rapidata 2.31.1__py3-none-any.whl → 2.32.0__py3-none-any.whl

rapidata/api_client/models/__init__.py

@@ -109,6 +109,7 @@ from rapidata.api_client.models.create_order_model_referee import CreateOrderMod
 from rapidata.api_client.models.create_order_model_workflow import CreateOrderModelWorkflow
 from rapidata.api_client.models.create_order_result import CreateOrderResult
 from rapidata.api_client.models.create_rapid_result import CreateRapidResult
+from rapidata.api_client.models.create_sample_model import CreateSampleModel
 from rapidata.api_client.models.create_simple_pipeline_model import CreateSimplePipelineModel
 from rapidata.api_client.models.create_simple_pipeline_model_artifacts_inner import CreateSimplePipelineModelArtifactsInner
 from rapidata.api_client.models.create_simple_pipeline_model_pipeline_steps_inner import CreateSimplePipelineModelPipelineStepsInner

rapidata/api_client/models/create_sample_model.py

@@ -0,0 +1,87 @@
+# 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
+from typing import Any, ClassVar, Dict, List
+from typing import Optional, Set
+from typing_extensions import Self
+
+class CreateSampleModel(BaseModel):
+    """
+    The model used to create a sample to a participant.
+    """ # noqa: E501
+    identifier: StrictStr = Field(description="The identifier used to correlate samples of different participants.")
+    __properties: ClassVar[List[str]] = ["identifier"]
+
+    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 CreateSampleModel 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 CreateSampleModel from a dict"""
+        if obj is None:
+            return None
+
+        if not isinstance(obj, dict):
+            return cls.model_validate(obj)
+
+        _obj = cls.model_validate({
+            "identifier": obj.get("identifier")
+        })
+        return _obj
+
+

