rapidata 1.4.3__py3-none-any.whl → 1.4.5__py3-none-any.whl

rapidata/__init__.py

@@ -10,7 +10,7 @@ from .rapidata_client import (
     ConditionalValidationSelection,
     CappedSelection,
     NaiveReferee,
-    ClassifyEarlyStoppingReferee,
+    EarlyStoppingReferee,
     PrivateTextMetadata,
     PublicTextMetadata,
     PromptMetadata,

rapidata/rapidata_client/__init__.py

@@ -12,7 +12,7 @@ from .selection import (
     ConditionalValidationSelection,
     CappedSelection,
 )
-from .referee import NaiveReferee, ClassifyEarlyStoppingReferee
+from .referee import NaiveReferee, EarlyStoppingReferee
 from .metadata import (
     PrivateTextMetadata,
     PublicTextMetadata,

rapidata/rapidata_client/order/rapidata_order.py

@@ -1,4 +1,4 @@
-import time
+from time import sleep
 from rapidata.rapidata_client.dataset.rapidata_dataset import RapidataDataset
 from rapidata.service.openapi_service import OpenAPIService
 import json
@@ -61,11 +61,11 @@ class RapidataOrder:
         :param refresh_rate: How often to refresh the progress bar, in seconds.
         :type refresh_rate: float
         """
-        total_rapids = self._get_total_rapids()
+        total_rapids = self._get_workflow_progress().total
         with tqdm(total=total_rapids, desc="Processing order", unit="rapids") as pbar:
             completed_rapids = 0
             while True:
-                current_completed = self._get_completed_rapids()
+                current_completed = self._get_workflow_progress().completed
                 if current_completed > completed_rapids:
                     pbar.update(current_completed - completed_rapids)
                     completed_rapids = current_completed
@@ -73,7 +73,7 @@ class RapidataOrder:
                 if completed_rapids >= total_rapids:
                     break
 
-                time.sleep(refresh_rate)
+                sleep(refresh_rate)
 
     def _get_workflow_id(self):
         if self._workflow_id:
@@ -86,31 +86,38 @@ class RapidataOrder:
                 self._workflow_id = cast(WorkflowArtifactModel, pipeline.artifacts["workflow-artifact"].actual_instance).workflow_id
                 break
             except Exception:
-                time.sleep(2)
+                sleep(2)
         if not self._workflow_id:
-            raise Exception("Order has not started yet. Please wait for a few seconds and try again.")
+            raise Exception("Order has not started yet. Please start it or 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):
+    
+    def _get_workflow_progress(self):
         workflow_id = self._get_workflow_id()
-        return self.openapi_service.workflow_api.workflow_get_progress_get(workflow_id).completed
+        progress = None
+        for _ in range(2):
+            try:
+                progress = self.openapi_service.workflow_api.workflow_get_progress_get(workflow_id)
+                break
+            except Exception:
+                sleep(5)
 
-    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
+        if not progress:
+            raise Exception(f"Failed to get progress. Please try again in a few seconds.")
+        
+        return progress
 
+        
     def get_results(self):
         """
-        Gets the results of the order.
+        Gets the results of the order. 
+        If the order is still processing, this method will block until the order is completed and then return the results.
 
         :return: The results of the order.
         :rtype: dict
         """
+        while self.get_status().state == "Processing":
+            sleep(5)
+
         try:
             # Get the raw result string
             result_str = self.openapi_service.order_api.order_result_get(id=self.order_id)

rapidata/rapidata_client/rapidata_client.py

@@ -28,8 +28,8 @@ class RapidataClient:
         self,
         client_id: str,
         client_secret: str,
-        endpoint: str = "https://api.app.rapidata.ai",
-        token_url: str = "https://api.app.rapidata.ai/connect/token",
+        endpoint: str = "https://auth.rapidata.ai",
+        token_url: str = "https://auth.rapidata.ai/connect/token",
         oauth_scope: str = "openid",
         cert_path: str | None = None,
     ):

rapidata/rapidata_client/referee/__init__.py

@@ -1,3 +1,3 @@
 from .base_referee import Referee
 from .naive_referee import NaiveReferee #as MaxVoteReferee
-from .classify_early_stopping_referee import ClassifyEarlyStoppingReferee
+from .early_stopping_referee import EarlyStoppingReferee

rapidata/rapidata_client/referee/{classify_early_stopping_referee.py → early_stopping_referee.py}

