arize-phoenix 4.22.0__tar.gz → 4.23.0__tar.gz

{arize_phoenix-4.22.0 → arize_phoenix-4.23.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: arize-phoenix
-Version: 4.22.0
+Version: 4.23.0
 Summary: AI Observability and Evaluation
 Project-URL: Documentation, https://docs.arize.com/phoenix/
 Project-URL: Issues, https://github.com/Arize-ai/phoenix/issues

{arize_phoenix-4.22.0 → arize_phoenix-4.23.0}/examples/manually-instrumented-chatbot/frontend/requirements.txt

@@ -103,7 +103,7 @@ sniffio==1.3.1
     # via
     #   anyio
     #   httpx
-streamlit==1.33.0
+streamlit==1.37.0
 tenacity==8.2.3
     # via streamlit
 toml==0.10.2

{arize_phoenix-4.22.0 → arize_phoenix-4.23.0}/src/phoenix/experiments/functions.py

@@ -447,14 +447,14 @@ def evaluate_experiment(
         )
         if not dataset.examples:
             raise ValueError(f"Dataset has no examples: {dataset_id=}, {dataset_version_id=}")
-        experiment_runs = tuple(
-            ExperimentRun.from_dict(exp_run)
+        experiment_runs = {
+            exp_run["id"]: ExperimentRun.from_dict(exp_run)
             for exp_run in sync_client.get(f"/v1/experiments/{experiment.id}/runs").json()["data"]
-        )
+        }
         if not experiment_runs:
             raise ValueError("Experiment has not been run")
         params = ExperimentParameters(n_examples=len(dataset.examples))
-        task_summary = TaskSummary.from_task_runs(params, experiment_runs)
+        task_summary = TaskSummary.from_task_runs(params, experiment_runs.values())
         ran_experiment = object.__new__(RanExperiment)
         ran_experiment.__init__(  # type: ignore[misc]
             dataset=dataset,

{arize_phoenix-4.22.0 → arize_phoenix-4.23.0}/src/phoenix/experiments/types.py

@@ -167,7 +167,7 @@ class Dataset:
     def from_dict(cls, obj: Mapping[str, Any]) -> Dataset:
         examples = tuple(map(Example.from_dict, obj.get("examples") or ()))
         return cls(
-            id=obj["id"],
+            id=obj["dataset_id"],
             version_id=obj["version_id"],
             examples={ex.id: ex for ex in examples},
         )
@@ -225,7 +225,7 @@ class ExperimentRun:
         )
 
     def __post_init__(self) -> None:
-        if bool(self.output) == bool(self.error):
+        if self.output is None and self.error is None:
             raise ValueError("Must specify exactly one of experiment_run_output or error")
 
 
@@ -284,7 +284,7 @@ class ExperimentEvaluationRun:
         )
 
     def __post_init__(self) -> None:
-        if bool(self.result) == bool(self.error):
+        if self.result is None and self.error is None:
             raise ValueError("Must specify either result or error")
 
 

{arize_phoenix-4.22.0 → arize_phoenix-4.23.0}/src/phoenix/server/api/context.py

@@ -14,7 +14,6 @@ from phoenix.server.api.dataloaders import (
     DocumentEvaluationsDataLoader,
     DocumentEvaluationSummaryDataLoader,
     DocumentRetrievalMetricsDataLoader,
-    EvaluationSummaryDataLoader,
     ExperimentAnnotationSummaryDataLoader,
     ExperimentErrorRatesDataLoader,
     ExperimentRunCountsDataLoader,
@@ -26,10 +25,8 @@ from phoenix.server.api.dataloaders import (
     SpanAnnotationsDataLoader,
     SpanDatasetExamplesDataLoader,
     SpanDescendantsDataLoader,
-    SpanEvaluationsDataLoader,
     SpanProjectsDataLoader,
     TokenCountDataLoader,
-    TraceEvaluationsDataLoader,
     TraceRowIdsDataLoader,
 )
 from phoenix.server.dml_event import DmlEvent
@@ -45,7 +42,6 @@ class DataLoaders:
     document_evaluations: DocumentEvaluationsDataLoader
     document_retrieval_metrics: DocumentRetrievalMetricsDataLoader
     annotation_summaries: AnnotationSummaryDataLoader
-    evaluation_summaries: EvaluationSummaryDataLoader
     experiment_annotation_summaries: ExperimentAnnotationSummaryDataLoader
     experiment_error_rates: ExperimentErrorRatesDataLoader
     experiment_run_counts: ExperimentRunCountsDataLoader
@@ -56,10 +52,8 @@ class DataLoaders:
     span_annotations: SpanAnnotationsDataLoader
     span_dataset_examples: SpanDatasetExamplesDataLoader
     span_descendants: SpanDescendantsDataLoader
-    span_evaluations: SpanEvaluationsDataLoader
     span_projects: SpanProjectsDataLoader
     token_counts: TokenCountDataLoader
-    trace_evaluations: TraceEvaluationsDataLoader
     trace_row_ids: TraceRowIdsDataLoader
     project_by_name: ProjectByNameDataLoader
 

