rapidata 2.36.1__py3-none-any.whl → 2.36.2__py3-none-any.whl

rapidata/rapidata_client/validation/validation_set_manager.py

@@ -18,6 +18,8 @@ from rapidata.api_client.models.page_info import PageInfo
 from rapidata.api_client.models.root_filter import RootFilter
 from rapidata.api_client.models.filter import Filter
 from rapidata.api_client.models.sort_criterion import SortCriterion
+from rapidata.api_client.models.sort_direction import SortDirection
+from rapidata.api_client.models.filter_operator import FilterOperator
 
 from rapidata.rapidata_client.validation.rapids.box import Box
 
@@ -27,6 +29,11 @@ from rapidata.rapidata_client.logging import (
     RapidataOutputManager,
 )
 from tqdm import tqdm
+from rapidata.rapidata_client.workflow import Workflow
+from rapidata.rapidata_client.datapoints._datapoint import Datapoint
+from rapidata.rapidata_client.validation.rapids.rapids import Rapid
+from rapidata.rapidata_client.settings._rapidata_setting import RapidataSetting
+from typing import Sequence
 
 
 class ValidationSetManager:
@@ -42,6 +49,25 @@ class ValidationSetManager:
         self.rapid = RapidsManager()
         logger.debug("ValidationSetManager initialized")
 
+    def _create_order_validation_set(
+        self,
+        workflow: Workflow,
+        order_name: str,
+        datapoints: list[Datapoint],
+        settings: Sequence[RapidataSetting] | None = None,
+    ) -> RapidataValidationSet:
+        rapids: list[Rapid] = []
+        for datapoint in datapoints:
+            rapids.append(
+                Rapid(
+                    asset=datapoint.asset,
+                    payload=workflow._to_payload(datapoint),
+                    metadata=datapoint.metadata,
+                    settings=settings,
+                )
+            )
+        return self._submit(name=order_name, rapids=rapids, dimensions=[])
+
     def create_classification_set(
         self,
         name: str,
@@ -543,7 +569,10 @@ class ValidationSetManager:
         return self._submit(name=name, rapids=rapids, dimensions=dimensions)
 
     def _submit(
-        self, name: str, rapids: list[Rapid], dimensions: list[str] | None
+        self,
+        name: str,
+        rapids: list[Rapid],
+        dimensions: list[str] | None,
     ) -> RapidataValidationSet:
         logger.debug("Creating validation set")
         validation_set_id = (
@@ -590,7 +619,7 @@ class ValidationSetManager:
         managed_print()
         managed_print(
             f"Validation set '{name}' created with ID {validation_set_id}\n",
-            f"Now viewable under: https://app.{self.__openapi_service.environment}/validation-set/detail/{validation_set_id}",
+            f"Now viewable under: {validation_set.validation_set_details_page}",
             sep="",
         )
 
@@ -637,10 +666,18 @@ class ValidationSetManager:
                 QueryModel(
                     page=PageInfo(index=1, size=amount),
                     filter=RootFilter(
-                        filters=[Filter(field="Name", operator="Contains", value=name)]
+                        filters=[
+                            Filter(
+                                field="Name",
+                                operator=FilterOperator.CONTAINS,
+                                value=name,
+                            )
+                        ]
                     ),
                     sortCriteria=[
-                        SortCriterion(direction="Desc", propertyName="CreatedAt")
+                        SortCriterion(
+                            direction=SortDirection.DESC, propertyName="CreatedAt"
+                        )
                     ],
                 )
             )

rapidata/rapidata_client/workflow/_base_workflow.py

@@ -5,9 +5,21 @@ from rapidata.api_client.models.simple_workflow_model import SimpleWorkflowModel
 from rapidata.api_client.models.evaluation_workflow_model import EvaluationWorkflowModel
 from rapidata.api_client.models.compare_workflow_model import CompareWorkflowModel
 from rapidata.rapidata_client.referee._base_referee import Referee
+from rapidata.api_client import (
+    ClassifyPayload,
+    ComparePayload,
+    LocatePayload,
+    ScrubPayload,
+    TranscriptionPayload,
+    LinePayload,
+    FreeTextPayload,
+)
+from rapidata.rapidata_client.datapoints._datapoint import Datapoint
+from rapidata.api_client.models.rapid_modality import RapidModality
 
 
 class Workflow(ABC):
+    modality: RapidModality
 
     def __init__(self, type: str):
         self._type = type
@@ -18,6 +30,21 @@ class Workflow(ABC):
             "_t": self._type,
         }
 
