rapidata 1.0.0__py3-none-any.whl → 1.2.0__py3-none-any.whl

rapidata/api_client/models/legacy_submit_password_reset_command.py

@@ -0,0 +1,102 @@
+# coding: utf-8
+
+"""
+    Rapidata.Dataset
+
+    No description provided (generated by Openapi Generator https://github.com/openapitools/openapi-generator)
+
+    The version of the OpenAPI document: v1
+    Generated by OpenAPI Generator (https://openapi-generator.tech)
+
+    Do not edit the class manually.
+"""  # noqa: E501
+
+
+from __future__ import annotations
+import pprint
+import re  # noqa: F401
+import json
+
+from pydantic import BaseModel, ConfigDict, Field, StrictStr, field_validator
+from typing import Any, ClassVar, Dict, List
+from typing import Optional, Set
+from typing_extensions import Self
+
+class LegacySubmitPasswordResetCommand(BaseModel):
+    """
+    LegacySubmitPasswordResetCommand
+    """ # noqa: E501
+    t: StrictStr = Field(description="Discriminator value for LegacySubmitPasswordResetCommand", alias="_t")
+    user_id: StrictStr = Field(alias="userId")
+    password: StrictStr
+    password_repeated: StrictStr = Field(alias="passwordRepeated")
+    reset_token: StrictStr = Field(alias="resetToken")
+    __properties: ClassVar[List[str]] = ["_t", "userId", "password", "passwordRepeated", "resetToken"]
+
+    @field_validator('t')
+    def t_validate_enum(cls, value):
+        """Validates the enum"""
+        if value not in set(['LegacySubmitPasswordResetCommand']):
+            raise ValueError("must be one of enum values ('LegacySubmitPasswordResetCommand')")
+        return value
+
+    model_config = ConfigDict(
+        populate_by_name=True,
+        validate_assignment=True,
+        protected_namespaces=(),
+    )
+
+
+    def to_str(self) -> str:
+        """Returns the string representation of the model using alias"""
+        return pprint.pformat(self.model_dump(by_alias=True))
+
+    def to_json(self) -> str:
+        """Returns the JSON representation of the model using alias"""
+        # TODO: pydantic v2: use .model_dump_json(by_alias=True, exclude_unset=True) instead
+        return json.dumps(self.to_dict())
+
+    @classmethod
+    def from_json(cls, json_str: str) -> Optional[Self]:
+        """Create an instance of LegacySubmitPasswordResetCommand from a JSON string"""
+        return cls.from_dict(json.loads(json_str))
+
+    def to_dict(self) -> Dict[str, Any]:
+        """Return the dictionary representation of the model using alias.
+
+        This has the following differences from calling pydantic's
+        `self.model_dump(by_alias=True)`:
+
+        * `None` is only added to the output dict for nullable fields that
+          were set at model initialization. Other fields with value `None`
+          are ignored.
+        """
+        excluded_fields: Set[str] = set([
+        ])
+
+        _dict = self.model_dump(
+            by_alias=True,
+            exclude=excluded_fields,
+            exclude_none=True,
+        )
+        return _dict
+
+    @classmethod
+    def from_dict(cls, obj: Optional[Dict[str, Any]]) -> Optional[Self]:
+        """Create an instance of LegacySubmitPasswordResetCommand from a dict"""
+        if obj is None:
+            return None
+
+        if not isinstance(obj, dict):
+            return cls.model_validate(obj)
+
+        _obj = cls.model_validate({
+            "_t": obj.get("_t") if obj.get("_t") is not None else 'LegacySubmitPasswordResetCommand',
+            "userId": obj.get("userId"),
+            "password": obj.get("password"),
+            "passwordRepeated": obj.get("passwordRepeated"),
+            "resetToken": obj.get("resetToken")
+        })
+        return _obj
+
+

