learning](https://labelstud.io/guide/ml.html) for more details about a Label Studio ML backend.
Reference the ML backend ID in the path of this API call. Get the ML backend ID by [listing the ML backends for a project](https://labelstud.io/api/#operation/api_ml_list).
Parameters
----------
id : int
A unique integer value identifying this ML backend.
batch_size : typing.Optional[int]
Computed number of tasks without predictions that the ML backend needs to predict.
request_options : typing.Optional[RequestOptions]
Request-specific configuration.
Returns
-------
None
Examples
--------
from label_studio_sdk import LabelStudio
client = LabelStudio(
api_key="YOUR_API_KEY",
)
client.ml.predict_all_tasks(
id=1,
)
"""
_response = self._raw_client.predict_all_tasks(id, batch_size=batch_size, request_options=request_options)
return _response.data
def train(
self,
id: int,
*,
use_ground_truth: typing.Optional[bool] = OMIT,
request_options: typing.Optional[RequestOptions] = None,
) -> None:
"""
After you add an ML backend, call this API with the ML backend ID to start training with
already-labeled tasks.
Get the ML backend ID by [listing the ML backends for a project](https://labelstud.io/api/#operation/api_ml_list).
Parameters
----------
id : int
A unique integer value identifying this ML backend.
use_ground_truth : typing.Optional[bool]
Whether to include ground truth annotations in training
request_options : typing.Optional[RequestOptions]
Request-specific configuration.
Returns
-------
None
Examples
--------
from label_studio_sdk import LabelStudio
client = LabelStudio(
api_key="YOUR_API_KEY",
)
client.ml.train(
id=1,
)
"""
_response = self._raw_client.train(id, use_ground_truth=use_ground_truth, request_options=request_options)
return _response.data
def list_model_versions(
self, id: int, *, request_options: typing.Optional[RequestOptions] = None
) -> ListModelVersionsMlResponse:
"""
Get available versions of the model.
Parameters
----------
id : int
request_options : typing.Optional[RequestOptions]
Request-specific configuration.
Returns
-------
ListModelVersionsMlResponse
List of available versions.
Examples
--------
from label_studio_sdk import LabelStudio
client = LabelStudio(
api_key="YOUR_API_KEY",
)
client.ml.list_model_versions(
id=1,
)
"""
_response = self._raw_client.list_model_versions(id, request_options=request_options)
return _response.data
class AsyncMlClient:
def __init__(self, *, client_wrapper: AsyncClientWrapper):
self._raw_client = AsyncRawMlClient(client_wrapper=client_wrapper)
@property
def with_raw_response(self) -> AsyncRawMlClient:
"""
Retrieves a raw implementation of this client that returns raw responses.
Returns
-------
AsyncRawMlClient
"""
return self._raw_client
async def list(
self, *, project: typing.Optional[int] = None, request_options: typing.Optional[RequestOptions] = None
) -> typing.List[MlBackend]:
"""
List all configured ML backends for a specific project by ID.
Use the following cURL command:
```bash
curl http://localhost:8000/api/ml?project={project_id} -H 'Authorization: Token abc123'
Parameters
----------
project : typing.Optional[int]
Project ID
request_options : typing.Optional[RequestOptions]
Request-specific configuration.
Returns
-------
typing.List[MlBackend]
Examples
--------
import asyncio
from label_studio_sdk import AsyncLabelStudio
client = AsyncLabelStudio(
api_key="YOUR_API_KEY",
)
async def main() -> None:
await client.ml.list()
asyncio.run(main())
"""
_response = await self._raw_client.list(project=project, request_options=request_options)
return _response.data
async def create(
self,
*,
auth_method: typing.Optional[CreateMlRequestAuthMethod] = OMIT,
basic_auth_pass: typing.Optional[str] = OMIT,
basic_auth_user: typing.Optional[str] = OMIT,
description: typing.Optional[str] = OMIT,
extra_params: typing.Optional[typing.Dict[str, typing.Any]] = OMIT,
is_interactive: typing.Optional[bool] = OMIT,
project: typing.Optional[int] = OMIT,
timeout: typing.Optional[int] = OMIT,
title: typing.Optional[str] = OMIT,
url: typing.Optional[str] = OMIT,
request_options: typing.Optional[RequestOptions] = None,
) -> MlBackend:
"""
Add an ML backend to a project using the Label Studio UI or by sending a POST request using the following cURL
command:
```bash
curl -X POST -H 'Content-type: application/json' http://localhost:8000/api/ml -H 'Authorization: Token abc123'\\
--data '{"url": "http://localhost:9090", "project": {project_id}}'
Parameters
----------
auth_method : typing.Optional[CreateMlRequestAuthMethod]
Auth method
basic_auth_pass : typing.Optional[str]
Basic auth password
basic_auth_user : typing.Optional[str]
Basic auth user
description : typing.Optional[str]
Description
extra_params : typing.Optional[typing.Dict[str, typing.Any]]
Extra parameters
is_interactive : typing.Optional[bool]
Is interactive
project : typing.Optional[int]
Project ID
timeout : typing.Optional[int]
Response model timeout
title : typing.Optional[str]
Title
url : typing.Optional[str]
ML backend URL
request_options : typing.Optional[RequestOptions]
Request-specific configuration.