{arize_phoenix-4.22.0 → arize_phoenix-4.23.0}/src/phoenix/server/api/dataloaders/__init__.py

@@ -10,7 +10,6 @@ from .document_evaluation_summaries import (
 )
 from .document_evaluations import DocumentEvaluationsDataLoader
 from .document_retrieval_metrics import DocumentRetrievalMetricsDataLoader
-from .evaluation_summaries import EvaluationSummaryCache, EvaluationSummaryDataLoader
 from .experiment_annotation_summaries import ExperimentAnnotationSummaryDataLoader
 from .experiment_error_rates import ExperimentErrorRatesDataLoader
 from .experiment_run_counts import ExperimentRunCountsDataLoader
@@ -22,10 +21,8 @@ from .record_counts import RecordCountCache, RecordCountDataLoader
 from .span_annotations import SpanAnnotationsDataLoader
 from .span_dataset_examples import SpanDatasetExamplesDataLoader
 from .span_descendants import SpanDescendantsDataLoader
-from .span_evaluations import SpanEvaluationsDataLoader
 from .span_projects import SpanProjectsDataLoader
 from .token_counts import TokenCountCache, TokenCountDataLoader
-from .trace_evaluations import TraceEvaluationsDataLoader
 from .trace_row_ids import TraceRowIdsDataLoader
 
 __all__ = [
@@ -37,7 +34,6 @@ __all__ = [
     "DocumentEvaluationsDataLoader",
     "DocumentRetrievalMetricsDataLoader",
     "AnnotationSummaryDataLoader",
-    "EvaluationSummaryDataLoader",
     "ExperimentAnnotationSummaryDataLoader",
     "ExperimentErrorRatesDataLoader",
     "ExperimentRunCountsDataLoader",
@@ -47,10 +43,8 @@ __all__ = [
     "RecordCountDataLoader",
     "SpanDatasetExamplesDataLoader",
     "SpanDescendantsDataLoader",
-    "SpanEvaluationsDataLoader",
     "SpanProjectsDataLoader",
     "TokenCountDataLoader",
-    "TraceEvaluationsDataLoader",
     "TraceRowIdsDataLoader",
     "ProjectByNameDataLoader",
     "SpanAnnotationsDataLoader",
@@ -65,9 +59,6 @@ class CacheForDataLoaders:
     annotation_summary: AnnotationSummaryCache = field(
         default_factory=AnnotationSummaryCache,
     )
-    evaluation_summary: EvaluationSummaryCache = field(
-        default_factory=EvaluationSummaryCache,
-    )
     latency_ms_quantile: LatencyMsQuantileCache = field(
         default_factory=LatencyMsQuantileCache,
     )

{arize_phoenix-4.22.0 → arize_phoenix-4.23.0}/src/phoenix/server/api/mutations/dataset_mutations.py