rapidata/api_client_README.md

@@ -97,6 +97,7 @@ Class | Method | HTTP request | Description
 *IdentityApi* | [**identity_get_client_auth_token_post**](rapidata/api_client/docs/IdentityApi.md#identity_get_client_auth_token_post) | **POST** /Identity/GetClientAuthToken | Issues a new auth token using the client credentials.
 *IdentityApi* | [**identity_index_post**](rapidata/api_client/docs/IdentityApi.md#identity_index_post) | **POST** /Identity/Index | Logs in a user by username or email and password.
 *IdentityApi* | [**identity_logout_post**](rapidata/api_client/docs/IdentityApi.md#identity_logout_post) | **POST** /Identity/Logout | Logs out the current user by deleting the refresh token cookie.
+*IdentityApi* | [**identity_register_temporary_post**](rapidata/api_client/docs/IdentityApi.md#identity_register_temporary_post) | **POST** /Identity/RegisterTemporary | Registers and logs in a temporary customer.
 *IdentityApi* | [**identity_request_reset_post**](rapidata/api_client/docs/IdentityApi.md#identity_request_reset_post) | **POST** /Identity/RequestReset | Request a password reset for a user.
 *IdentityApi* | [**identity_signup_post**](rapidata/api_client/docs/IdentityApi.md#identity_signup_post) | **POST** /Identity/Signup | Signs up a new user.
 *IdentityApi* | [**identity_submit_reset_post**](rapidata/api_client/docs/IdentityApi.md#identity_submit_reset_post) | **POST** /Identity/SubmitReset | Updates the password of a user after a password reset request.
@@ -131,7 +132,9 @@ Class | Method | HTTP request | Description
 *RapidApi* | [**rapid_query_validation_rapids_get**](rapidata/api_client/docs/RapidApi.md#rapid_query_validation_rapids_get) | **GET** /Rapid/QueryValidationRapids | Queries the validation rapids for a specific validation set.
 *RapidApi* | [**rapid_skip_user_guess_post**](rapidata/api_client/docs/RapidApi.md#rapid_skip_user_guess_post) | **POST** /Rapid/SkipUserGuess | Skips a Rapid for the user.
 *RapidApi* | [**rapid_validate_current_rapid_bag_get**](rapidata/api_client/docs/RapidApi.md#rapid_validate_current_rapid_bag_get) | **GET** /Rapid/ValidateCurrentRapidBag | Validates that the rapids associated with the current user are active.
+*RapidataIdentityAPIApi* | [**root_get**](rapidata/api_client/docs/RapidataIdentityAPIApi.md#root_get) | **GET** / | 
 *SimpleWorkflowApi* | [**simple_workflow_get_result_overview_get**](rapidata/api_client/docs/SimpleWorkflowApi.md#simple_workflow_get_result_overview_get) | **GET** /SimpleWorkflow/GetResultOverview | Get the result overview for a simple workflow.
+*UserInfoApi* | [**connect_userinfo_get**](rapidata/api_client/docs/UserInfoApi.md#connect_userinfo_get) | **GET** /connect/userinfo | Retrieves information about the authenticated user.
 *ValidationApi* | [**validation_add_validation_rapid_post**](rapidata/api_client/docs/ValidationApi.md#validation_add_validation_rapid_post) | **POST** /Validation/AddValidationRapid | Adds a new validation rapid to the specified validation set.
 *ValidationApi* | [**validation_add_validation_text_rapid_post**](rapidata/api_client/docs/ValidationApi.md#validation_add_validation_text_rapid_post) | **POST** /Validation/AddValidationTextRapid | Adds a new validation rapid to the specified validation set.
 *ValidationApi* | [**validation_create_validation_set_post**](rapidata/api_client/docs/ValidationApi.md#validation_create_validation_set_post) | **POST** /Validation/CreateValidationSet | Creates a new empty validation set.
@@ -273,9 +276,11 @@ Class | Method | HTTP request | Description
  - [ImportValidationSetFromFileResult](rapidata/api_client/docs/ImportValidationSetFromFileResult.md)
  - [InProgressRapidModel](rapidata/api_client/docs/InProgressRapidModel.md)
  - [IssueAuthTokenResult](rapidata/api_client/docs/IssueAuthTokenResult.md)
- - [IssueClientAuthTokenResult](rapidata/api_client/docs/IssueClientAuthTokenResult.md)
  - [LabelingSelection](rapidata/api_client/docs/LabelingSelection.md)
  - [LanguageUserFilterModel](rapidata/api_client/docs/LanguageUserFilterModel.md)
+ - [LegacyIssueClientAuthTokenResult](rapidata/api_client/docs/LegacyIssueClientAuthTokenResult.md)
+ - [LegacyRequestPasswordResetCommand](rapidata/api_client/docs/LegacyRequestPasswordResetCommand.md)
+ - [LegacySubmitPasswordResetCommand](rapidata/api_client/docs/LegacySubmitPasswordResetCommand.md)
  - [Line](rapidata/api_client/docs/Line.md)
  - [LinePayload](rapidata/api_client/docs/LinePayload.md)
  - [LinePoint](rapidata/api_client/docs/LinePoint.md)
@@ -334,7 +339,6 @@ Class | Method | HTTP request | Description
  - [RapidResultModel](rapidata/api_client/docs/RapidResultModel.md)
  - [RapidResultModelResult](rapidata/api_client/docs/RapidResultModelResult.md)
  - [RapidSkippedModel](rapidata/api_client/docs/RapidSkippedModel.md)
- - [RequestPasswordResetCommand](rapidata/api_client/docs/RequestPasswordResetCommand.md)
  - [RootFilter](rapidata/api_client/docs/RootFilter.md)
  - [SendCompletionMailStepModel](rapidata/api_client/docs/SendCompletionMailStepModel.md)
  - [Shape](rapidata/api_client/docs/Shape.md)
@@ -354,7 +358,6 @@ Class | Method | HTTP request | Description
  - [StaticSelection](rapidata/api_client/docs/StaticSelection.md)
  - [SubmitCocoModel](rapidata/api_client/docs/SubmitCocoModel.md)
  - [SubmitCocoResult](rapidata/api_client/docs/SubmitCocoResult.md)
- - [SubmitPasswordResetCommand](rapidata/api_client/docs/SubmitPasswordResetCommand.md)
  - [TextAsset](rapidata/api_client/docs/TextAsset.md)
  - [TextAssetModel](rapidata/api_client/docs/TextAssetModel.md)
  - [TextMetadata](rapidata/api_client/docs/TextMetadata.md)
@@ -403,6 +406,10 @@ Authentication schemes defined for the API:
 - **API key parameter name**: Authorization
 - **Location**: HTTP header
 