rapidata/api_client_README.md

@@ -76,7 +76,6 @@ Class | Method | HTTP request | Description
 *BenchmarkApi* | [**benchmark_benchmark_id_get**](rapidata/api_client/docs/BenchmarkApi.md#benchmark_benchmark_id_get) | **GET** /benchmark/{benchmarkId} | Returns a single benchmark by its ID.
 *BenchmarkApi* | [**benchmark_benchmark_id_name_put**](rapidata/api_client/docs/BenchmarkApi.md#benchmark_benchmark_id_name_put) | **PUT** /benchmark/{benchmarkId}/name | Updates the name of a benchmark.
 *BenchmarkApi* | [**benchmark_benchmark_id_participant_participant_id_delete**](rapidata/api_client/docs/BenchmarkApi.md#benchmark_benchmark_id_participant_participant_id_delete) | **DELETE** /benchmark/{benchmarkId}/participant/{participantId} | Deletes a participant on a benchmark.
-*BenchmarkApi* | [**benchmark_benchmark_id_participant_participant_id_get**](rapidata/api_client/docs/BenchmarkApi.md#benchmark_benchmark_id_participant_participant_id_get) | **GET** /benchmark/{benchmarkId}/participant/{participantId} | Gets a participant by it's Id.
 *BenchmarkApi* | [**benchmark_benchmark_id_participants_get**](rapidata/api_client/docs/BenchmarkApi.md#benchmark_benchmark_id_participants_get) | **GET** /benchmark/{benchmarkId}/participants | Query all participants within a benchmark
 *BenchmarkApi* | [**benchmark_benchmark_id_participants_participant_id_disable_post**](rapidata/api_client/docs/BenchmarkApi.md#benchmark_benchmark_id_participants_participant_id_disable_post) | **POST** /benchmark/{benchmarkId}/participants/{participantId}/disable | This endpoint disables a participant in a benchmark. this means that the participant will no longer actively be matched up against other participants and not collect further results. It will still be visible in the leaderboard.
 *BenchmarkApi* | [**benchmark_benchmark_id_participants_participant_id_submit_post**](rapidata/api_client/docs/BenchmarkApi.md#benchmark_benchmark_id_participants_participant_id_submit_post) | **POST** /benchmark/{benchmarkId}/participants/{participantId}/submit | Submits a participant to a benchmark.
@@ -163,6 +162,11 @@ Class | Method | HTTP request | Description
 *OrderApi* | [**order_unsupported_post**](rapidata/api_client/docs/OrderApi.md#order_unsupported_post) | **POST** /order/unsupported | Notifies the admins that a user wants to create an order with an unsupported label type or data type.
 *OrderApi* | [**orders_get**](rapidata/api_client/docs/OrderApi.md#orders_get) | **GET** /orders | Queries orders based on a filter, page, and sort criteria.
 *OrderApi* | [**orders_public_get**](rapidata/api_client/docs/OrderApi.md#orders_public_get) | **GET** /orders/public | Retrieves orders that are public and can be cloned by any user.
+*ParticipantApi* | [**participant_participant_id_delete**](rapidata/api_client/docs/ParticipantApi.md#participant_participant_id_delete) | **DELETE** /participant/{participantId} | Deletes a participant on a benchmark.
+*ParticipantApi* | [**participant_participant_id_get**](rapidata/api_client/docs/ParticipantApi.md#participant_participant_id_get) | **GET** /participant/{participantId} | Gets a participant by it's Id.
+*ParticipantApi* | [**participant_participant_id_sample_post**](rapidata/api_client/docs/ParticipantApi.md#participant_participant_id_sample_post) | **POST** /participant/{participantId}/sample | Adds a sample to a participant.
+*ParticipantApi* | [**participant_sample_sample_id_delete**](rapidata/api_client/docs/ParticipantApi.md#participant_sample_sample_id_delete) | **DELETE** /participant-sample/{sampleId} | Deletes a sample.
+*ParticipantApi* | [**participants_participant_id_submit_post**](rapidata/api_client/docs/ParticipantApi.md#participants_participant_id_submit_post) | **POST** /participants/{participantId}/submit | Submits a participant to a benchmark.
 *PipelineApi* | [**pipeline_id_workflow_config_artifact_id_put**](rapidata/api_client/docs/PipelineApi.md#pipeline_id_workflow_config_artifact_id_put) | **PUT** /pipeline/{id}/workflow-config/{artifactId} | Updates the workflow configuration for a pipeline.
 *PipelineApi* | [**pipeline_id_workflow_config_put**](rapidata/api_client/docs/PipelineApi.md#pipeline_id_workflow_config_put) | **PUT** /pipeline/{id}/workflow-config | Updates the workflow configuration for a pipeline.
 *PipelineApi* | [**pipeline_pipeline_id_campaign_artifact_id_put**](rapidata/api_client/docs/PipelineApi.md#pipeline_pipeline_id_campaign_artifact_id_put) | **PUT** /pipeline/{pipelineId}/campaign/{artifactId} | Updates a specific campaign for a pipeline.
@@ -297,6 +301,7 @@ Class | Method | HTTP request | Description
  - [CreateOrderModelWorkflow](rapidata/api_client/docs/CreateOrderModelWorkflow.md)
  - [CreateOrderResult](rapidata/api_client/docs/CreateOrderResult.md)
  - [CreateRapidResult](rapidata/api_client/docs/CreateRapidResult.md)
+ - [CreateSampleModel](rapidata/api_client/docs/CreateSampleModel.md)
  - [CreateSimplePipelineModel](rapidata/api_client/docs/CreateSimplePipelineModel.md)
  - [CreateSimplePipelineModelArtifactsInner](rapidata/api_client/docs/CreateSimplePipelineModelArtifactsInner.md)
  - [CreateSimplePipelineModelPipelineStepsInner](rapidata/api_client/docs/CreateSimplePipelineModelPipelineStepsInner.md)

rapidata/rapidata_client/__init__.py

@@ -9,12 +9,19 @@ from .selection import (
     RetrievalMode,
     EffortEstimationSelection,
 )
-from .metadata import (
+from .datapoints import Datapoint
+from .datapoints.metadata import (
     PrivateTextMetadata,
     PublicTextMetadata,
     PromptMetadata,
     SelectWordsMetadata,
 )
+from .datapoints.assets import (
+    MediaAsset, 
+    TextAsset, 
+    MultiAsset, 
+    RapidataDataTypes
+)
 from .settings import (
     RapidataSettings, 
     TranslationBehaviourOptions,
@@ -27,12 +34,6 @@ from .settings import (
     AllowNeitherBoth,
     )
 from .country_codes import CountryCodes
-from .assets import (
-    MediaAsset, 
-    TextAsset, 
-    MultiAsset, 
-    RapidataDataTypes
-)
 from .filter import (
     CountryFilter,
     LanguageFilter,
@@ -55,3 +56,4 @@ from .logging import (
 )
 