@@ -230,7 +230,7 @@ class DatasetMutationMixin:
     ) -> DatasetMutationPayload:
         dataset_id = input.dataset_id
         # Extract the span rowids from the input examples if they exist
-        span_ids = span_ids = [example.span_id for example in input.examples if example.span_id]
+        span_ids = [example.span_id for example in input.examples if example.span_id]
         span_rowids = {
             from_global_id_with_expected_type(global_id=span_id, expected_type_name=Span.__name__)
             for span_id in set(span_ids)
@@ -260,6 +260,8 @@ class DatasetMutationMixin:
                 )
                 .returning(models.DatasetVersion.id)
             )
+
+            # Fetch spans and span annotations
             spans = (
                 await session.execute(
                     select(models.Span.id)
@@ -267,9 +269,36 @@ class DatasetMutationMixin:
                     .where(models.Span.id.in_(span_rowids))
                 )
             ).all()
-            # Just validate that the number of spans matches the number of span_ids
-            # to ensure that the span_ids are valid
-            assert len(spans) == len(span_rowids)
+
+            span_annotations = (
+                await session.execute(
+                    select(
+                        models.SpanAnnotation.span_rowid,
+                        models.SpanAnnotation.name,
+                        models.SpanAnnotation.label,
+                        models.SpanAnnotation.score,
+                        models.SpanAnnotation.explanation,
+                        models.SpanAnnotation.metadata_,
+                        models.SpanAnnotation.annotator_kind,
+                    )
+                    .select_from(models.SpanAnnotation)
+                    .where(models.SpanAnnotation.span_rowid.in_(span_rowids))
+                )
+            ).all()
+
+            span_annotations_by_span: Dict[int, Dict[Any, Any]] = {span.id: {} for span in spans}
+            for annotation in span_annotations:
+                span_id = annotation.span_rowid
+                if span_id not in span_annotations_by_span:
+                    span_annotations_by_span[span_id] = dict()
+                span_annotations_by_span[span_id][annotation.name] = {
+                    "label": annotation.label,
+                    "score": annotation.score,
+                    "explanation": annotation.explanation,
+                    "metadata": annotation.metadata_,
+                    "annotator_kind": annotation.annotator_kind,
+                }
+
             DatasetExample = models.DatasetExample
             dataset_example_rowids = (
                 await session.scalars(
@@ -291,21 +320,32 @@ class DatasetMutationMixin:
             assert len(dataset_example_rowids) == len(input.examples)
             assert all(map(lambda id: isinstance(id, int), dataset_example_rowids))
             DatasetExampleRevision = models.DatasetExampleRevision
-            await session.execute(
-                insert(DatasetExampleRevision),
-                [
+
+            dataset_example_revisions = []
+            for dataset_example_rowid, example in zip(dataset_example_rowids, input.examples):
+                span_annotation = {}
+                if example.span_id:
+                    span_id = from_global_id_with_expected_type(
+                        global_id=example.span_id,
+                        expected_type_name=Span.__name__,
+                    )
+                    span_annotation = span_annotations_by_span.get(span_id, {})
+                dataset_example_revisions.append(
                     {
                         DatasetExampleRevision.dataset_example_id.key: dataset_example_rowid,
                         DatasetExampleRevision.dataset_version_id.key: dataset_version_rowid,
                         DatasetExampleRevision.input.key: example.input,
                         DatasetExampleRevision.output.key: example.output,
-                        DatasetExampleRevision.metadata_.key: example.metadata,
+                        DatasetExampleRevision.metadata_.key: {
+                            **(example.metadata or {}),
+                            "annotations": span_annotation,
+                        },
                         DatasetExampleRevision.revision_kind.key: "CREATE",
                     }
-                    for dataset_example_rowid, example in zip(
-                        dataset_example_rowids, input.examples
-                    )
-                ],
+                )
+            await session.execute(
+                insert(DatasetExampleRevision),
+                dataset_example_revisions,
             )
         info.context.event_queue.put(DatasetInsertEvent((dataset.id,)))
         return DatasetMutationPayload(dataset=to_gql_dataset(dataset))

{arize_phoenix-4.22.0 → arize_phoenix-4.23.0}/src/phoenix/server/api/queries.py

@@ -67,10 +67,40 @@ from phoenix.server.api.types.Project import Project
 from phoenix.server.api.types.SortDir import SortDir
 from phoenix.server.api.types.Span import Span, to_gql_span
 from phoenix.server.api.types.Trace import Trace
+from phoenix.server.api.types.User import User
 
