Jacques/method list datasets

sdks/python/src/opik/api_objects/dataset/rest_operations.py

@@ -0,0 +1,87 @@
+from typing import List
+from opik.rest_api import OpikApi
+from opik import exceptions
+from . import dataset
+from .. import experiment
+from ...rest_api.core.api_error import ApiError
+
+
+def get_datasets(
+    rest_client: OpikApi, max_results: int = 1000, sync_items: bool = True
+) -> List[dataset.Dataset]:
+    page_size = 100
+    datasets: List[dataset.Dataset] = []
+
+    page = 1
+    while len(datasets) < max_results:
+        page_datasets = rest_client.datasets.find_datasets(
+            page=page,
+            size=page_size,
+        )
+
+        if len(page_datasets.content) == 0:
+            break
+
+        for dataset_fern in page_datasets.content[: (max_results - len(datasets))]:
+            dataset_ = dataset.Dataset(
+                name=dataset_fern.name,
+                description=dataset_fern.description,
+                rest_client=rest_client,
+            )
+
+            if sync_items:
+                dataset_.__internal_api__sync_hashes__()
+
+            datasets.append(dataset_)
+
+        page += 1
+
+    return datasets
+
+
+def get_dataset_id(rest_client: OpikApi, dataset_name: str) -> str:
+    try:
+        dataset_id = rest_client.datasets.get_dataset_by_identifier(
+            dataset_name=dataset_name
+        ).id
+    except ApiError as e:
+        if e.status_code == 404:
+            raise exceptions.DatasetNotFound(
+                f"Dataset with the name {dataset_name} not found."
+            ) from e
+        raise
+
+    return dataset_id
+
+
+def get_dataset_experiments(
+    rest_client: OpikApi, dataset_id: str, max_results: int = 1000
+) -> List[experiment.Experiment]:
+    page_size = 100
+    experiments: List[experiment.Experiment] = []
+
+    page = 1
+    while len(experiments) < max_results:
+        page_experiments = rest_client.experiments.find_experiments(
+            page=page,
+            size=page_size,
+            dataset_id=dataset_id,
+        )
+
+        if len(page_experiments.content) == 0:
+            break
+
+        for experiment_ in page_experiments.content[: max_results - len(experiments)]:
+            experiments.append(
+                experiment.Experiment(
+                    id=experiment_.id,
+                    name=experiment_.name,
+                    dataset_name=experiment_.dataset_name,
+                    rest_client=rest_client,
+                    # TODO: add prompt if exists
+                )
+            )
+
+        page += 1
+
+    return experiments

sdks/python/src/opik/api_objects/experiment/experiment_item.py

@@ -25,15 +25,19 @@ def from_rest_experiment_item_compare(
         cls,
         value: experiment_item_compare.ExperimentItemCompare,
     ) -> "ExperimentItemContent":
-        feedback_scores: List[FeedbackScoreDict] = [
-            {
-                "category_name": rest_feedback_score.category_name,
-                "name": rest_feedback_score.name,
-                "reason": rest_feedback_score.reason,
-                "value": rest_feedback_score.value,
-            }
-            for rest_feedback_score in value.feedback_scores
-        ]
+        if value.feedback_scores is None:
+            feedback_scores: List[FeedbackScoreDict] = []
+        else:
+            feedback_scores = [
+                {
+                    "category_name": rest_feedback_score.category_name,
+                    "name": rest_feedback_score.name,
+                    "reason": rest_feedback_score.reason,
+                    "value": rest_feedback_score.value,
+                }
+                for rest_feedback_score in value.feedback_scores
+            ]
+
         return ExperimentItemContent(
             id=value.id,
             trace_id=value.trace_id,

sdks/python/src/opik/api_objects/experiment/helpers.py

@@ -1,9 +1,6 @@
 import logging
+from opik import jsonable_encoder
 from typing import Any, Dict, List, Mapping, Optional, Tuple
-
-from opik import exceptions, jsonable_encoder
-from opik.rest_api import OpikApi
-from opik.rest_api.types import experiment_public
 from .. import prompt
 
 LOGGER = logging.getLogger(__name__)
@@ -49,24 +46,6 @@ def build_metadata_and_prompt_versions(
     return metadata, prompt_versions
 