 from .validation import Box
+from .exceptions import FailedUploadException

rapidata/rapidata_client/benchmark/rapidata_benchmark.py

@@ -15,9 +15,10 @@ from rapidata.rapidata_client.logging import logger
 from rapidata.service.openapi_service import OpenAPIService
 
 from rapidata.rapidata_client.benchmark.leaderboard.rapidata_leaderboard import RapidataLeaderboard
-from rapidata.rapidata_client.metadata import PromptIdentifierMetadata
-from rapidata.rapidata_client.assets import MediaAsset
+from rapidata.rapidata_client.datapoints.metadata import PromptIdentifierMetadata
+from rapidata.rapidata_client.datapoints.assets import MediaAsset
 from rapidata.rapidata_client.order._rapidata_dataset import RapidataDataset
+from rapidata.rapidata_client.datapoints.datapoint import Datapoint
 
 class RapidataBenchmark:
     """
@@ -287,7 +288,7 @@ class RapidataBenchmark:
         dataset = RapidataDataset(participant_result.dataset_id, self.__openapi_service)
         
         try:
-            dataset._add_datapoints(assets, prompts_metadata)
+            dataset.add_datapoints([Datapoint(asset=asset, metadata=metadata) for asset, metadata in zip(assets, prompts_metadata)])
         except Exception as e:
             logger.warning(f"An error occurred while adding datapoints to the dataset: {e}")
             upload_progress = self.__openapi_service.dataset_api.dataset_dataset_id_progress_get(

rapidata/rapidata_client/datapoints/__init__.py

@@ -0,0 +1,3 @@
+from .datapoint import Datapoint
+from .assets import MediaAsset, MultiAsset, TextAsset
+from .metadata import Metadata, PromptMetadata, PrivateTextMetadata, PublicTextMetadata, SelectWordsMetadata, MediaAssetMetadata, PromptIdentifierMetadata

rapidata/rapidata_client/{assets → datapoints/assets}/_media_asset.py

@@ -7,7 +7,7 @@ Implements lazy loading for URL-based media to prevent unnecessary downloads.
 from typing import Optional, cast
 import os
 from io import BytesIO
-from rapidata.rapidata_client.assets._base_asset import BaseAsset
+from rapidata.rapidata_client.datapoints.assets._base_asset import BaseAsset
 import requests
 import re
 from PIL import Image
@@ -16,7 +16,7 @@ import tempfile
 from pydantic import StrictStr, StrictBytes
 import logging
 from functools import cached_property
-from rapidata.rapidata_client.assets._sessions import SessionManager
+from rapidata.rapidata_client.datapoints.assets._sessions import SessionManager
 from rapidata.rapidata_client.logging import logger
 
 class MediaAsset(BaseAsset):

rapidata/rapidata_client/{assets → datapoints/assets}/_multi_asset.py

@@ -3,8 +3,8 @@
 Defines the MultiAsset class for handling multiple BaseAsset instances.
 """
 
-from rapidata.rapidata_client.assets._base_asset import BaseAsset
-from rapidata.rapidata_client.assets import MediaAsset, TextAsset
+from rapidata.rapidata_client.datapoints.assets._base_asset import BaseAsset
+from rapidata.rapidata_client.datapoints.assets import MediaAsset, TextAsset
 from typing import Iterator, Sequence, cast
 
 

rapidata/rapidata_client/{assets → datapoints/assets}/_text_asset.py

@@ -3,7 +3,7 @@
 Defines the TextAsset class for handling textual data within assets.
 """
 