@@ -5,7 +5,7 @@ from rapidata.api_client.models.early_stopping_referee_model import (
 )
 
 
-class ClassifyEarlyStoppingReferee(Referee):
+class EarlyStoppingReferee(Referee):
     """A referee that stops the task when confidence in the winning category exceeds a threshold.
 
     This referee implements an early stopping mechanism for classification tasks.
@@ -15,6 +15,8 @@ class ClassifyEarlyStoppingReferee(Referee):
     The threshold behaves logarithmically, meaning small increments (e.g., from 0.99
     to 0.999) can significantly impact the stopping criteria.
 
+    This referee is supported for the classification and compare tasks.
+
     Attributes:
         threshold (float): The confidence threshold for early stopping.
         max_vote_count (int): The maximum number of votes allowed before stopping.

rapidata/rapidata_client/referee/naive_referee.py

@@ -3,31 +3,31 @@ from rapidata.rapidata_client.referee.base_referee import Referee
 
 
 class NaiveReferee(Referee):
-    """A simple referee that completes a task after a fixed number of guesses.
+    """A simple referee that completes a task after a fixed number of responses.
 
     This referee implements a straightforward approach to task completion,
     where the task is considered finished after a predetermined number of
-    guesses have been made, regardless of the content or quality of those guesses.
+    responses have been made, regardless of the content or quality of those responses.
 
     Attributes:
-        required_guesses (int): The number of guesses required to complete the task.
+        responses (int): The number of responses required to complete the task.
     """
 
-    def __init__(self, required_guesses: int = 10):
+    def __init__(self, responses: int = 10):
         """Initialize the NaiveReferee.
 
         Args:
-            required_guesses (int, optional): The number of guesses required
-                to complete the task. Defaults to 10.
+            responses (int, optional): The number of responses required
+                to complete the task. Defaults to 10. This is per media item.
         """
         super().__init__()
-        self.required_guesses = required_guesses
+        self.responses = responses
 
     def to_dict(self):
         return {
             "_t": "NaiveRefereeConfig",
-            "guessesRequired": self.required_guesses,
+            "guessesRequired": self.responses,
         }
 
     def to_model(self):
-        return NaiveRefereeModel(_t="NaiveReferee", totalVotes=self.required_guesses)
+        return NaiveRefereeModel(_t="NaiveReferee", totalVotes=self.responses)

rapidata/rapidata_client/simple_builders/simple_classification_builders.py

@@ -1,7 +1,7 @@
 from rapidata.rapidata_client.order.rapidata_order_builder import RapidataOrderBuilder
 from rapidata.rapidata_client.metadata.base_metadata import Metadata
 from rapidata.rapidata_client.referee.naive_referee import NaiveReferee
-from rapidata.rapidata_client.referee.classify_early_stopping_referee import ClassifyEarlyStoppingReferee
+from rapidata.rapidata_client.referee.early_stopping_referee import EarlyStoppingReferee
 from rapidata.rapidata_client.selection.base_selection import Selection
 from rapidata.rapidata_client.workflow.classify_workflow import ClassifyWorkflow
 from rapidata.rapidata_client.selection.validation_selection import ValidationSelection
@@ -43,13 +43,13 @@ class ClassificationOrderBuilder:
 
