IL-407 raise error on non existent datasets

src/intelligence_layer/evaluation/dataset/in_memory_dataset_repository.py

@@ -63,7 +63,7 @@ def examples(
         input_type: type[Input],
         expected_output_type: type[ExpectedOutput],
     ) -> Iterable[Example[Input, ExpectedOutput]]:
-        if dataset_id not in self._datasets_and_examples:
+        if dataset_id not in self._datasets_and_examples.keys():
             raise ValueError(
                 f"Repository does not contain a dataset with id: {dataset_id}"
             )

src/intelligence_layer/evaluation/run/in_memory_run_repository.py

@@ -67,6 +67,9 @@ def example_tracer(self, run_id: str, example_id: str) -> Tracer:
     def example_outputs(
         self, run_id: str, output_type: type[Output]
     ) -> Iterable[ExampleOutput[Output]]:
+        if run_id not in self._run_overviews.keys():
+            raise ValueError(f"Repository does not contain a run with id: {run_id}")
+
         return (
             cast(ExampleOutput[Output], example_output)
             for example_output in sorted(

tests/connectors/argilla/test_argilla_client.py

@@ -2,9 +2,11 @@
 from typing import Callable, Iterable, Sequence, TypeVar
 from uuid import uuid4
 
+import pytest
 from dotenv import load_dotenv
 from pydantic import BaseModel
 from pytest import fixture
+from requests import HTTPError
 
 from intelligence_layer.connectors.argilla.argilla_client import (
     ArgillaClient,
@@ -110,6 +112,13 @@ def long_qa_records(
     return records
 
 
+def test_error_on_non_existent_dataset(
+    argilla_client: DefaultArgillaClient,
+) -> None:
+    with pytest.raises(HTTPError):
+        list(argilla_client.records("non_existent_dataset_id"))
+
+
 def test_records_returns_records_previously_added(
     argilla_client: DefaultArgillaClient,
     qa_dataset_id: str,

tests/evaluation/test_hugging_face_dataset_repository.py

@@ -108,12 +108,10 @@ def test_examples_returns_an_empty_list_for_not_existing_dataset_id(
 def test_example_returns_none_for_not_existing_ids(
     hugging_face_dataset_repository: HuggingFaceDatasetRepository,
 ) -> None:
-    assert (
+    with pytest.raises(ValueError):
         hugging_face_dataset_repository.example(
             "not-existing-dataset-id", "not-existing-example-id", str, str
         )
-        is None
-    )
 
 
 def test_delete_dataset_does_not_fail_for_not_existing_dataset_id(
@@ -165,7 +163,8 @@ def test_hugging_face_repository_supports_all_operations_for_created_dataset(
 
     # deleting an existing dataset works
     hugging_face_repository_.delete_dataset(dataset.id)
-    assert hugging_face_repository_.examples(dataset.id, str, str) == []
+    with pytest.raises(ValueError):
+        hugging_face_repository_.examples(dataset.id, str, str)
     assert hugging_face_repository_.dataset(dataset.id) is None