 
 @strawberry.type
 class Query:
+    @strawberry.field
+    async def users(
+        self,
+        info: Info[Context, None],
+        first: Optional[int] = 50,
+        last: Optional[int] = UNSET,
+        after: Optional[CursorString] = UNSET,
+        before: Optional[CursorString] = UNSET,
+    ) -> Connection[User]:
+        args = ConnectionArgs(
+            first=first,
+            after=after if isinstance(after, CursorString) else None,
+            last=last,
+            before=before if isinstance(before, CursorString) else None,
+        )
+        stmt = select(models.User).order_by(models.User.email)
+        async with info.context.db() as session:
+            users = await session.stream_scalars(stmt)
+            data = [
+                User(
+                    id_attr=user.id,
+                    email=user.email,
+                    username=user.username,
+                    created_at=user.created_at,
+                )
+                async for user in users
+            ]
+        return connection_from_list(data=data, args=args)
+
     @strawberry.field
     async def projects(
         self,

{arize_phoenix-4.22.0 → arize_phoenix-4.23.0}/src/phoenix/server/api/routers/v1/experiment_runs.py

@@ -108,7 +108,12 @@ async def create_experiment_run(
     return CreateExperimentResponseBody(data=CreateExperimentRunResponseBodyData(id=str(run_gid)))
 
 
-class ListExperimentRunsResponseBody(ResponseBody[List[ExperimentRun]]):
+class ExperimentRunResponse(ExperimentRun):
+    id: str = Field(description="The ID of the experiment run")
+    experiment_id: str = Field(description="The ID of the experiment")
+
+
+class ListExperimentRunsResponseBody(ResponseBody[List[ExperimentRunResponse]]):
     pass
 
 
@@ -147,7 +152,7 @@ async def list_experiment_runs(
             experiment_gid = GlobalID("Experiment", str(exp_run.experiment_id))
             example_gid = GlobalID("DatasetExample", str(exp_run.dataset_example_id))
             runs.append(
-                ExperimentRun(
+                ExperimentRunResponse(
                     start_time=exp_run.start_time,
                     end_time=exp_run.end_time,
                     experiment_id=str(experiment_gid),

{arize_phoenix-4.22.0 → arize_phoenix-4.23.0}/src/phoenix/server/api/types/Evaluation.py

@@ -1,7 +1,7 @@
 import strawberry
 
 import phoenix.trace.v1 as pb
-from phoenix.db.models import DocumentAnnotation, SpanAnnotation, TraceAnnotation
+from phoenix.db.models import DocumentAnnotation, TraceAnnotation
 
 from .Annotation import Annotation
 
@@ -31,31 +31,6 @@ class TraceEvaluation(Annotation):
         )
 
 
-@strawberry.type
-class SpanEvaluation(Annotation):
-    @staticmethod
-    def from_pb_evaluation(evaluation: pb.Evaluation) -> "SpanEvaluation":
-        result = evaluation.result
-        score = result.score.value if result.HasField("score") else None
-        label = result.label.value if result.HasField("label") else None
-        explanation = result.explanation.value if result.HasField("explanation") else None
-        return SpanEvaluation(
-            name=evaluation.name,
-            score=score,
-            label=label,
-            explanation=explanation,
-        )
-
-    @staticmethod
-    def from_sql_span_annotation(annotation: SpanAnnotation) -> "SpanEvaluation":
-        return SpanEvaluation(
-            name=annotation.name,
-            score=annotation.score,
-            label=annotation.label,
-            explanation=annotation.explanation,
-        )
-
-
 @strawberry.type
 class DocumentEvaluation(Annotation):
     document_position: int = strawberry.field(

{arize_phoenix-4.22.0 → arize_phoenix-4.23.0}/src/phoenix/server/api/types/Project.py

@@ -24,7 +24,6 @@ from phoenix.server.api.input_types.SpanSort import SpanSort, SpanSortConfig
 from phoenix.server.api.input_types.TimeRange import TimeRange
 from phoenix.server.api.types.AnnotationSummary import AnnotationSummary
 from phoenix.server.api.types.DocumentEvaluationSummary import DocumentEvaluationSummary
-from phoenix.server.api.types.EvaluationSummary import EvaluationSummary
 from phoenix.server.api.types.pagination import (
     Cursor,
     CursorSortColumn,
@@ -249,23 +248,6 @@ class Project(Node):
             has_next_page=has_next_page,
         )
 