     def create(self, submit: bool = True, max_upload_workers: int = 10):
         if self._probability_threshold and self._responses_required:
-            referee = ClassifyEarlyStoppingReferee(
+            referee = EarlyStoppingReferee(
                 max_vote_count=self._responses_required,
                 threshold=self._probability_threshold
             )
 
         else:
-            referee = NaiveReferee(required_guesses=self._responses_required)
+            referee = NaiveReferee(responses=self._responses_required)
 
         assets = [MediaAsset(path=media_path) for media_path in self._media_paths]
 

rapidata/rapidata_client/simple_builders/simple_compare_builders.py

@@ -2,7 +2,7 @@ from rapidata.service.openapi_service import OpenAPIService
 from rapidata.rapidata_client.metadata import Metadata
 from rapidata.rapidata_client.order.rapidata_order_builder import RapidataOrderBuilder
 from rapidata.rapidata_client.workflow.compare_workflow import CompareWorkflow
-from rapidata.rapidata_client.referee.naive_referee import NaiveReferee
+from rapidata.rapidata_client.referee import NaiveReferee, EarlyStoppingReferee
 from rapidata.rapidata_client.selection.validation_selection import ValidationSelection
 from rapidata.rapidata_client.selection.labeling_selection import LabelingSelection
 from rapidata.rapidata_client.selection.base_selection import Selection
@@ -18,6 +18,7 @@ class CompareOrderBuilder:
         self._responses_required = 10
         self._metadata = None
         self._validation_set_id = None
+        self._probability_threshold = None
 
     def responses(self, responses_required: int) -> 'CompareOrderBuilder':
         """Set the number of resoonses required per matchup/pairing for the comparison order."""
@@ -34,7 +35,20 @@ class CompareOrderBuilder:
         self._validation_set_id = validation_set_id
         return self
     
+    def probability_threshold(self, probability_threshold: float) -> 'CompareOrderBuilder':
+        """Set the probability threshold for early stopping."""
+        self._probability_threshold = probability_threshold
+        return self
+    
     def create(self, submit: bool = True, max_upload_workers: int = 10):
+        if self._probability_threshold and self._responses_required:
+            referee = EarlyStoppingReferee(
+                max_vote_count=self._responses_required,
+                threshold=self._probability_threshold
+            )
+
+        else:
+            referee = NaiveReferee(responses=self._responses_required)
         selection: list[Selection] = ([ValidationSelection(amount=1, validation_set_id=self._validation_set_id), LabelingSelection(amount=2)] 
                      if self._validation_set_id 
                      else [LabelingSelection(amount=3)])
@@ -46,7 +60,7 @@ class CompareOrderBuilder:
                     criteria=self._criteria
                 )
             )
-            .referee(NaiveReferee(required_guesses=self._responses_required))
+            .referee(referee)
             .media(media_paths, metadata=self._metadata) # type: ignore
             .selections(selection)
             .create(submit=submit, max_workers=max_upload_workers))

{rapidata-1.4.3.dist-info → rapidata-1.4.5.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: rapidata
-Version: 1.4.3
+Version: 1.4.5
 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

