arize-phoenix 4.26.0__py3-none-any.whl → 4.28.1__py3-none-any.whl

phoenix/server/app.py

@@ -2,6 +2,7 @@ import asyncio
 import contextlib
 import json
 import logging
+from datetime import datetime, timedelta, timezone
 from functools import cached_property
 from pathlib import Path
 from types import MethodType
@@ -10,12 +11,14 @@ from typing import (
     Any,
     AsyncContextManager,
     AsyncIterator,
+    Awaitable,
     Callable,
     Dict,
     Iterable,
     List,
     NamedTuple,
     Optional,
+    Set,
     Tuple,
     Union,
     cast,
@@ -26,11 +29,8 @@ from fastapi import APIRouter, FastAPI
 from fastapi.middleware.gzip import GZipMiddleware
 from fastapi.responses import FileResponse
 from fastapi.utils import is_body_allowed_for_status_code
-from sqlalchemy.ext.asyncio import (
-    AsyncEngine,
-    AsyncSession,
-    async_sessionmaker,
-)
+from sqlalchemy import select
+from sqlalchemy.ext.asyncio import AsyncEngine, AsyncSession, async_sessionmaker
 from starlette.datastructures import State as StarletteState
 from starlette.exceptions import HTTPException
 from starlette.middleware import Middleware
@@ -46,12 +46,9 @@ from typing_extensions import TypeAlias
 
 import phoenix
 import phoenix.trace.v1 as pb
-from phoenix.config import (
-    DEFAULT_PROJECT_NAME,
-    SERVER_DIR,
-    server_instrumentation_is_enabled,
-)
+from phoenix.config import DEFAULT_PROJECT_NAME, SERVER_DIR, server_instrumentation_is_enabled
 from phoenix.core.model_schema import Model
+from phoenix.db import models
 from phoenix.db.bulk_inserter import BulkInserter
 from phoenix.db.engines import create_engine
 from phoenix.db.helpers import SupportedSQLDialect
@@ -82,7 +79,6 @@ from phoenix.server.api.dataloaders import (
     TokenCountDataLoader,
     TraceRowIdsDataLoader,
 )
-from phoenix.server.api.routers.auth import router as auth_router
 from phoenix.server.api.routers.v1 import REST_API_VERSION
 from phoenix.server.api.routers.v1 import router as v1_router
 from phoenix.server.api.schema import schema
@@ -93,9 +89,17 @@ from phoenix.server.telemetry import initialize_opentelemetry_tracer_provider
 from phoenix.server.types import (
     CanGetLastUpdatedAt,
     CanPutItem,
+    DaemonTask,
     DbSessionFactory,
     LastUpdatedAt,
 )
+from phoenix.trace.fixtures import (
+    get_evals_from_fixture,
+    get_trace_fixture_by_name,
+    load_example_traces,
+    reset_fixture_span_ids_and_timestamps,
+)
+from phoenix.trace.otel import decode_otlp_span, encode_span_to_otlp
 from phoenix.trace.schemas import Span
 from phoenix.utilities.client import PHOENIX_SERVER_VERSION_HEADER
 
@@ -103,12 +107,23 @@ if TYPE_CHECKING:
     from opentelemetry.trace import TracerProvider
 
 logger = logging.getLogger(__name__)
+logger.setLevel(logging.INFO)
 logger.addHandler(logging.NullHandler())
 
 router = APIRouter(include_in_schema=False)
 
 templates = Jinja2Templates(directory=SERVER_DIR / "templates")
 
+"""
+Threshold (in minutes) to determine if database is booted up for the first time.
+
+Used to assess whether the `default` project was created recently.
+If so, demo data is automatically ingested upon initial boot up to populate the database.
+"""
+NEW_DB_AGE_THRESHOLD_MINUTES = 2
+
+ProjectName: TypeAlias = str
+
 
 class AppConfig(NamedTuple):
     has_inferences: bool
@@ -226,9 +241,99 @@ def _db(engine: AsyncEngine) -> Callable[[], AsyncContextManager[AsyncSession]]:
     return factory
 