-    @strawberry.field(
-        description="Names of all available evaluations for traces. "
-        "(The list contains no duplicates.)"
-    )  # type: ignore
-    async def trace_evaluation_names(
-        self,
-        info: Info[Context, None],
-    ) -> List[str]:
-        stmt = (
-            select(distinct(models.TraceAnnotation.name))
-            .join(models.Trace)
-            .where(models.Trace.project_rowid == self.id_attr)
-            .where(models.TraceAnnotation.annotator_kind == "LLM")
-        )
-        async with info.context.db() as session:
-            return list(await session.scalars(stmt))
-
     @strawberry.field(
         description="Names of all available annotations for traces. "
         "(The list contains no duplicates.)"
@@ -282,24 +264,6 @@ class Project(Node):
         async with info.context.db() as session:
             return list(await session.scalars(stmt))
 
-    @strawberry.field(
-        description="Names of all available evaluations for spans. "
-        "(The list contains no duplicates.)"
-    )  # type: ignore
-    async def span_evaluation_names(
-        self,
-        info: Info[Context, None],
-    ) -> List[str]:
-        stmt = (
-            select(distinct(models.SpanAnnotation.name))
-            .join(models.Span)
-            .join(models.Trace, models.Span.trace_rowid == models.Trace.id)
-            .where(models.Trace.project_rowid == self.id_attr)
-            .where(models.SpanAnnotation.annotator_kind == "LLM")
-        )
-        async with info.context.db() as session:
-            return list(await session.scalars(stmt))
-
     @strawberry.field(
         description="Names of all available annotations for spans. "
         "(The list contains no duplicates.)"
@@ -337,29 +301,6 @@ class Project(Node):
         async with info.context.db() as session:
             return list(await session.scalars(stmt))
 
-    @strawberry.field
-    async def trace_evaluation_summary(
-        self,
-        info: Info[Context, None],
-        evaluation_name: str,
-        time_range: Optional[TimeRange] = UNSET,
-    ) -> Optional[EvaluationSummary]:
-        return await info.context.data_loaders.evaluation_summaries.load(
-            ("trace", self.id_attr, time_range, None, evaluation_name),
-        )
-
-    @strawberry.field
-    async def span_evaluation_summary(
-        self,
-        info: Info[Context, None],
-        evaluation_name: str,
-        time_range: Optional[TimeRange] = UNSET,
-        filter_condition: Optional[str] = UNSET,
-    ) -> Optional[EvaluationSummary]:
-        return await info.context.data_loaders.evaluation_summaries.load(
-            ("span", self.id_attr, time_range, filter_condition, evaluation_name),
-        )
-
     @strawberry.field
     async def trace_annotation_summary(
         self,
@@ -405,7 +346,7 @@ class Project(Node):
     @strawberry.field
     async def validate_span_filter_condition(self, condition: str) -> ValidationResult:
         # This query is too expensive to run on every validation
-        # valid_eval_names = await self.span_evaluation_names()
+        # valid_eval_names = await self.span_annotation_names()
         try:
             SpanFilter(
                 condition=condition,

{arize_phoenix-4.22.0 → arize_phoenix-4.23.0}/src/phoenix/server/api/types/Span.py

@@ -28,7 +28,7 @@ from phoenix.server.api.types.SpanAnnotation import to_gql_span_annotation
 from phoenix.trace.attributes import get_attribute_value
 
 from .DocumentRetrievalMetrics import DocumentRetrievalMetrics
-from .Evaluation import DocumentEvaluation, SpanEvaluation
+from .Evaluation import DocumentEvaluation
 from .ExampleRevisionInterface import ExampleRevision
 from .MimeType import MimeType
 from .SpanAnnotation import SpanAnnotation
@@ -170,14 +170,6 @@ class Span(Node):
         "codes from descendant spans (children, grandchildren, etc.)",
     )
 