{rapidata-1.4.3.dist-info → rapidata-1.4.5.dist-info}/RECORD

@@ -1,4 +1,4 @@
-rapidata/__init__.py,sha256=vjmq4p45annpd9K5QbD_0RdcLIkDnNFkNb55ZVws56A,587
+rapidata/__init__.py,sha256=xkV2cgV_RBZCW8CNxWFMlf4EFJUWXcJnqfzQtJsG6NA,579
 rapidata/api_client/__init__.py,sha256=kVvFCI9LJojctHrtdas3VH3EPpbuT8S2OrERxSew5v4,23913
 rapidata/api_client/api/__init__.py,sha256=S0oVoAVMys10M-Z1SqirMdnHMYSHH3Lz6iph1CfILc0,1004
 rapidata/api_client/api/campaign_api.py,sha256=DxPFqt9F6c9OpXu_Uxhsrib2NVwnbcZFa3Vkrj7cIuA,40474
@@ -310,7 +310,7 @@ rapidata/api_client/models/workflow_split_model_filter_configs_inner.py,sha256=1
 rapidata/api_client/models/workflow_state.py,sha256=5LAK1se76RCoozeVB6oxMPb8p_5bhLZJqn7q5fFQWis,850
 rapidata/api_client/rest.py,sha256=zmCIFQC2l1t-KZcq-TgEm3vco3y_LK6vRm3Q07K-xRI,9423
 rapidata/api_client_README.md,sha256=58aoLkfxLpoYpBu7uOQkdfPB5wwrkW1cix_LZ1GLQQQ,37571
-rapidata/rapidata_client/__init__.py,sha256=mX25RNyeNqboIzCMz_Lm3SnMTXM8LJ_jpXuIRraafww,803
+rapidata/rapidata_client/__init__.py,sha256=2SsfOqLpkTh6b0wTKwUaBboxwuU9_SvP4TK-eUoocGo,795
 rapidata/rapidata_client/assets/__init__.py,sha256=T-XKvMSkmyI8iYLUYDdZ3LrrSInHsGMUY_Tz77hhnlE,240
 rapidata/rapidata_client/assets/base_asset.py,sha256=B2YWH1NgaeYUYHDW3OPpHM_bqawHbH4EjnRCE2BYwiM,298
 rapidata/rapidata_client/assets/media_asset.py,sha256=4xU1k2abdHGwbkJAYNOZYyOPB__5VBRDvRjklegFufQ,887
@@ -341,13 +341,13 @@ rapidata/rapidata_client/metadata/prompt_metadata.py,sha256=_FypjKWrC3iKUO_G2CVw
 rapidata/rapidata_client/metadata/public_text_metadata.py,sha256=LTiBQHs6izxQ6-C84d6Pf7lL4ENTDgg__HZnDKvzvMc,511
 rapidata/rapidata_client/metadata/transcription_metadata.py,sha256=THtDEVCON4UlcXHmXrjilaOLHys4TrktUOPGWnXaCcc,631
 rapidata/rapidata_client/order/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-rapidata/rapidata_client/order/rapidata_order.py,sha256=u0VhNgnbT0dn3AJiITqjtqn_OWgdwUcXrgU4iblPXZI,4663
+rapidata/rapidata_client/order/rapidata_order.py,sha256=IMozRlrYK_yqGLk0GQbLNYL_RRb7gUcSgDe6NUjOEEE,4812
 rapidata/rapidata_client/order/rapidata_order_builder.py,sha256=QvICzduLAuAkf8qFKxHV3zAag838WnV9lEzWdD4dxI0,15926
-rapidata/rapidata_client/rapidata_client.py,sha256=UKFDbck3TOcSIRM31gHpl7vpSCHd3jPrjiTODfBPH44,6956
-rapidata/rapidata_client/referee/__init__.py,sha256=Ow9MQsONhF4sX2wFK9jbvSBrpcJgtq3OglIQMkBUdIY,167
+rapidata/rapidata_client/rapidata_client.py,sha256=lXOubU-d1AWX28fhFPdLgIM1p6CCLOGJPJGQNehEqX0,6950
+rapidata/rapidata_client/referee/__init__.py,sha256=E1VODxTjoQRnxzdgMh3aRlDLouxe1nWuvozEHXD2gq4,150
 rapidata/rapidata_client/referee/base_referee.py,sha256=bMy7cw0a-pGNbFu6u_1_Jplu0A483Ubj4oDQzh8vu8k,493
-rapidata/rapidata_client/referee/classify_early_stopping_referee.py,sha256=B5wsqKM3_Oc1TU_MFGiIyiXjwK1LcmaVjhzLdaL8Cgw,1797
