arize-phoenix 4.22.1__tar.gz → 4.24.0__tar.gz

{arize_phoenix-4.22.1 → arize_phoenix-4.24.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: arize-phoenix
-Version: 4.22.1
+Version: 4.24.0
 Summary: AI Observability and Evaluation
 Project-URL: Documentation, https://docs.arize.com/phoenix/
 Project-URL: Issues, https://github.com/Arize-ai/phoenix/issues
@@ -80,7 +80,7 @@ Requires-Dist: jupyter; extra == 'dev'
 Requires-Dist: langchain>=0.0.334; extra == 'dev'
 Requires-Dist: litellm>=1.0.3; extra == 'dev'
 Requires-Dist: llama-index>=0.10.3; extra == 'dev'
-Requires-Dist: mypy==1.11.0; extra == 'dev'
+Requires-Dist: mypy==1.11.1; extra == 'dev'
 Requires-Dist: nbqa; extra == 'dev'
 Requires-Dist: pandas-stubs==2.0.3.230814; (python_version < '3.9') and extra == 'dev'
 Requires-Dist: pandas-stubs==2.2.2.240603; (python_version >= '3.9') and extra == 'dev'

{arize_phoenix-4.22.1 → arize_phoenix-4.24.0}/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-4.22.1 → arize_phoenix-4.24.0}/pyproject.toml

@@ -71,7 +71,7 @@ dev = [
   "jupyter",
   "nbqa",
   "ruff==0.5.4",
-  "mypy==1.11.0",
+  "mypy==1.11.1",
   "pandas>=1.0",
   "tabulate",  # used by DataFrame.to_markdown()
   "types-tabulate",

{arize_phoenix-4.22.1 → arize_phoenix-4.24.0}/src/phoenix/db/migrations/future_versions/cd164e83824f_users_and_tokens.py

@@ -150,7 +150,7 @@ async def insert_roles_and_users(connection: AsyncConnection) -> None:
     user to the `users` table.
     """
     await connection.execute(
-        insert(UserRole).values([{"role": "SYSTEM"}, {"role": "ADMIN"}, {"role": "GENERAL"}])
+        insert(UserRole).values([{"role": "SYSTEM"}, {"role": "ADMIN"}, {"role": "MEMBER"}])
     )
     system_user_role_id = sa.select(UserRole.id).where(UserRole.role == "SYSTEM").scalar_subquery()
     admin_user_role_id = sa.select(UserRole.id).where(UserRole.role == "ADMIN").scalar_subquery()

{arize_phoenix-4.22.1 → arize_phoenix-4.24.0}/src/phoenix/db/models.py

@@ -626,6 +626,7 @@ if ENABLE_AUTH:
         __tablename__ = "user_roles"
         id: Mapped[int] = mapped_column(primary_key=True)
         role: Mapped[str] = mapped_column(unique=True)
+        users: Mapped[List["User"]] = relationship("User", back_populates="role")
 
     class User(Base):
         __tablename__ = "users"
@@ -634,6 +635,7 @@ if ENABLE_AUTH:
             ForeignKey("user_roles.id"),
             index=True,
         )
+        role: Mapped["UserRole"] = relationship("UserRole", back_populates="users")
         username: Mapped[Optional[str]] = mapped_column(nullable=True, unique=True, index=True)
         email: Mapped[str] = mapped_column(nullable=False, unique=True, index=True)
         auth_method: Mapped[str] = mapped_column(
@@ -646,6 +648,7 @@ if ENABLE_AUTH:
             UtcTimeStamp, server_default=func.now(), onupdate=func.now()
         )
         deleted_at: Mapped[Optional[datetime]] = mapped_column(UtcTimeStamp)
+        api_keys: Mapped[List["APIKey"]] = relationship("APIKey", back_populates="user")
 
     class APIKey(Base):
         __tablename__ = "api_keys"
@@ -654,6 +657,7 @@ if ENABLE_AUTH:
             ForeignKey("users.id"),
             index=True,
         )
+        user: Mapped["User"] = relationship("User", back_populates="api_keys")
         name: Mapped[str]
         description: Mapped[Optional[str]]
         created_at: Mapped[datetime] = mapped_column(UtcTimeStamp, server_default=func.now())

{arize_phoenix-4.22.1 → arize_phoenix-4.24.0}/src/phoenix/experiments/functions.py

@@ -108,6 +108,7 @@ def run_experiment(
     rate_limit_errors: Optional[RateLimitErrors] = None,
     dry_run: Union[bool, int] = False,
     print_summary: bool = True,
+    concurrency: int = 3,
 ) -> RanExperiment:
     """
     Runs an experiment using a given set of dataset of examples.
@@ -158,6 +159,9 @@ def run_experiment(
             examples of the given size. Defaults to False.
         print_summary (bool): Whether to print a summary of the experiment and evaluation results.
             Defaults to True.
+        concurrency (int): Specifies the concurrency for task execution. In order to enable
+            concurrent task execution, the task callable must be a coroutine function.
+            Defaults to 3.
 