+<a id="oauth2"></a>
+### oauth2
+
+
 
 ## Author
 

rapidata/rapidata_client/__init__.py

@@ -1,13 +1,24 @@
 from .rapidata_client import RapidataClient
-from .workflow import ClassifyWorkflow, TranscriptionWorkflow, CompareWorkflow, FreeTextWorkflow
+from .workflow import (
+    ClassifyWorkflow,
+    TranscriptionWorkflow,
+    CompareWorkflow,
+    FreeTextWorkflow,
+)
 from .selection import (
     DemographicSelection,
     LabelingSelection,
     ValidationSelection,
     ConditionalValidationSelection,
+    CappedSelection,
 )
 from .referee import NaiveReferee, ClassifyEarlyStoppingReferee
-from .metadata import PrivateTextMetadata, PublicTextMetadata, PromptMetadata, TranscriptionMetadata
+from .metadata import (
+    PrivateTextMetadata,
+    PublicTextMetadata,
+    PromptMetadata,
+    TranscriptionMetadata,
+)
 from .feature_flags import FeatureFlags
 from .country_codes import CountryCodes
 from .assets import MediaAsset, TextAsset, MultiAsset

rapidata/rapidata_client/assets/multi_asset.py

@@ -23,6 +23,8 @@ class MultiAsset(BaseAsset):
         Args:
             assets (List[BaseAsset]): A list of BaseAsset instances to be managed together.
         """
+        if len(assets) != 2:
+            raise ValueError("Assets must come in pairs for comparison tasks.")
         self.assets = assets
 
     def __len__(self) -> int:

rapidata/rapidata_client/dataset/rapidata_dataset.py

@@ -9,6 +9,7 @@ from rapidata.api_client.models.upload_text_sources_to_dataset_model import (
     UploadTextSourcesToDatasetModel,
 )
 from rapidata.rapidata_client.metadata.base_metadata import Metadata
+from rapidata.rapidata_client.assets import TextAsset, MediaAsset, MultiAsset
 from rapidata.service import LocalFileService
 from rapidata.service.openapi_service import OpenAPIService
 from concurrent.futures import ThreadPoolExecutor, as_completed
@@ -22,7 +23,8 @@ class RapidataDataset:
         self.openapi_service = openapi_service
         self.local_file_service = LocalFileService()
 