+    @abstractmethod
+    def _to_payload(
+        self,
+        datapoint: Datapoint,
+    ) -> (
+        ClassifyPayload
+        | ComparePayload
+        | LocatePayload
+        | ScrubPayload
+        | LinePayload
+        | FreeTextPayload
+        | TranscriptionPayload
+    ):
+        pass
+
     @abstractmethod
     def _to_model(
         self,

rapidata/rapidata_client/workflow/_classify_workflow.py

@@ -1,8 +1,15 @@
 from typing import Any
-from rapidata.api_client.models.attach_category_rapid_blueprint import AttachCategoryRapidBlueprint
+from rapidata.api_client.models.attach_category_rapid_blueprint import (
+    AttachCategoryRapidBlueprint,
+)
 from rapidata.api_client.models.simple_workflow_model import SimpleWorkflowModel
-from rapidata.api_client.models.simple_workflow_model_blueprint import SimpleWorkflowModelBlueprint
+from rapidata.api_client.models.simple_workflow_model_blueprint import (
+    SimpleWorkflowModelBlueprint,
+)
 from rapidata.rapidata_client.workflow import Workflow
+from rapidata.api_client import ClassifyPayload
+from rapidata.rapidata_client.datapoints._datapoint import Datapoint
+from rapidata.api_client.models.rapid_modality import RapidModality
 
 
 class ClassifyWorkflow(Workflow):
@@ -21,29 +28,38 @@ class ClassifyWorkflow(Workflow):
         _options (list[str]): The list of classification options.
     """
 
+    modality = RapidModality.CLASSIFY
+
     def __init__(self, instruction: str, answer_options: list[str]):
         super().__init__(type="SimpleWorkflowConfig")
-        self._question = instruction
-        self._options = answer_options
+        self._instruction = instruction
+        self._answer_options = answer_options
 
     def _to_dict(self) -> dict[str, Any]:
         return {
             **super()._to_dict(),
             "blueprint": {
                 "_t": "ClassifyBlueprint",
-                "title": self._question,
-                "possibleCategories": self._options,
-            }
+                "title": self._instruction,
+                "possibleCategories": self._answer_options,
+            },
         }
 
     def _to_model(self) -> SimpleWorkflowModel:
         blueprint = AttachCategoryRapidBlueprint(
             _t="ClassifyBlueprint",
-            title=self._question,
-            possibleCategories=self._options,
+            title=self._instruction,
+            possibleCategories=self._answer_options,
         )
 
         return SimpleWorkflowModel(
             _t="SimpleWorkflow",
             blueprint=SimpleWorkflowModelBlueprint(blueprint),
         )
+
+    def _to_payload(self, datapoint: Datapoint) -> ClassifyPayload:
+        return ClassifyPayload(
+            _t="ClassifyPayload",
+            possibleCategories=self._answer_options,
+            title=self._instruction,
+        )

rapidata/rapidata_client/workflow/_compare_workflow.py

@@ -5,6 +5,9 @@ from rapidata.api_client.models.simple_workflow_model_blueprint import (
 from rapidata.rapidata_client.workflow import Workflow
 from rapidata.api_client.models.compare_rapid_blueprint import CompareRapidBlueprint
 from rapidata.api_client.models.simple_workflow_model import SimpleWorkflowModel
+from rapidata.api_client import ComparePayload
+from rapidata.rapidata_client.datapoints._datapoint import Datapoint
+from rapidata.api_client.models.rapid_modality import RapidModality
 
 
 class CompareWorkflow(Workflow):
@@ -21,6 +24,8 @@ class CompareWorkflow(Workflow):
         instruction (str): The instruction to be used for comparison.
     """
 