-rapidata/rapidata_client/referee/naive_referee.py,sha256=KWMLSc73gOdM8YT_ciFhfN7J4eKgtOFphBG9tIra9g0,1179
+rapidata/rapidata_client/referee/early_stopping_referee.py,sha256=Dg2Kk7OiLBtS3kknsLxyJIlS27xmPvsikFR6g4xlbTE,1862
+rapidata/rapidata_client/referee/naive_referee.py,sha256=uOmrzCfpDopm6ww8u4Am5vz5qGNM7621HNJ-QIeH9wM,1164
 rapidata/rapidata_client/selection/__init__.py,sha256=RFdVUeo2bjCL1gPIL6HXzbpOBqY9kYedkNgb2_ANNLs,321
 rapidata/rapidata_client/selection/base_selection.py,sha256=Y3HkROPm4I4HLNiR0HuHKpvk236KkRlsoDxQATm_chY,138
 rapidata/rapidata_client/selection/capped_selection.py,sha256=ikEIT1sKbwWSK8zVqabT-se5LwsEbFs9dkMbO3u_ERk,807
@@ -356,8 +356,8 @@ rapidata/rapidata_client/selection/demographic_selection.py,sha256=DnMbLhzRItZ0t
 rapidata/rapidata_client/selection/labeling_selection.py,sha256=cqDMQEXfQGMmgIvPgGOYgIGaXflV_J7LZsGOsakLXqo,425
 rapidata/rapidata_client/selection/validation_selection.py,sha256=HswzD2SvZZWisNLoGj--0sT_TIK8crYp3xGGndo6aLY,523
 rapidata/rapidata_client/simple_builders/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-rapidata/rapidata_client/simple_builders/simple_classification_builders.py,sha256=kaa8DNQJ3WOd_LdXItNkENMCID5v5WThbAEQVBNoNyc,5552
-rapidata/rapidata_client/simple_builders/simple_compare_builders.py,sha256=MOb_QwwjaERRks82LemtG46ovoMLNQmNLN_A7JpZK0M,4240
+rapidata/rapidata_client/simple_builders/simple_classification_builders.py,sha256=gTb1OzJZjNW48ltrp0v9tcDjLee9_ZiLw78zYziVu_g,5520
+rapidata/rapidata_client/simple_builders/simple_compare_builders.py,sha256=GmNIUoAkoAHp0OnWLOgeI5-wnDiycuye0o1CUa9c3Q0,4808
 rapidata/rapidata_client/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 rapidata/rapidata_client/utils/utils.py,sha256=Fl99gCnh_HnieIp099xEvEv4g2kEIKiFcUp0G2iz6x8,815
 rapidata/rapidata_client/workflow/__init__.py,sha256=xWuzAhBnbcUFfWcgYrzj8ZYLSOXyFtgfepgMrf0hNhU,290
@@ -372,7 +372,7 @@ rapidata/service/local_file_service.py,sha256=pgorvlWcx52Uh3cEG6VrdMK_t__7dacQ_5
 rapidata/service/openapi_service.py,sha256=pGcOCttKZW0PVCSM7Kfehe5loh7CxmmDDbu4UJbamnI,2770
 rapidata/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 rapidata/utils/image_utils.py,sha256=TldO3eJWG8IhfJjm5MfNGO0mEDm1mQTsRoA0HLU1Uxs,404
-rapidata-1.4.3.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-rapidata-1.4.3.dist-info/METADATA,sha256=AgMjhm057wT3bZqvDtZZstb1jHm8o3OKxMhiiVL18og,1012
-rapidata-1.4.3.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
-rapidata-1.4.3.dist-info/RECORD,,
+rapidata-1.4.5.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+rapidata-1.4.5.dist-info/METADATA,sha256=4tGW6aFOk8Zw8_fi7hFzdIWKoimwxRwy0VLmpVCpSW0,1012
+rapidata-1.4.5.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
+rapidata-1.4.5.dist-info/RECORD,,