-    def add_texts(self, texts: list[str]):
+    def add_texts(self, text_assets: list[TextAsset]):
+        texts = [text.text for text in text_assets]
         model = UploadTextSourcesToDatasetModel(
             datasetId=self.dataset_id, textSources=texts
         )
@@ -32,24 +34,26 @@ class RapidataDataset:
 
     def add_media_from_paths(
         self,
-        media_paths: list[str | list[str]],
+        media_paths: list[MediaAsset | MultiAsset],
         metadata: list[Metadata] | None = None,
         max_workers: int = 10,
     ):
         if metadata is not None and len(metadata) != len(media_paths):
             raise ValueError(
-                "metadata must be None or have the same length as image_paths"
+                "metadata must be None or have the same length as media_paths"
             )
 
-        def upload_datapoint(media_paths_rapid: str | list[str], meta: Metadata | None) -> None:
-            if isinstance(media_paths_rapid, list) and not all(
-                os.path.exists(media_path) for media_path in media_paths_rapid
-            ):
-                raise FileNotFoundError(f"File not found: {media_paths_rapid}")
-            elif isinstance(media_paths_rapid, str) and not os.path.exists(
-                media_paths_rapid
-            ):
-                raise FileNotFoundError(f"File not found: {media_paths_rapid}")
+        def upload_datapoint(media_asset: MediaAsset | MultiAsset, meta: Metadata | None) -> None:
+            if isinstance(media_asset, MediaAsset):
+                paths = [media_asset.path]
+            elif isinstance(media_asset, MultiAsset):
+                paths = [asset.path for asset in media_asset.assets if isinstance(asset, MediaAsset)]
+            else:
+                raise ValueError(f"Unsupported asset type: {type(media_asset)}")
+            
+            assert all(
+                os.path.exists(media_path) for media_path in paths
+            ), "All media paths must exist on the local filesystem."
 
             meta_model = meta.to_model() if meta else None
             model = DatapointMetadataModel(
@@ -63,14 +67,14 @@ class RapidataDataset:
 
             self.openapi_service.dataset_api.dataset_create_datapoint_post(
                 model=model,
-                files=media_paths_rapid if isinstance(media_paths_rapid, list) else [media_paths_rapid] # type: ignore
+                files=paths # type: ignore
             )
 
         total_uploads = len(media_paths)
         with ThreadPoolExecutor(max_workers=max_workers) as executor:
             futures = [
-                executor.submit(upload_datapoint, media_paths, meta)
-                for media_paths, meta in zip_longest(media_paths, metadata or [])
+                executor.submit(upload_datapoint, media_asset, meta)
+                for media_asset, meta in zip_longest(media_paths, metadata or [])
             ]
 
             with tqdm(total=total_uploads, desc="Uploading datapoints") as pbar:

rapidata/rapidata_client/dataset/validation_set_builder.py

@@ -160,8 +160,8 @@ class ValidationSetBuilder:
         self,
         asset: MediaAsset | TextAsset,
         question: str,
-        transcription: list[str],
-        correct_words: list[str],
+        transcription: str,
+        truths: list[int],
         strict_grading: bool | None = None,
         metadata: list[Metadata] = [],
     ):
@@ -171,9 +171,9 @@ class ValidationSetBuilder:
             asset (MediaAsset | TextAsset): The asset for the rapid.
             question (str): The question for the rapid.
             transcription (list[str]): The transcription for the rapid.
-            correct_words (list[str]): The list of correct words for the rapid.
-            strict_grading (bool | None, optional): The strict grading flag for the rapid. Defaults to None.
-            metadata (list[Metadata], optional): The metadata for the rapid. Defaults to an empty list.
+            truths (list[int]): The list of indices of the true word selections.
+            strict_grading (bool | None, optional): The strict grading for the rapid. Defaults to None.
+            metadata (list[Metadata], optional): The metadata for the rapid.
 