+    modality = RapidModality.COMPARE
+
     def __init__(self, instruction: str, a_b_names: list[str] | None = None):
         super().__init__(type="CompareWorkflowConfig")
         self._instruction = instruction
@@ -44,3 +49,9 @@ class CompareWorkflow(Workflow):
             _t="SimpleWorkflow",
             blueprint=SimpleWorkflowModelBlueprint(blueprint),
         )
+
+    def _to_payload(self, datapoint: Datapoint) -> ComparePayload:
+        return ComparePayload(
+            _t="ComparePayload",
+            criteria=self._instruction,
+        )

rapidata/rapidata_client/workflow/_draw_workflow.py

@@ -1,22 +1,30 @@
 from rapidata.api_client.models.simple_workflow_model import SimpleWorkflowModel
-from rapidata.api_client.models.simple_workflow_model_blueprint import SimpleWorkflowModelBlueprint
+from rapidata.api_client.models.simple_workflow_model_blueprint import (
+    SimpleWorkflowModelBlueprint,
+)
 from rapidata.api_client.models.line_rapid_blueprint import LineRapidBlueprint
 from rapidata.rapidata_client.workflow._base_workflow import Workflow
+from rapidata.api_client import LinePayload
+from rapidata.rapidata_client.datapoints._datapoint import Datapoint
+from rapidata.api_client.models.rapid_modality import RapidModality
 
 
 class DrawWorkflow(Workflow):
+    modality = RapidModality.LINE
 
     def __init__(self, target: str):
         super().__init__(type="SimpleWorkflowConfig")
         self._target = target
 
     def _to_model(self) -> SimpleWorkflowModel:
-        blueprint = LineRapidBlueprint(
-            _t="LineBlueprint",
-            target=self._target
-        )
+        blueprint = LineRapidBlueprint(_t="LineBlueprint", target=self._target)
 
         return SimpleWorkflowModel(
-            _t="SimpleWorkflow",
-            blueprint=SimpleWorkflowModelBlueprint(blueprint)
+            _t="SimpleWorkflow", blueprint=SimpleWorkflowModelBlueprint(blueprint)
+        )
+
+    def _to_payload(self, datapoint: Datapoint) -> LinePayload:
+        return LinePayload(
+            _t="LinePayload",
+            target=self._target,
         )

rapidata/rapidata_client/workflow/_evaluation_workflow.py

@@ -1,5 +1,7 @@
 from rapidata.api_client.models.evaluation_workflow_model import EvaluationWorkflowModel
 from rapidata.rapidata_client.workflow._base_workflow import Workflow
+from rapidata.rapidata_client.datapoints._datapoint import Datapoint
+from rapidata.api_client.models.rapid_modality import RapidModality
 
 
 class EvaluationWorkflow(Workflow):
@@ -7,12 +9,14 @@ class EvaluationWorkflow(Workflow):
     A workflow to run evaluation orders.
 
     This is used internally only and should not be necessary to be used by clients.
-    
+
     Args:
         validation_set_id (str): a source for the tasks that will be sent to the user
         should_accept_incorrect (bool): indicates if the user should get feedback on their answers if they answer wrong. If set to true the user will not notice that he was tested.
     """
 