     Returns:
         RanExperiment: The results of the experiment and evaluation. Additional evaluations can be
@@ -389,6 +393,7 @@ def run_experiment(
         exit_on_error=False,
         fallback_return_value=None,
         tqdm_bar_format=get_tqdm_progress_bar_formatter("running tasks"),
+        concurrency=concurrency,
     )
 
     test_cases = [
@@ -414,6 +419,7 @@ def run_experiment(
             dry_run=dry_run,
             print_summary=print_summary,
             rate_limit_errors=rate_limit_errors,
+            concurrency=concurrency,
         )
     if print_summary:
         print(ran_experiment)
@@ -427,6 +433,7 @@ def evaluate_experiment(
     dry_run: Union[bool, int] = False,
     print_summary: bool = True,
     rate_limit_errors: Optional[RateLimitErrors] = None,
+    concurrency: int = 3,
 ) -> RanExperiment:
     if not dry_run and _is_dry_run(experiment):
         dry_run = True
@@ -628,6 +635,7 @@ def evaluate_experiment(
         exit_on_error=False,
         fallback_return_value=None,
         tqdm_bar_format=get_tqdm_progress_bar_formatter("running experiment evaluations"),
+        concurrency=concurrency,
     )
     eval_runs, _execution_details = executor.run(evaluation_input)
     eval_summary = EvaluationSummary.from_eval_runs(

{arize_phoenix-4.22.1 → arize_phoenix-4.24.0}/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-4.22.1 → arize_phoenix-4.24.0}/src/phoenix/server/api/queries.py

@@ -66,11 +66,115 @@ from phoenix.server.api.types.pagination import (
 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.SystemApiKey import SystemApiKey
 from phoenix.server.api.types.Trace import Trace
+from phoenix.server.api.types.User import User
+from phoenix.server.api.types.UserApiKey import UserApiKey
+from phoenix.server.api.types.UserRole import UserRole
 
 
 @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)
+            .where(models.UserRole.role != "SYSTEM")
+            .order_by(models.User.email)
+            .options(joinedload(models.User.role))
+        )
+        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,
+                    role=UserRole(
+                        id_attr=user.role.id,
+                        role=user.role.role,
+                    ),
+                )
+                async for user in users
+            ]
+        return connection_from_list(data=data, args=args)
+
+    @strawberry.field
+    async def user_roles(
+        self,
+        info: Info[Context, None],
+    ) -> List[UserRole]:
+        async with info.context.db() as session:
+            roles = await session.scalars(
+                select(models.UserRole).where(models.UserRole.role != "SYSTEM")
+            )
+        return [
+            UserRole(
+                id_attr=role.id,
+                role=role.role,
+            )
+            for role in roles
+        ]
+
+    @strawberry.field
+    async def user_api_keys(self, info: Info[Context, None]) -> List[UserApiKey]:
+        # TODO(auth): add access control
+        stmt = (
+            select(models.APIKey)
+            .join(models.User)
+            .join(models.UserRole)
+            .where(models.UserRole.role != "SYSTEM")
+        )
+        async with info.context.db() as session:
+            api_keys = await session.scalars(stmt)
+        return [
+            UserApiKey(
+                id_attr=api_key.id,
+                user_id=api_key.user_id,
+                name=api_key.name,
+                description=api_key.description,
+                created_at=api_key.created_at,
+                expires_at=api_key.expires_at,
+            )
+            for api_key in api_keys
+        ]
+
+    @strawberry.field
+    async def system_api_keys(self, info: Info[Context, None]) -> List[SystemApiKey]:
+        # TODO(auth): add access control
+        stmt = (
+            select(models.APIKey)
+            .join(models.User)
+            .join(models.UserRole)
+            .where(models.UserRole.role == "SYSTEM")
+        )
+        async with info.context.db() as session:
+            api_keys = await session.scalars(stmt)
+        return [
+            SystemApiKey(
+                id_attr=api_key.id,
+                name=api_key.name,
+                description=api_key.description,
+                created_at=api_key.created_at,
+                expires_at=api_key.expires_at,
+            )
+            for api_key in api_keys
+        ]
+
     @strawberry.field
     async def projects(
         self,

arize_phoenix-4.24.0/src/phoenix/server/api/types/ApiKey.py

@@ -0,0 +1,16 @@
+from datetime import datetime
+from typing import Optional
+
+import strawberry
+
+
+@strawberry.interface
+class ApiKey:
+    name: str = strawberry.field(description="Name of the API key.")
+    description: Optional[str] = strawberry.field(description="Description of the API key.")
+    created_at: datetime = strawberry.field(
+        description="The date and time the API key was created."
+    )
+    expires_at: Optional[datetime] = strawberry.field(
+        description="The date and time the API key will expire."
+    )

{arize_phoenix-4.22.1 → arize_phoenix-4.24.0}/src/phoenix/server/api/types/Span.py