         Returns:
             ValidationSetBuilder: The ValidationSetBuilder instance.
@@ -183,16 +183,14 @@ class ValidationSetBuilder:
         """
         transcription_words = [
             TranscriptionWord(word=word, wordIndex=i)
-            for i, word in enumerate(transcription)
+            for i, word in enumerate(transcription.split())
         ]
 
-        correct_transcription_words = []
-        for word in correct_words:
-            if word not in transcription:
-                raise ValueError(f"Correct word '{word}' not found in transcription")
-            correct_transcription_words.append(
-                TranscriptionWord(word=word, wordIndex=transcription.index(word))
-            )
+        true_words = []
+        for idx in truths:
+            if idx > len(transcription_words) - 1:
+                raise ValueError(f"Index {idx} is out of bounds")
+            true_words.append(transcription_words[idx])
 
         payload = TranscriptionPayload(
             _t="TranscriptionPayload", title=question, transcription=transcription_words
@@ -200,7 +198,7 @@ class ValidationSetBuilder:
 
         model_truth = TranscriptionTruth(
             _t="TranscriptionTruth",
-            correctWords=correct_transcription_words,
+            correctWords=true_words,
             strictGrading=strict_grading,
         )
 
@@ -211,7 +209,7 @@ class ValidationSetBuilder:
                 payload=payload,
                 truths=model_truth,
                 metadata=metadata,
-                randomCorrectProbability=1 / len(transcription),
+                randomCorrectProbability = 1 / len(transcription_words),
             )
         )
 

rapidata/rapidata_client/order/rapidata_order.py

@@ -3,7 +3,9 @@ from rapidata.rapidata_client.dataset.rapidata_dataset import RapidataDataset
 from rapidata.service.openapi_service import OpenAPIService
 import json
 from rapidata.api_client.exceptions import ApiException
-
+from typing import cast
+from rapidata.api_client.models.workflow_artifact_model import WorkflowArtifactModel
+from tqdm import tqdm
 
 class RapidataOrder:
     """
@@ -26,6 +28,7 @@ class RapidataOrder:
         self.openapi_service = openapi_service
         self.order_id = order_id
         self._dataset = dataset
+        self._workflow_id = None
 
     def submit(self):
         """
@@ -49,27 +52,55 @@ class RapidataOrder:
         """
         return self.openapi_service.order_api.order_get_by_id_get(self.order_id)
 
-    def wait_for_done(self):
+    def display_progress_bar(self, refresh_rate=5):
         """
-        Blocking call that waits for the order to be done. Exponential backoff is used to check the status of the order.
+        Displays a progress bar for the order processing using tqdm.
+        
+        :param refresh_rate: How often to refresh the progress bar, in seconds.
+        :type refresh_rate: float
         """
-        wait_time = 1
-        back_off_factor = 1.1
-        minimum_poll_interval = 60  # 1 minute
-
-        while True:
-            time.sleep(wait_time)
-            result = self.get_status()
-            if result.state == "ManualReview":
-                print(
-                    "Order is in manual review. Please contact support for approval. Will continue polling."
-                )
+        total_rapids = self._get_total_rapids()
+        with tqdm(total=total_rapids, desc="Processing order", unit="rapids") as pbar:
+            completed_rapids = 0
+            while True:
+                current_completed = self._get_completed_rapids()
+                if current_completed > completed_rapids:
+                    pbar.update(current_completed - completed_rapids)
+                    completed_rapids = current_completed
+                
+                if completed_rapids >= total_rapids:
+                    break
+                
+                time.sleep(refresh_rate)
 
-            if result.state == "Completed" or result.state == "Failed":
+    def _get_workflow_id(self):
+        if self._workflow_id:
+            return self._workflow_id
+        
+        for _ in range(2):
+            try:
+                order_result = self.openapi_service.order_api.order_get_by_id_get(self.order_id)
+                pipeline = self.openapi_service.pipeline_api.pipeline_id_get(order_result.pipeline_id)
+                self._workflow_id = cast(WorkflowArtifactModel, pipeline.artifacts["workflow-artifact"].actual_instance).workflow_id
                 break
-            wait_time = max(
-                minimum_poll_interval, wait_time * back_off_factor
-            )  # poll at least every 10 minutes
+            except Exception:
+                time.sleep(2)
+        if not self._workflow_id:
+            raise Exception("Order has not started yet. Please wait for a few seconds and try again.")
+        return self._workflow_id
+    
+    def _get_total_rapids(self):
+        workflow_id = self._get_workflow_id()
+        return self.openapi_service.workflow_api.workflow_get_progress_get(workflow_id).total
+
+    def _get_completed_rapids(self):
+        workflow_id = self._get_workflow_id()
+        return self.openapi_service.workflow_api.workflow_get_progress_get(workflow_id).completed
+
+    def get_progress_percentage(self):
+        workflow_id = self._get_workflow_id()
+        progress = self.openapi_service.workflow_api.workflow_get_progress_get(workflow_id)
+        return progress.completion_percentage
 