+    modality = RapidModality.NONE
+
     def __init__(self, validation_set_id: str, should_accept_incorrect: bool):
         super().__init__("EvaluationWorkflow")
         self.validation_set_id = validation_set_id
@@ -24,3 +28,6 @@ class EvaluationWorkflow(Workflow):
             validationSetId=self.validation_set_id,
             shouldAcceptIncorrect=self.should_accept_incorrect,
         )
+
+    def _to_payload(self, datapoint: Datapoint):
+        raise NotImplementedError("EvaluationWorkflow does not have a payload")

rapidata/rapidata_client/workflow/_free_text_workflow.py

@@ -5,6 +5,9 @@ from rapidata.api_client.models.simple_workflow_model_blueprint import (
 )
 from rapidata.rapidata_client.workflow import Workflow
 from rapidata.api_client.models.free_text_rapid_blueprint import FreeTextRapidBlueprint
+from rapidata.api_client import FreeTextPayload
+from rapidata.rapidata_client.datapoints._datapoint import Datapoint
+from rapidata.api_client.models.rapid_modality import RapidModality
 
 
 class FreeTextWorkflow(Workflow):
@@ -23,6 +26,8 @@ class FreeTextWorkflow(Workflow):
             Should always specify that the LLM should respond with 'not spam' or 'spam'.
     """
 
+    modality = RapidModality.FREETEXT
+
     def __init__(self, instruction: str, validation_system_prompt: str | None = None):
         super().__init__(type="SimpleWorkflowConfig")
         self._instruction = instruction
@@ -49,3 +54,9 @@ class FreeTextWorkflow(Workflow):
             _t="SimpleWorkflow",
             blueprint=SimpleWorkflowModelBlueprint(blueprint),
         )
+
+    def _to_payload(self, datapoint: Datapoint) -> FreeTextPayload:
+        return FreeTextPayload(
+            _t="FreeTextPayload",
+            question=self._instruction,
+        )

rapidata/rapidata_client/workflow/_locate_workflow.py

@@ -1,22 +1,30 @@
 from rapidata.api_client.models.simple_workflow_model import SimpleWorkflowModel
-from rapidata.api_client.models.simple_workflow_model_blueprint import SimpleWorkflowModelBlueprint
+from rapidata.api_client.models.simple_workflow_model_blueprint import (
+    SimpleWorkflowModelBlueprint,
+)
 from rapidata.api_client.models.locate_rapid_blueprint import LocateRapidBlueprint
 from rapidata.rapidata_client.workflow._base_workflow import Workflow
+from rapidata.api_client import LocatePayload
+from rapidata.rapidata_client.datapoints._datapoint import Datapoint
+from rapidata.api_client.models.rapid_modality import RapidModality
 
 
 class LocateWorkflow(Workflow):
+    modality = RapidModality.LOCATE
 
     def __init__(self, target: str):
         super().__init__(type="SimpleWorkflowConfig")
         self._target = target
 
     def _to_model(self) -> SimpleWorkflowModel:
-        blueprint = LocateRapidBlueprint(
-            _t="LocateBlueprint",
-            target=self._target
-        )
+        blueprint = LocateRapidBlueprint(_t="LocateBlueprint", target=self._target)
 
         return SimpleWorkflowModel(
-            _t="SimpleWorkflow",
-            blueprint=SimpleWorkflowModelBlueprint(blueprint)
+            _t="SimpleWorkflow", blueprint=SimpleWorkflowModelBlueprint(blueprint)
+        )
+
+    def _to_payload(self, datapoint: Datapoint) -> LocatePayload:
+        return LocatePayload(
+            _t="LocatePayload",
+            target=self._target,
         )

rapidata/rapidata_client/workflow/_ranking_workflow.py

@@ -1,30 +1,48 @@
-from rapidata.api_client import CompareWorkflowModelPairMakerConfig, OnlinePairMakerConfigModel, EloConfigModel
+from rapidata.api_client import (
+    CompareWorkflowModelPairMakerConfig,
+    OnlinePairMakerConfigModel,
+    EloConfigModel,
+)
 from rapidata.api_client.models.compare_workflow_model import CompareWorkflowModel
 from rapidata.rapidata_client.workflow._base_workflow import Workflow
 from rapidata.rapidata_client.datapoints.metadata import PromptMetadata
-from rapidata.api_client.models.dataset_dataset_id_datapoints_post_request_metadata_inner import DatasetDatasetIdDatapointsPostRequestMetadataInner
+from rapidata.api_client.models.dataset_dataset_id_datapoints_post_request_metadata_inner import (
+    DatasetDatasetIdDatapointsPostRequestMetadataInner,
+)
+from rapidata.api_client import ComparePayload
+from rapidata.rapidata_client.datapoints._datapoint import Datapoint
+from rapidata.api_client.models.rapid_modality import RapidModality
 