Returns
-------
MlBackend
Examples
--------
import asyncio
from label_studio_sdk import AsyncLabelStudio
client = AsyncLabelStudio(
api_key="YOUR_API_KEY",
)
async def main() -> None:
await client.ml.create()
asyncio.run(main())
"""
_response = await self._raw_client.create(
auth_method=auth_method,
basic_auth_pass=basic_auth_pass,
basic_auth_user=basic_auth_user,
description=description,
extra_params=extra_params,
is_interactive=is_interactive,
project=project,
timeout=timeout,
title=title,
url=url,
request_options=request_options,
)
return _response.data
async def get(self, id: int, *, request_options: typing.Optional[RequestOptions] = None) -> MlBackend:
"""
Get details about a specific ML backend connection by ID. For example, make a GET request using the
following cURL command:
```bash
curl http://localhost:8000/api/ml/{ml_backend_ID} -H 'Authorization: Token abc123'
Parameters
----------
id : int
request_options : typing.Optional[RequestOptions]
Request-specific configuration.
Returns
-------
MlBackend
Examples
--------
import asyncio
from label_studio_sdk import AsyncLabelStudio
client = AsyncLabelStudio(
api_key="YOUR_API_KEY",
)
async def main() -> None:
await client.ml.get(
id=1,
)
asyncio.run(main())
"""
_response = await self._raw_client.get(id, request_options=request_options)
return _response.data
async def delete(self, id: int, *, request_options: typing.Optional[RequestOptions] = None) -> None:
"""
Remove an existing ML backend connection by ID. For example, use the
following cURL command:
```bash
curl -X DELETE http://localhost:8000/api/ml/{ml_backend_ID} -H 'Authorization: Token abc123'
Parameters
----------
id : int
request_options : typing.Optional[RequestOptions]
Request-specific configuration.
Returns
-------
None
Examples
--------
import asyncio
from label_studio_sdk import AsyncLabelStudio
client = AsyncLabelStudio(
api_key="YOUR_API_KEY",
)
async def main() -> None:
await client.ml.delete(
id=1,
)
asyncio.run(main())
"""
_response = await self._raw_client.delete(id, request_options=request_options)
return _response.data
async def update(
self,
id: int,
*,
auth_method: typing.Optional[UpdateMlRequestAuthMethod] = OMIT,
basic_auth_pass: typing.Optional[str] = OMIT,
basic_auth_user: typing.Optional[str] = OMIT,
description: typing.Optional[str] = OMIT,
extra_params: typing.Optional[typing.Dict[str, typing.Any]] = OMIT,
is_interactive: typing.Optional[bool] = OMIT,
project: typing.Optional[int] = OMIT,
timeout: typing.Optional[int] = OMIT,
title: typing.Optional[str] = OMIT,
url: typing.Optional[str] = OMIT,
request_options: typing.Optional[RequestOptions] = None,
) -> MlBackend:
"""
Update ML backend parameters using the Label Studio UI or by sending a PATCH request using the following cURL command:
```bash
curl -X PATCH -H 'Content-type: application/json' http://localhost:8000/api/ml/{ml_backend_ID} -H 'Authorization: Token abc123'\\
--data '{"url": "http://localhost:9091"}'
Parameters
----------
id : int
auth_method : typing.Optional[UpdateMlRequestAuthMethod]
Auth method
basic_auth_pass : typing.Optional[str]
Basic auth password
basic_auth_user : typing.Optional[str]
Basic auth user
description : typing.Optional[str]
Description
extra_params : typing.Optional[typing.Dict[str, typing.Any]]
Extra parameters
is_interactive : typing.Optional[bool]
Is interactive
project : typing.Optional[int]
Project ID
timeout : typing.Optional[int]
Response model timeout
title : typing.Optional[str]
Title
url : typing.Optional[str]
ML backend URL
request_options : typing.Optional[RequestOptions]
Request-specific configuration.