     def get_results(self):
         """

rapidata/rapidata_client/order/rapidata_order_builder.py

@@ -13,6 +13,9 @@ from rapidata.api_client.models.create_order_model_workflow import (
     CreateOrderModelWorkflow,
 )
 from rapidata.api_client.models.country_user_filter_model import CountryUserFilterModel
+from rapidata.api_client.models.language_user_filter_model import (
+    LanguageUserFilterModel,
+)
 from rapidata.rapidata_client.feature_flags import FeatureFlags
 from rapidata.rapidata_client.metadata.base_metadata import Metadata
 from rapidata.rapidata_client.dataset.rapidata_dataset import RapidataDataset
@@ -25,6 +28,10 @@ from rapidata.service.openapi_service import OpenAPIService
 
 from rapidata.rapidata_client.workflow.compare_workflow import CompareWorkflow
 
+from rapidata.rapidata_client.assets import MediaAsset, TextAsset, MultiAsset
+
+from typing import cast, Sequence
+
 
 class RapidataOrderBuilder:
     """Builder object for creating Rapidata orders.
@@ -52,17 +59,16 @@ class RapidataOrderBuilder:
         self._openapi_service = openapi_service
         self._workflow: Workflow | None = None
         self._referee: Referee | None = None
-        self._media_paths: list[str | list[str]] = []
         self._metadata: list[Metadata] | None = None
         self._aggregator: AggregatorType | None = None
         self._validation_set_id: str | None = None
         self._feature_flags: FeatureFlags | None = None
         self._country_codes: list[str] | None = None
+        self._language_codes: list[str] | None = None
         self._selections: list[Selection] = []
         self._rapids_per_bag: int = 2
         self._priority: int = 50
-        self._texts: list[str] | None = None
-        self._media_paths: list[str | list[str]] = []
+        self._assets: list[MediaAsset] | list[TextAsset] | list[MultiAsset] = []
 
     def _to_model(self) -> CreateOrderModel:
         """
@@ -80,22 +86,32 @@ class RapidataOrderBuilder:
         if self._referee is None:
             print("No referee provided, using default NaiveReferee.")
             self._referee = NaiveReferee()
-        if self._country_codes is None:
-            country_filter = None
-        else:
-            country_filter = CountryUserFilterModel(
-                _t="CountryFilter", countries=self._country_codes
+
+        user_filters = []
+
+        if self._country_codes is not None:
+            user_filters.append(
+                CreateOrderModelUserFiltersInner(
+                    CountryUserFilterModel(
+                        _t="CountryFilter", countries=self._country_codes
+                    )
+                )
+            )
+
+        if self._language_codes is not None:
+            user_filters.append(
+                CreateOrderModelUserFiltersInner(
+                    LanguageUserFilterModel(
+                        _t="LanguageFilter", languages=self._language_codes
+                    )
+                )
             )
 
         return CreateOrderModel(
             _t="CreateOrderModel",
             orderName=self._name,
             workflow=CreateOrderModelWorkflow(self._workflow.to_model()),
-            userFilters=(
-                [CreateOrderModelUserFiltersInner(country_filter)]
-                if country_filter
-                else []
-            ),
+            userFilters=user_filters,
             referee=CreateOrderModelReferee(self._referee.to_model()),
             validationSetId=self._validation_set_id,
             featureFlags=(
@@ -129,8 +145,12 @@ class RapidataOrderBuilder:
         if isinstance(
             self._workflow, CompareWorkflow
         ):  # Temporary fix; will be handled by backend in the future
+            assert all(isinstance(item, MultiAsset) for item in self._assets), (
+                "The media paths must be of type MultiAsset for comparison tasks."
+            )
+            media_paths = cast(list[MultiAsset], self._assets)
             assert all(
-                [len(path) == 2 for path in self._media_paths]
+                [len(path) == 2 for path in media_paths]
             ), "The media paths must come in pairs for comparison tasks."
 
         result = self._openapi_service.order_api.order_create_post(
@@ -145,22 +165,18 @@ class RapidataOrderBuilder:
             openapi_service=self._openapi_service,
         )
 