 
-def get_experiment_data_by_name(
-    rest_client: OpikApi, name: str
-) -> experiment_public.ExperimentPublic:
-    page = 0
-
-    while True:
-        page += 1
-        experiment_page_public = rest_client.experiments.find_experiments(name=name)
-        if len(experiment_page_public.content) == 0:
-            raise exceptions.ExperimentNotFound(
-                f"Experiment with the name {name} not found."
-            )
-
-        for experiment in experiment_page_public.content:
-            if experiment.name == name:
-                return experiment
-
-
 def handle_prompt_args(
     prompt: Optional[prompt.Prompt] = None,
     prompts: Optional[List[prompt.Prompt]] = None,

sdks/python/src/opik/api_objects/experiment/rest_operations.py

@@ -0,0 +1,21 @@
+from opik.rest_api import OpikApi
+from opik.rest_api.types import experiment_public
+from opik import exceptions
+
+
+def get_experiment_data_by_name(
+    rest_client: OpikApi, name: str
+) -> experiment_public.ExperimentPublic:
+    page = 0
+
+    while True:
+        page += 1
+        experiment_page_public = rest_client.experiments.find_experiments(name=name)
+        if len(experiment_page_public.content) == 0:
+            raise exceptions.ExperimentNotFound(
+                f"Experiment with the name {name} not found."
+            )
+
+        for experiment in experiment_page_public.content:
+            if experiment.name == name:
+                return experiment

sdks/python/src/opik/api_objects/opik_client.py

@@ -20,6 +20,8 @@
     validation_helpers,
 )
 from .experiment import helpers as experiment_helpers
+from .experiment import rest_operations as experiment_rest_operations
+from .dataset import rest_operations as dataset_rest_operations
 from ..message_processing import streamer_constructors, messages
 from ..message_processing.batching import sequence_splitter
 
@@ -437,6 +439,52 @@ def get_dataset(self, name: str) -> dataset.Dataset:
 
         return dataset_
 
+    def get_datasets(
+        self,
+        max_results: int = 100,
+        sync_items: bool = True,
+    ) -> List[dataset.Dataset]:
+        """
+        Returns all datasets up to the specified limit.
+
+        Args:
+            max_results: The maximum number of datasets to return.
+            sync_items: Whether to sync the hashes of the dataset items. This is used to deduplicate items when fetching the dataset but it can be an expensive operation.
+
+        Returns:
+            List[dataset.Dataset]: A list of dataset objects that match the filter string.
+        """
+        datasets = dataset_rest_operations.get_datasets(
+            self._rest_client, max_results, sync_items
+        )
+
+        return datasets
+
+    def get_dataset_experiments(
+        self,
+        dataset_name: str,
+        max_results: int = 100,
+    ) -> List[experiment.Experiment]:
+        """
+        Returns all experiments up to the specified limit.
+
+        Args:
+            dataset_name: The name of the dataset
+            max_results: The maximum number of experiments to return.
+
+        Returns:
+            List[experiment.Experiment]: A list of experiment objects.
+        """
+        dataset_id = dataset_rest_operations.get_dataset_id(
+            self._rest_client, dataset_name
+        )
+
+        experiments = dataset_rest_operations.get_dataset_experiments(
+            self._rest_client, dataset_id, max_results
+        )
+
+        return experiments
+
     def delete_dataset(self, name: str) -> None:
         """
         Delete dataset by name
@@ -552,7 +600,7 @@ def get_experiment_by_name(self, name: str) -> experiment.Experiment:
         Returns:
             experiment.Experiment: the API object for an existing experiment.
         """
-        experiment_public = experiment_helpers.get_experiment_data_by_name(
+        experiment_public = experiment_rest_operations.get_experiment_data_by_name(
             rest_client=self._rest_client, name=name
         )
 

sdks/python/src/opik/exceptions.py

@@ -40,3 +40,7 @@ def __str__(self) -> str:
 
 class ExperimentNotFound(OpikException):
     pass
+
+
+class DatasetNotFound(OpikException):
+    pass

sdks/python/tests/e2e/test_experiment.py

@@ -531,3 +531,38 @@ def test_experiment__get_experiment_by_name__experiment_not_found__ExperimentNot
 ):
     with pytest.raises(exceptions.ExperimentNotFound):
         opik_client.get_experiment_by_id("not-existing-name")
+
+
+def test_experiment__get_experiment_items__no_feedback_scores(
+    opik_client: opik.Opik, dataset_name: str, experiment_name: str
+):
+    dataset = opik_client.create_dataset(dataset_name)
+
+    dataset.insert(
+        [
+            {
+                "input": {"question": "What is the of capital of France?"},
+                "expected_model_output": {"output": "Paris"},
+            },
+        ]
+    )
+
+    def task(item: Dict[str, Any]) -> Dict[str, Any]:
+        return {
+            "output": "Paris",
+        }
+
+    opik.evaluate(
+        dataset=dataset,
+        task=task,
+        scoring_metrics=[],
+        experiment_name=experiment_name,
+    )
+
+    opik.flush_tracker()
+
+    experiment = opik_client.get_experiment_by_name(experiment_name)
+    items = experiment.get_items()
+
+    assert len(items) == 1
+    assert items[0].feedback_scores == []