@@ -232,7 +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(
@@ -248,10 +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-4.24.0/src/phoenix/server/api/types/SystemApiKey.py

@@ -0,0 +1,9 @@
+import strawberry
+from strawberry.relay import Node, NodeID
+
+from .ApiKey import ApiKey
+
+
+@strawberry.type
+class SystemApiKey(ApiKey, Node):
+    id_attr: NodeID[int]

arize_phoenix-4.24.0/src/phoenix/server/api/types/User.py

@@ -0,0 +1,16 @@
+from datetime import datetime
+from typing import Optional
+
+import strawberry
+from strawberry.relay import Node, NodeID
+
+from .UserRole import UserRole
+
+
+@strawberry.type
+class User(Node):
+    id_attr: NodeID[int]
+    email: str
+    username: Optional[str]
+    created_at: datetime
+    role: UserRole

arize_phoenix-4.24.0/src/phoenix/server/api/types/UserApiKey.py

@@ -0,0 +1,11 @@
+import strawberry
+from strawberry import Private
+from strawberry.relay.types import Node, NodeID
+
+from .ApiKey import ApiKey
+
+
+@strawberry.type
+class UserApiKey(ApiKey, Node):
+    id_attr: NodeID[int]
+    user_id: Private[int]

arize_phoenix-4.24.0/src/phoenix/server/api/types/UserRole.py

@@ -0,0 +1,8 @@
+import strawberry
+from strawberry.relay import Node, NodeID
+
+
+@strawberry.type
+class UserRole(Node):
+    id_attr: NodeID[int]
+    role: str

{arize_phoenix-4.22.1 → arize_phoenix-4.24.0}/src/phoenix/server/app.py

@@ -115,6 +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):
@@ -162,6 +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:
@@ -395,6 +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,
@@ -515,6 +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-4.22.1 → arize_phoenix-4.24.0}/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()

arize_phoenix-4.24.0/src/phoenix/server/static/.vite/manifest.json

@@ -0,0 +1,78 @@
+{
+  "_components-DzA9gIHT.js": {
+    "file": "assets/components-DzA9gIHT.js",
+    "name": "components",
+    "imports": [
+      "_vendor-CIqy43_9.js",
+      "_vendor-arizeai-B1YgcWL8.js",
+      "_pages-DzkUGFGV.js",
+      "_vendor-three-DwGkEfCM.js",
+      "_vendor-codemirror-_bcwCA1C.js"
+    ]
+  },
+  "_pages-DzkUGFGV.js": {
+    "file": "assets/pages-DzkUGFGV.js",
+    "name": "pages",
+    "imports": [
+      "_vendor-CIqy43_9.js",
+      "_components-DzA9gIHT.js",
+      "_vendor-arizeai-B1YgcWL8.js",
+      "_vendor-recharts-C3pM_Wlg.js",
+      "_vendor-codemirror-_bcwCA1C.js"
+    ]
+  },
+  "_vendor-!~{003}~.js": {
+    "file": "assets/vendor-DxkFTwjz.css",
+    "src": "_vendor-!~{003}~.js"
+  },
+  "_vendor-CIqy43_9.js": {
+    "file": "assets/vendor-CIqy43_9.js",
+    "name": "vendor",
+    "imports": [
+      "_vendor-three-DwGkEfCM.js"
+    ],
+    "css": [
+      "assets/vendor-DxkFTwjz.css"
+    ]
+  },
+  "_vendor-arizeai-B1YgcWL8.js": {
+    "file": "assets/vendor-arizeai-B1YgcWL8.js",
+    "name": "vendor-arizeai",
+    "imports": [
+      "_vendor-CIqy43_9.js"
+    ]
+  },
+  "_vendor-codemirror-_bcwCA1C.js": {
+    "file": "assets/vendor-codemirror-_bcwCA1C.js",
+    "name": "vendor-codemirror",
+    "imports": [
+      "_vendor-CIqy43_9.js"
+    ]
+  },
+  "_vendor-recharts-C3pM_Wlg.js": {
+    "file": "assets/vendor-recharts-C3pM_Wlg.js",
+    "name": "vendor-recharts",
+    "imports": [
+      "_vendor-CIqy43_9.js"
+    ]
+  },
+  "_vendor-three-DwGkEfCM.js": {
+    "file": "assets/vendor-three-DwGkEfCM.js",
+    "name": "vendor-three"
+  },
+  "index.tsx": {
+    "file": "assets/index-BuTlV4Gk.js",
+    "name": "index",
+    "src": "index.tsx",
+    "isEntry": true,
+    "imports": [
+      "_vendor-CIqy43_9.js",
+      "_vendor-arizeai-B1YgcWL8.js",
+      "_pages-DzkUGFGV.js",
+      "_components-DzA9gIHT.js",
+      "_vendor-three-DwGkEfCM.js",
+      "_vendor-recharts-C3pM_Wlg.js",
+      "_vendor-codemirror-_bcwCA1C.js"
+    ]
+  }
+}