-from rapidata.rapidata_client.assets._base_asset import BaseAsset
+from rapidata.rapidata_client.datapoints.assets._base_asset import BaseAsset
 
 
 class TextAsset(BaseAsset):

rapidata/rapidata_client/datapoints/datapoint.py

@@ -0,0 +1,108 @@
+from typing import Sequence, cast
+from rapidata.rapidata_client.datapoints.assets import MediaAsset, TextAsset, MultiAsset, BaseAsset
+from rapidata.rapidata_client.datapoints.metadata import Metadata
+from rapidata.api_client.models.dataset_dataset_id_datapoints_post_request_metadata_inner import DatasetDatasetIdDatapointsPostRequestMetadataInner
+from rapidata.api_client.models.create_datapoint_from_text_sources_model import CreateDatapointFromTextSourcesModel
+from pydantic import StrictStr, StrictBytes
+
+class Datapoint:
+    def __init__(self, asset: MediaAsset | TextAsset | MultiAsset, metadata: Sequence[Metadata] | None = None):
+        if not isinstance(asset, (MediaAsset, TextAsset, MultiAsset)):
+            raise TypeError("Asset must be of type MediaAsset, TextAsset, or MultiAsset.")
+        
+        if metadata and not isinstance(metadata, Sequence):
+            raise TypeError("Metadata must be a list of Metadata objects.")
+        
+        if metadata and not all(isinstance(m, Metadata) for m in metadata):
+            raise TypeError("All metadata objects must be of type Metadata.")
+        
+        self.asset = asset
+        self.metadata = metadata
+
+    def _get_effective_asset_type(self) -> type:
+        """Get the effective asset type, handling MultiAsset by looking at its first asset."""
+        if isinstance(self.asset, MultiAsset):
+            return type(self.asset.assets[0])
+        return type(self.asset)
+
+    def is_media_asset(self) -> bool:
+        """Check if this datapoint contains media assets."""
+        effective_type = self._get_effective_asset_type()
+        return issubclass(effective_type, MediaAsset)
+
+    def is_text_asset(self) -> bool:
+        """Check if this datapoint contains text assets."""
+        effective_type = self._get_effective_asset_type()
+        return issubclass(effective_type, TextAsset)
+
+    def get_texts(self) -> list[str]:
+        """Extract text content from the asset(s)."""
+        if isinstance(self.asset, TextAsset):
+            return [self.asset.text]
+        elif isinstance(self.asset, MultiAsset):
+            texts = []
+            for asset in self.asset.assets:
+                if isinstance(asset, TextAsset):
+                    texts.append(asset.text)
+            return texts
+        else:
+            raise ValueError(f"Cannot extract text from asset type: {type(self.asset)}")
+
+    def get_media_assets(self) -> list[MediaAsset]:
+        """Extract media assets from the datapoint."""
+        if isinstance(self.asset, MediaAsset):
+            return [self.asset]
+        elif isinstance(self.asset, MultiAsset):
+            media_assets = []
+            for asset in self.asset.assets:
+                if isinstance(asset, MediaAsset):
+                    media_assets.append(asset)
+            return media_assets
+        else:
+            raise ValueError(f"Cannot extract media assets from asset type: {type(self.asset)}")
+
+    def get_local_file_paths(self) -> list[StrictStr | tuple[StrictStr, StrictBytes] | StrictBytes]:
+        """Get local file paths for media assets that are stored locally."""
+        if not self.is_media_asset():
+            return []
+        
+        media_assets = self.get_media_assets()
+        return [asset.to_file() for asset in media_assets if asset.is_local()]
+
+    def get_urls(self) -> list[str]:
+        """Get URLs for media assets that are remote."""
+        if not self.is_media_asset():
+            return []
+        
+        media_assets = self.get_media_assets()
+        return [asset.path for asset in media_assets if not asset.is_local()]
+
+    def get_prepared_metadata(self) -> list[DatasetDatasetIdDatapointsPostRequestMetadataInner]:
+        """Prepare metadata for API upload."""
+        metadata: list[DatasetDatasetIdDatapointsPostRequestMetadataInner] = []
+        if self.metadata:
+            for meta in self.metadata:
+                meta_model = meta.to_model() if meta else None
+                if meta_model:
+                    metadata.append(DatasetDatasetIdDatapointsPostRequestMetadataInner(meta_model))
+        return metadata
+
+    def create_text_upload_model(self, index: int) -> CreateDatapointFromTextSourcesModel:
+        """Create the model for uploading text datapoints."""
+        if not self.is_text_asset():
+            raise ValueError("Cannot create text upload model for non-text asset")
+        
+        texts = self.get_texts()
+        metadata = self.get_prepared_metadata()
+        
+        return CreateDatapointFromTextSourcesModel(
+            textSources=texts,
+            sortIndex=index,
+            metadata=metadata,
+        )
+
+    def __str__(self):
+        return f"Datapoint(asset={self.asset})"
+
+    def __repr__(self):
+        return self.__str__()

rapidata/rapidata_client/{metadata → datapoints/metadata}/_media_asset_metadata.py