Returns
-------
MlBackend
Examples
--------
import asyncio
from label_studio_sdk import AsyncLabelStudio
client = AsyncLabelStudio(
api_key="YOUR_API_KEY",
)
async def main() -> None:
await client.ml.update(
id=1,
)
asyncio.run(main())
"""
_response = await self._raw_client.update(
id,
auth_method=auth_method,
basic_auth_pass=basic_auth_pass,
basic_auth_user=basic_auth_user,
description=description,
extra_params=extra_params,
is_interactive=is_interactive,
project=project,
timeout=timeout,
title=title,
url=url,
request_options=request_options,
)
return _response.data
async def predict_interactive(
self,
id: int,
*,
task: int,
context: typing.Optional[typing.Any] = OMIT,
request_options: typing.Optional[RequestOptions] = None,
) -> None:
"""
Send a request to the machine learning backend set up to be used for interactive preannotations to retrieve a
predicted region based on annotator input.
See [set up machine learning](https://labelstud.io/guide/ml.html#Get-interactive-preannotations) for more.
Parameters
----------
id : int
A unique integer value identifying this ML backend.
task : int
ID of task to annotate
context : typing.Optional[typing.Any]
Context for ML model
request_options : typing.Optional[RequestOptions]
Request-specific configuration.
Returns
-------
None
Examples
--------
import asyncio
from label_studio_sdk import AsyncLabelStudio
client = AsyncLabelStudio(
api_key="YOUR_API_KEY",
)
async def main() -> None:
await client.ml.predict_interactive(
id=1,
task=1,
)
asyncio.run(main())
"""
_response = await self._raw_client.predict_interactive(
id, task=task, context=context, request_options=request_options
)
return _response.data
async def predict_all_tasks(
self,
id: int,
*,
batch_size: typing.Optional[int] = None,
request_options: typing.Optional[RequestOptions] = None,
) -> None:
"""
This endpoint is not available in Label Studio Community Edition. [Learn more about Label Studio Enterprise](https://humansignal.com/goenterprise)
Create predictions for all tasks using a specific ML backend so that you can set up an active learning strategy based on the confidence or uncertainty scores associated with the predictions. Creating predictions requires a Label Studio ML backend set up and configured for your project.
See [Set up machine learning](https://labelstud.io/guide/ml.html) for more details about a Label Studio ML backend.
Reference the ML backend ID in the path of this API call. Get the ML backend ID by [listing the ML backends for a project](https://labelstud.io/api/#operation/api_ml_list).
Parameters
----------
id : int
A unique integer value identifying this ML backend.
batch_size : typing.Optional[int]
Computed number of tasks without predictions that the ML backend needs to predict.
request_options : typing.Optional[RequestOptions]
Request-specific configuration.
Returns
-------
None
Examples
--------
import asyncio
from label_studio_sdk import AsyncLabelStudio
client = AsyncLabelStudio(
api_key="YOUR_API_KEY",
)
async def main() -> None:
await client.ml.predict_all_tasks(
id=1,
)
asyncio.run(main())
"""
_response = await self._raw_client.predict_all_tasks(id, batch_size=batch_size, request_options=request_options)
return _response.data
async def train(
self,
id: int,
*,
use_ground_truth: typing.Optional[bool] = OMIT,
request_options: typing.Optional[RequestOptions] = None,
) -> None:
"""
After you add an ML backend, call this API with the ML backend ID to start training with
already-labeled tasks.
Get the ML backend ID by [listing the ML backends for a project](https://labelstud.io/api/#operation/api_ml_list).
Parameters
----------
id : int
A unique integer value identifying this ML backend.
use_ground_truth : typing.Optional[bool]
Whether to include ground truth annotations in training
request_options : typing.Optional[RequestOptions]
Request-specific configuration.
Returns
-------
None
Examples
--------
import asyncio
from label_studio_sdk import AsyncLabelStudio
client = AsyncLabelStudio(
api_key="YOUR_API_KEY",
)
async def main() -> None:
await client.ml.train(
id=1,
)
asyncio.run(main())
"""
_response = await self._raw_client.train(id, use_ground_truth=use_ground_truth, request_options=request_options)
return _response.data
async def list_model_versions(
self, id: int, *, request_options: typing.Optional[RequestOptions] = None
) -> ListModelVersionsMlResponse:
"""
Get available versions of the model.
Parameters
----------
id : int
request_options : typing.Optional[RequestOptions]
Request-specific configuration.
Returns
-------
ListModelVersionsMlResponse
List of available versions.
Examples
--------
import asyncio
from label_studio_sdk import AsyncLabelStudio
client = AsyncLabelStudio(
api_key="YOUR_API_KEY",
)
async def main() -> None:
await client.ml.list_model_versions(
id=1,
)
asyncio.run(main())
"""
_response = await self._raw_client.list_model_versions(id, request_options=request_options)
return _response.data