-        if self._media_paths and self._texts:
+        if not self._assets:
             raise ValueError(
-                "You cannot provide both media paths and texts to the same order."
+                "You must provide assets to start the order."
             )
+        if all(isinstance(item, TextAsset) for item in self._assets):
+            assets = cast(list[TextAsset], self._assets)
+            order.dataset.add_texts(assets)
 
-        if not self._media_paths and not self._texts:
-            raise ValueError(
-                "You must provide either media paths or texts to the order."
-            )
-
-        if self._texts:
-            order.dataset.add_texts(self._texts)
-
-        if self._media_paths:
+        elif all(isinstance(item, (MediaAsset, MultiAsset)) for item in self._assets):
+            assets = cast(list[MediaAsset | MultiAsset], self._assets)
             order.dataset.add_media_from_paths(
-                self._media_paths, self._metadata, max_workers
+                assets, self._metadata, max_workers
             )
 
         if submit:
@@ -196,60 +212,60 @@ class RapidataOrderBuilder:
 
     def media(
         self,
-        media_paths: list[str | list[str]],
-        metadata: list[Metadata] | None = None,
+        asset: list[MediaAsset] | list[TextAsset] | list[MultiAsset],
+        metadata: Sequence[Metadata] | None = None,
     ) -> "RapidataOrderBuilder":
         """
         Set the media assets for the order.
 
         Args:
-            media_paths (list[str | list[str]]): The paths of the media assets to be set.
+            media_paths (list[MediaAsset] | list[TextAsset] | list[MultiAsset]): The paths of the media assets to be set.
             metadata (list[Metadata] | None, optional): Metadata for the media assets. Defaults to None.
 
         Returns:
             RapidataOrderBuilder: The updated RapidataOrderBuilder instance.
         """
-        self._media_paths = media_paths
-        self._metadata = metadata
+        self._assets = asset
+        self._metadata = metadata # type: ignore
         return self
 
-    def texts(self, texts: list[str]) -> "RapidataOrderBuilder":
+    def feature_flags(self, feature_flags: FeatureFlags) -> "RapidataOrderBuilder":
         """
-        Set the TextAssets for the order.
+        Set the feature flags for the order.
 
         Args:
-            texts (list[str]): The texts to be set.
+            feature_flags (FeatureFlags): The feature flags to be set.
 
         Returns:
             RapidataOrderBuilder: The updated RapidataOrderBuilder instance.
         """
-        self._texts = texts
+        self._feature_flags = feature_flags
         return self
 
-    def feature_flags(self, feature_flags: FeatureFlags) -> "RapidataOrderBuilder":
+    def country_filter(self, country_codes: list[str]) -> "RapidataOrderBuilder":
         """
-        Set the feature flags for the order.
+        Set the target country codes for the order. E.g. `country_codes=["DE", "CH", "AT"]` for Germany, Switzerland, and Austria.
 
         Args:
-            feature_flags (FeatureFlags): The feature flags to be set.
+            country_codes (list[str]): The country codes to be set.
 
         Returns:
             RapidataOrderBuilder: The updated RapidataOrderBuilder instance.
         """
-        self._feature_flags = feature_flags
+        self._country_codes = country_codes
         return self
 
-    def country_filter(self, country_codes: list[str]) -> "RapidataOrderBuilder":
+    def language_filter(self, language_codes: list[str]) -> "RapidataOrderBuilder":
         """
-        Set the target country codes for the order.
+        Set the target language codes for the order. E.g. `language_codes=["de", "fr", "it"]` for German, French, and Italian.
 
         Args:
-            country_codes (list[str]): The country codes to be set.
+            language_codes (list[str]): The language codes to be set.
 
         Returns:
             RapidataOrderBuilder: The updated RapidataOrderBuilder instance.
         """
-        self._country_codes = country_codes
+        self._language_codes = language_codes
         return self
 
     def aggregator(self, aggregator: AggregatorType) -> "RapidataOrderBuilder":

rapidata/rapidata_client/selection/__init__.py

@@ -3,3 +3,4 @@ from .demographic_selection import DemographicSelection
 from .labeling_selection import LabelingSelection
 from .validation_selection import ValidationSelection
 from .conditional_validation_selection import ConditionalValidationSelection
+from .capped_selection import CappedSelection