@@ -1,6 +1,6 @@
 from rapidata.api_client.models.prompt_asset_metadata_input import PromptAssetMetadataInput
 from rapidata.api_client.models.url_asset_input import UrlAssetInput
-from rapidata.rapidata_client.metadata._base_metadata import Metadata
+from rapidata.rapidata_client.datapoints.metadata._base_metadata import Metadata
 from rapidata.api_client.models.prompt_asset_metadata_input_asset import PromptAssetMetadataInputAsset
 
 

rapidata/rapidata_client/{metadata → datapoints/metadata}/_private_text_metadata.py

@@ -1,7 +1,7 @@
 from rapidata.api_client.models.private_text_metadata_input import (
     PrivateTextMetadataInput,
 )
-from rapidata.rapidata_client.metadata._base_metadata import Metadata
+from rapidata.rapidata_client.datapoints.metadata._base_metadata import Metadata
 
 
 class PrivateTextMetadata(Metadata):

rapidata/rapidata_client/{metadata → datapoints/metadata}/_prompt_identifier_metadata.py

@@ -1,4 +1,4 @@
-from rapidata.rapidata_client.metadata._base_metadata import Metadata
+from rapidata.rapidata_client.datapoints.metadata._base_metadata import Metadata
 from rapidata.api_client.models.private_text_metadata_input import (
     PrivateTextMetadataInput,
 )

rapidata/rapidata_client/{metadata → datapoints/metadata}/_prompt_metadata.py

@@ -1,4 +1,4 @@
-from rapidata.rapidata_client.metadata._base_metadata import Metadata
+from rapidata.rapidata_client.datapoints.metadata._base_metadata import Metadata
 from rapidata.api_client.models.prompt_metadata_input import PromptMetadataInput
 
 

rapidata/rapidata_client/{metadata → datapoints/metadata}/_public_text_metadata.py

@@ -1,7 +1,7 @@
 from rapidata.api_client.models.public_text_metadata_input import (
     PublicTextMetadataInput,
 )
-from rapidata.rapidata_client.metadata._base_metadata import Metadata
+from rapidata.rapidata_client.datapoints.metadata._base_metadata import Metadata
 
 
 class PublicTextMetadata(Metadata):

rapidata/rapidata_client/{metadata → datapoints/metadata}/_select_words_metadata.py

@@ -1,7 +1,7 @@
 from rapidata.api_client.models.transcription_metadata_input import (
     TranscriptionMetadataInput,
 )
-from rapidata.rapidata_client.metadata._base_metadata import Metadata
+from rapidata.rapidata_client.datapoints.metadata._base_metadata import Metadata
 
 
 class SelectWordsMetadata(Metadata):

rapidata/rapidata_client/demographic/demographic_manager.py

@@ -1,5 +1,5 @@
 from rapidata.service.openapi_service import OpenAPIService
-from rapidata.rapidata_client.assets import MediaAsset
+from rapidata.rapidata_client.datapoints.assets import MediaAsset
 from rapidata.api_client.models.create_demographic_rapid_model import CreateDemographicRapidModel
 from rapidata.api_client.models.classify_payload import ClassifyPayload
 from rapidata.rapidata_client.logging import logger

rapidata/rapidata_client/exceptions/__init__.py

@@ -0,0 +1 @@
+from .failed_upload_exception import FailedUploadException

rapidata/rapidata_client/exceptions/failed_upload_exception.py

@@ -0,0 +1,19 @@
+from rapidata.rapidata_client.datapoints.datapoint import Datapoint
+from rapidata.rapidata_client.order._rapidata_dataset import RapidataDataset
+from rapidata.rapidata_client.order.rapidata_order import RapidataOrder
+
+
+class FailedUploadException(Exception):
+    """Custom error class for Failed Uploads to the Rapidata order."""
+    def __init__(
+        self, 
+        dataset: RapidataDataset,
+        order: RapidataOrder,
+        failed_uploads: list[Datapoint]
+    ):
+        self.dataset = dataset
+        self.order = order
+        self.failed_uploads = failed_uploads
+
+    def __str__(self) -> str:
+        return f"Failed to upload {self.failed_uploads}"