-    @strawberry.field(
-        description="Evaluations associated with the span, e.g. if the span is "
-        "an LLM, an evaluation may assess the helpfulness of its response with "
-        "respect to its input."
-    )  # type: ignore
-    async def span_evaluations(self, info: Info[Context, None]) -> List[SpanEvaluation]:
-        return await info.context.data_loaders.span_evaluations.load(self.id_attr)
-
     @strawberry.field(
         description=(
             "Annotations associated with the span. This encompasses both "
@@ -240,7 +232,7 @@ class Span(Node):
     @strawberry.field(
         description="The span's attributes translated into an example revision for a dataset",
     )  # type: ignore
-    def as_example_revision(self) -> SpanAsExampleRevision:
+    async def as_example_revision(self, info: Info[Context, None]) -> SpanAsExampleRevision:
         db_span = self.db_span
         attributes = db_span.attributes
         span_io = _SpanIO(
@@ -256,10 +248,28 @@ class Span(Node):
             llm_output_messages=get_attribute_value(attributes, LLM_OUTPUT_MESSAGES),
             retrieval_documents=get_attribute_value(attributes, RETRIEVAL_DOCUMENTS),
         )
+
+        # Fetch annotations associated with this span
+        span_annotations = await self.span_annotations(info)
+        annotations = dict()
+        for annotation in span_annotations:
+            annotations[annotation.name] = {
+                "label": annotation.label,
+                "score": annotation.score,
+                "explanation": annotation.explanation,
+                "metadata": annotation.metadata,
+                "annotator_kind": annotation.annotator_kind.value,
+            }
+        # Merge annotations into the metadata
+        metadata = {
+            **attributes,
+            "annotations": annotations,
+        }
+
         return SpanAsExampleRevision(
             input=get_dataset_example_input(span_io),
             output=get_dataset_example_output(span_io),
-            metadata=attributes,
+            metadata=metadata,
         )
 
     @strawberry.field(description="The project that this span belongs to.")  # type: ignore

arize_phoenix-4.23.0/src/phoenix/server/api/types/User.py

@@ -0,0 +1,13 @@
+from datetime import datetime
+from typing import Optional
+
+import strawberry
+from strawberry.relay import Node, NodeID
+
+
+@strawberry.type
+class User(Node):
+    id_attr: NodeID[int]
+    email: str
+    username: Optional[str]
+    created_at: datetime

{arize_phoenix-4.22.0 → arize_phoenix-4.23.0}/src/phoenix/server/app.py

@@ -65,7 +65,6 @@ from phoenix.server.api.dataloaders import (
     DocumentEvaluationsDataLoader,
     DocumentEvaluationSummaryDataLoader,
     DocumentRetrievalMetricsDataLoader,
-    EvaluationSummaryDataLoader,
     ExperimentAnnotationSummaryDataLoader,
     ExperimentErrorRatesDataLoader,
     ExperimentRunCountsDataLoader,
@@ -77,10 +76,8 @@ from phoenix.server.api.dataloaders import (
     SpanAnnotationsDataLoader,
     SpanDatasetExamplesDataLoader,
     SpanDescendantsDataLoader,
-    SpanEvaluationsDataLoader,
     SpanProjectsDataLoader,
     TokenCountDataLoader,
-    TraceEvaluationsDataLoader,
     TraceRowIdsDataLoader,
 )
 from phoenix.server.api.routers.v1 import REST_API_VERSION
@@ -118,6 +115,8 @@ class AppConfig(NamedTuple):
     n_samples: int
     is_development: bool
     web_manifest_path: Path
+    authentication_enabled: bool
+    """ Whether authentication is enabled """
 
 
 class Static(StaticFiles):
@@ -165,6 +164,7 @@ class Static(StaticFiles):
                     "request": request,
                     "is_development": self._app_config.is_development,
                     "manifest": self._web_manifest,
+                    "authentication_enabled": self._app_config.authentication_enabled,
                 },
             )
         except Exception as e:
@@ -303,12 +303,6 @@ def create_graphql_router(
                     if cache_for_dataloaders
                     else None,
                 ),
-                evaluation_summaries=EvaluationSummaryDataLoader(
-                    db,
-                    cache_map=cache_for_dataloaders.evaluation_summary
-                    if cache_for_dataloaders
-                    else None,
-                ),
                 experiment_annotation_summaries=ExperimentAnnotationSummaryDataLoader(db),
                 experiment_error_rates=ExperimentErrorRatesDataLoader(db),
                 experiment_run_counts=ExperimentRunCountsDataLoader(db),
@@ -332,13 +326,11 @@ def create_graphql_router(
                 span_annotations=SpanAnnotationsDataLoader(db),
                 span_dataset_examples=SpanDatasetExamplesDataLoader(db),
                 span_descendants=SpanDescendantsDataLoader(db),
-                span_evaluations=SpanEvaluationsDataLoader(db),
                 span_projects=SpanProjectsDataLoader(db),
                 token_counts=TokenCountDataLoader(
                     db,
                     cache_map=cache_for_dataloaders.token_count if cache_for_dataloaders else None,
                 ),
-                trace_evaluations=TraceEvaluationsDataLoader(db),
                 trace_row_ids=TraceRowIdsDataLoader(db),
                 project_by_name=ProjectByNameDataLoader(db),
             ),
@@ -406,6 +398,7 @@ def create_app(
     db: DbSessionFactory,
     export_path: Path,
     model: Model,
+    authentication_enabled: bool,
     umap_params: UMAPParameters,
     corpus: Optional[Model] = None,
     debug: bool = False,
@@ -526,6 +519,7 @@ def create_app(
                     n_neighbors=umap_params.n_neighbors,
                     n_samples=umap_params.n_samples,
                     is_development=dev,
+                    authentication_enabled=authentication_enabled,
                     web_manifest_path=SERVER_DIR / "static" / ".vite" / "manifest.json",
                 ),
             ),