 
 class RankingWorkflow(Workflow):
-    def __init__(self,
-                 criteria: str,
-                 total_comparison_budget: int,
-                 random_comparisons_ratio,
-                 elo_start: int = 1200,
-                 elo_k_factor: int = 40,
-                 elo_scaling_factor: int = 400,
-                 context: str | None = None,
-                 ):
+    modality = RapidModality.COMPARE
+
+    def __init__(
+        self,
+        criteria: str,
+        total_comparison_budget: int,
+        random_comparisons_ratio,
+        elo_start: int = 1200,
+        elo_k_factor: int = 40,
+        elo_scaling_factor: int = 400,
+        context: str | None = None,
+    ):
         super().__init__(type="CompareWorkflowConfig")
 
-        self.context = [DatasetDatasetIdDatapointsPostRequestMetadataInner(
-            PromptMetadata(context).to_model())
-            ] if context else None
+        self.context = (
+            [
+                DatasetDatasetIdDatapointsPostRequestMetadataInner(
+                    PromptMetadata(context).to_model()
+                )
+            ]
+            if context
+            else None
+        )
 
         self.criteria = criteria
         self.pair_maker_config = CompareWorkflowModelPairMakerConfig(
             OnlinePairMakerConfigModel(
-                _t='OnlinePairMaker',
+                _t="OnlinePairMaker",
                 totalComparisonBudget=total_comparison_budget,
                 randomMatchesRatio=random_comparisons_ratio,
             )
@@ -45,3 +63,9 @@ class RankingWorkflow(Workflow):
             pairMakerConfig=self.pair_maker_config,
             metadata=self.context,
         )
+
+    def _to_payload(self, datapoint: Datapoint) -> ComparePayload:
+        return ComparePayload(
+            _t="ComparePayload",
+            criteria=self.criteria,
+        )

rapidata/rapidata_client/workflow/_select_words_workflow.py

@@ -1,14 +1,24 @@
 from rapidata.api_client.models.simple_workflow_model import SimpleWorkflowModel
-from rapidata.api_client.models.simple_workflow_model_blueprint import SimpleWorkflowModelBlueprint
-from rapidata.api_client.models.transcription_rapid_blueprint import TranscriptionRapidBlueprint
+from rapidata.api_client.models.simple_workflow_model_blueprint import (
+    SimpleWorkflowModelBlueprint,
+)
+from rapidata.api_client.models.transcription_rapid_blueprint import (
+    TranscriptionRapidBlueprint,
+)
 from rapidata.rapidata_client.workflow._base_workflow import Workflow
+from rapidata.api_client import TranscriptionPayload, TranscriptionWord
+from rapidata.rapidata_client.datapoints._datapoint import Datapoint
+from rapidata.rapidata_client.datapoints.metadata._select_words_metadata import (
+    SelectWordsMetadata,
+)
+from rapidata.api_client.models.rapid_modality import RapidModality
 
 
 class SelectWordsWorkflow(Workflow):
     """
     A workflow for select words tasks.
 
-    This class represents a select words workflow 
+    This class represents a select words workflow
     where datapoints have a sentence attached to them where words can be selected.
 