{arize_phoenix-4.22.0 → arize_phoenix-4.23.0}/src/phoenix/server/dml_event_handler.py

@@ -142,7 +142,6 @@ class _SpanDeleteEventHandler(_SpanDmlEventHandler):
     @staticmethod
     def _clear(cache: CacheForDataLoaders, project_id: int) -> None:
         cache.annotation_summary.invalidate_project(project_id)
-        cache.evaluation_summary.invalidate_project(project_id)
         cache.document_evaluation_summary.invalidate_project(project_id)
 
 
@@ -201,7 +200,6 @@ class _SpanAnnotationDmlEventHandler(_AnnotationDmlEventHandler[SpanAnnotationDm
     @staticmethod
     def _clear(cache: CacheForDataLoaders, project_id: int, name: str) -> None:
         cache.annotation_summary.invalidate((project_id, name, "span"))
-        cache.evaluation_summary.invalidate((project_id, name, "span"))
 
 
 class _TraceAnnotationDmlEventHandler(_AnnotationDmlEventHandler[TraceAnnotationDmlEvent]):
@@ -214,7 +212,6 @@ class _TraceAnnotationDmlEventHandler(_AnnotationDmlEventHandler[TraceAnnotation
     @staticmethod
     def _clear(cache: CacheForDataLoaders, project_id: int, name: str) -> None:
         cache.annotation_summary.invalidate((project_id, name, "trace"))
-        cache.evaluation_summary.invalidate((project_id, name, "trace"))
 
 
 class _DocumentAnnotationDmlEventHandler(_AnnotationDmlEventHandler[DocumentAnnotationDmlEvent]):

{arize_phoenix-4.22.0 → arize_phoenix-4.23.0}/src/phoenix/server/main.py

@@ -13,6 +13,7 @@ from uvicorn import Config, Server
 import phoenix.trace.v1 as pb
 from phoenix.config import (
     EXPORT_DIR,
+    get_auth_settings,
     get_env_database_connection_str,
     get_env_enable_prometheus,
     get_env_grpc_port,
@@ -80,6 +81,11 @@ _WELCOME_MESSAGE = """
 |  Storage: {storage}
 """
 
+_EXPERIMENTAL_WARNING = """
+🚨 WARNING: Phoenix is running in experimental mode. 🚨
+|  Authentication enabled: {auth_enabled}
+"""
+
 
 def _write_pid_file_when_ready(
     server: Server,
@@ -212,6 +218,8 @@ if __name__ == "__main__":
         reference_inferences,
     )
 
+    authentication_enabled, auth_secret = get_auth_settings()
+
     fixture_spans: List[Span] = []
     fixture_evals: List[pb.Evaluation] = []
     if trace_dataset_name is not None:
@@ -251,6 +259,7 @@ if __name__ == "__main__":
         db=factory,
         export_path=export_path,
         model=model,
+        authentication_enabled=authentication_enabled,
         umap_params=umap_params,
         corpus=None
         if corpus_inferences is None
@@ -278,5 +287,8 @@ if __name__ == "__main__":
         )
     )
 
+    if authentication_enabled:
+        print(_EXPERIMENTAL_WARNING.format(auth_enabled=authentication_enabled))
+
     # Start the server
     server.run()