     Attributes:
@@ -18,17 +28,41 @@ class SelectWordsWorkflow(Workflow):
         instruction (str): The instruction to be provided for the select words task.
     """
 
+    modality = RapidModality.TRANSCRIPTION
+
     def __init__(self, instruction: str):
         super().__init__(type="SimpleWorkflowConfig")
         self._instruction = instruction
 
     def _to_model(self) -> SimpleWorkflowModel:
         blueprint = TranscriptionRapidBlueprint(
-            _t="TranscriptionBlueprint",
-            title=self._instruction
+            _t="TranscriptionBlueprint", title=self._instruction
         )
 
         return SimpleWorkflowModel(
-            _t="SimpleWorkflow",
-            blueprint=SimpleWorkflowModelBlueprint(blueprint)
+            _t="SimpleWorkflow", blueprint=SimpleWorkflowModelBlueprint(blueprint)
+        )
+
+    def _to_payload(self, datapoint: Datapoint) -> TranscriptionPayload:
+        assert (
+            datapoint.metadata is not None
+        ), "SelectWordsWorkflow requires a metadata datapoint"
+
+        assert any(
+            isinstance(metadata, SelectWordsMetadata) for metadata in datapoint.metadata
+        ), "SelectWordsWorkflow requires a SelectWordsMetadata datapoint"
+
+        select_words_metadata = next(
+            metadata
+            for metadata in datapoint.metadata
+            if isinstance(metadata, SelectWordsMetadata)
+        )
+
+        return TranscriptionPayload(
+            _t="TranscriptionPayload",
+            title=self._instruction,
+            transcription=[
+                TranscriptionWord(word=word, wordIndex=i)
+                for i, word in enumerate(select_words_metadata.select_words.split())
+            ],
         )

rapidata/rapidata_client/workflow/_timestamp_workflow.py

@@ -1,14 +1,19 @@
 from rapidata.api_client.models.simple_workflow_model import SimpleWorkflowModel
-from rapidata.api_client.models.simple_workflow_model_blueprint import SimpleWorkflowModelBlueprint
+from rapidata.api_client.models.simple_workflow_model_blueprint import (
+    SimpleWorkflowModelBlueprint,
+)
 from rapidata.api_client.models.scrub_rapid_blueprint import ScrubRapidBlueprint
 from rapidata.rapidata_client.workflow._base_workflow import Workflow
+from rapidata.api_client import ScrubPayload
+from rapidata.rapidata_client.datapoints._datapoint import Datapoint
+from rapidata.api_client.models.rapid_modality import RapidModality
 
 
 class TimestampWorkflow(Workflow):
     """
     A workflow for timestamp tasks.
 
-    This class represents a timestamp workflow 
+    This class represents a timestamp workflow
     where audio or video content receives timestamps based on the instruction.
 
     Attributes:
@@ -18,17 +23,21 @@ class TimestampWorkflow(Workflow):
         instruction (str): The instruction to be provided for the timestamp task.
     """
 
+    modality = RapidModality.SCRUB
+
     def __init__(self, instruction: str):
         super().__init__(type="SimpleWorkflowConfig")
         self._instruction = instruction
 
     def _to_model(self) -> SimpleWorkflowModel:
-        blueprint = ScrubRapidBlueprint(
-            _t="ScrubBlueprint",
-            target=self._instruction
-        )
+        blueprint = ScrubRapidBlueprint(_t="ScrubBlueprint", target=self._instruction)
 
         return SimpleWorkflowModel(
-            _t="SimpleWorkflow",
-            blueprint=SimpleWorkflowModelBlueprint(blueprint)
+            _t="SimpleWorkflow", blueprint=SimpleWorkflowModelBlueprint(blueprint)
+        )
+
+    def _to_payload(self, datapoint: Datapoint) -> ScrubPayload:
+        return ScrubPayload(
+            _t="ScrubPayload",
+            target=self._instruction,
         )

rapidata/service/openapi_service.py

@@ -127,7 +127,7 @@ class OpenAPIService:
     @property
     def benchmark_api(self) -> BenchmarkApi:
         return BenchmarkApi(self.api_client)
-    
+
     @property
     def participant_api(self) -> ParticipantApi:
         return ParticipantApi(self.api_client)

{rapidata-2.36.1.dist-info → rapidata-2.36.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: rapidata
-Version: 2.36.1
+Version: 2.36.2
 Summary: Rapidata package containing the Rapidata Python Client to interact with the Rapidata Web API in an easy way.
 License: Apache-2.0
 Author: Rapidata AG