 
+class Scaffolder(DaemonTask):
+    def __init__(
+        self,
+        db: DbSessionFactory,
+        queue_span: Callable[[Span, ProjectName], Awaitable[None]],
+        queue_evaluation: Callable[[pb.Evaluation], Awaitable[None]],
+        tracing_fixture_names: Set[str] = set(),
+        force_fixture_ingestion: bool = False,
+    ) -> None:
+        super().__init__()
+        self._db = db
+        self._queue_span = queue_span
+        self._queue_evaluation = queue_evaluation
+        self._tracing_fixtures = set(
+            get_trace_fixture_by_name(name) for name in tracing_fixture_names
+        )
+        self._force_fixture_ingestion = force_fixture_ingestion
+
+    async def __aenter__(self) -> None:
+        await self.start()
+
+    async def __aexit__(self, *args: Any, **kwargs: Any) -> None:
+        await self.stop()
+
+    async def _run(self) -> None:
+        """
+        Main entry point for Scaffolder.
+        Determines whether to load fixtures and handles them.
+        """
+        if await self._should_load_fixtures():
+            logger.info("Loading trace fixtures.")
+            await self._handle_tracing_fixtures()
+            logger.info("Finished loading fixtures.")
+        else:
+            logger.info("DB is not new, avoid loading demo fixtures.")
+
+    async def _should_load_fixtures(self) -> bool:
+        if self._force_fixture_ingestion:
+            return True
+
+        async with self._db() as session:
+            created_at = await session.scalar(
+                select(models.Project.created_at).where(models.Project.name == "default")
+            )
+        if created_at is None:
+            return False
+
+        is_new_db = datetime.now(timezone.utc) - created_at < timedelta(
+            minutes=NEW_DB_AGE_THRESHOLD_MINUTES
+        )
+        return is_new_db
+
+    async def _handle_tracing_fixtures(self) -> None:
+        """
+        Main handler for processing trace fixtures. Process each fixture by
+        loading its trace dataframe, gettting and processings its
+        spans and evals, and queuing.
+        """
+        loop = asyncio.get_running_loop()
+        for fixture in self._tracing_fixtures:
+            try:
+                trace_ds = await loop.run_in_executor(None, load_example_traces, fixture.name)
+
+                fixture_spans, fixture_evals = await loop.run_in_executor(
+                    None,
+                    reset_fixture_span_ids_and_timestamps,
+                    (
+                        # Apply `encode` here because legacy jsonl files contains UUIDs as strings.
+                        # `encode` removes the hyphens in the UUIDs.
+                        decode_otlp_span(encode_span_to_otlp(span))
+                        for span in trace_ds.to_spans()
+                    ),
+                    get_evals_from_fixture(fixture.name),
+                )
+
+                project_name = fixture.project_name or fixture.name
+                logger.info(f"Loading '{project_name}' fixtures...")
+                for span in fixture_spans:
+                    await self._queue_span(span, project_name)
+                for evaluation in fixture_evals:
+                    await self._queue_evaluation(evaluation)
+
+            except FileNotFoundError:
+                logger.warning(f"Fixture file not found for '{fixture.name}'")
+            except ValueError as e:
+                logger.error(f"Error processing fixture '{fixture.name}': {e}")
+            except Exception as e:
+                logger.error(f"Unexpected error processing fixture '{fixture.name}': {e}")
+
+
 def _lifespan(
     *,
-    dialect: SupportedSQLDialect,
+    db: DbSessionFactory,
     bulk_inserter: BulkInserter,
     dml_event_handler: DmlEventHandler,
     tracer_provider: Optional["TracerProvider"] = None,
@@ -236,11 +341,13 @@ def _lifespan(
     startup_callbacks: Iterable[Callable[[], None]] = (),
     shutdown_callbacks: Iterable[Callable[[], None]] = (),
     read_only: bool = False,
+    tracing_fixture_names: Set[str] = set(),
+    force_fixture_ingestion: bool = False,
 ) -> StatefulLifespan[FastAPI]:
     @contextlib.asynccontextmanager
     async def lifespan(_: FastAPI) -> AsyncIterator[Dict[str, Any]]:
         global DB_MUTEX
-        DB_MUTEX = asyncio.Lock() if dialect is SupportedSQLDialect.SQLITE else None
+        DB_MUTEX = asyncio.Lock() if db.dialect is SupportedSQLDialect.SQLITE else None
         async with bulk_inserter as (
             enqueue,
             queue_span,
@@ -251,7 +358,13 @@ def _lifespan(
             disabled=read_only,
             tracer_provider=tracer_provider,
             enable_prometheus=enable_prometheus,
-        ), dml_event_handler:
+        ), dml_event_handler, Scaffolder(
+            db=db,
+            queue_span=queue_span,
+            queue_evaluation=queue_evaluation,
+            tracing_fixture_names=tracing_fixture_names,
+            force_fixture_ingestion=force_fixture_ingestion,
+        ):
             for callback in startup_callbacks:
                 callback()
             yield {
@@ -317,17 +430,19 @@ def create_graphql_router(
                 dataset_example_spans=DatasetExampleSpansDataLoader(db),
                 document_evaluation_summaries=DocumentEvaluationSummaryDataLoader(
                     db,
-                    cache_map=cache_for_dataloaders.document_evaluation_summary
-                    if cache_for_dataloaders
-                    else None,
+                    cache_map=(
+                        cache_for_dataloaders.document_evaluation_summary
+                        if cache_for_dataloaders
+                        else None
+                    ),
                 ),
                 document_evaluations=DocumentEvaluationsDataLoader(db),
                 document_retrieval_metrics=DocumentRetrievalMetricsDataLoader(db),
                 annotation_summaries=AnnotationSummaryDataLoader(
                     db,
-                    cache_map=cache_for_dataloaders.annotation_summary
-                    if cache_for_dataloaders
-                    else None,
+                    cache_map=(
+                        cache_for_dataloaders.annotation_summary if cache_for_dataloaders else None
+                    ),
                 ),
                 experiment_annotation_summaries=ExperimentAnnotationSummaryDataLoader(db),
                 experiment_error_rates=ExperimentErrorRatesDataLoader(db),
@@ -335,15 +450,17 @@ def create_graphql_router(
                 experiment_sequence_number=ExperimentSequenceNumberDataLoader(db),
                 latency_ms_quantile=LatencyMsQuantileDataLoader(
                     db,
-                    cache_map=cache_for_dataloaders.latency_ms_quantile
-                    if cache_for_dataloaders
-                    else None,
+                    cache_map=(
+                        cache_for_dataloaders.latency_ms_quantile if cache_for_dataloaders else None
+                    ),
                 ),
                 min_start_or_max_end_times=MinStartOrMaxEndTimeDataLoader(
                     db,
-                    cache_map=cache_for_dataloaders.min_start_or_max_end_time
-                    if cache_for_dataloaders
-                    else None,
+                    cache_map=(
+                        cache_for_dataloaders.min_start_or_max_end_time
+                        if cache_for_dataloaders
+                        else None
+                    ),
                 ),
                 record_counts=RecordCountDataLoader(
                     db,
@@ -438,6 +555,8 @@ def create_app(
     startup_callbacks: Iterable[Callable[[], None]] = (),
     shutdown_callbacks: Iterable[Callable[[], None]] = (),
     secret: Optional[str] = None,
+    tracing_fixture_names: Set[str] = set(),
+    force_fixture_ingestion: bool = False,
 ) -> FastAPI:
     startup_callbacks_list: List[Callable[[], None]] = list(startup_callbacks)
     shutdown_callbacks_list: List[Callable[[], None]] = list(shutdown_callbacks)
@@ -510,11 +629,12 @@ def create_app(
         prometheus_middlewares = [Middleware(PrometheusMiddleware)]
     else:
         prometheus_middlewares = []
+
     app = FastAPI(
         title="Arize-Phoenix REST API",
         version=REST_API_VERSION,
         lifespan=_lifespan(
-            dialect=db.dialect,
+            db=db,
             read_only=read_only,
             bulk_inserter=bulk_inserter,
             dml_event_handler=dml_event_handler,
@@ -522,6 +642,8 @@ def create_app(
             enable_prometheus=enable_prometheus,
             shutdown_callbacks=shutdown_callbacks_list,
             startup_callbacks=startup_callbacks_list,
+            tracing_fixture_names=tracing_fixture_names,
+            force_fixture_ingestion=force_fixture_ingestion,
         ),
         middleware=[
             Middleware(HeadersMiddleware),
@@ -539,8 +661,6 @@ def create_app(
     app.include_router(router)
     app.include_router(graphql_router)
     app.add_middleware(GZipMiddleware)
-    if authentication_enabled:
-        app.include_router(auth_router)
     if serve_ui:
         app.mount(
             "/",

phoenix/server/main.py

@@ -48,6 +48,7 @@ from phoenix.trace.fixtures import (
     TRACES_FIXTURES,
     get_dataset_fixtures,
     get_evals_from_fixture,
+    get_trace_fixtures_by_project_name,
     load_example_traces,
     reset_fixture_span_ids_and_timestamps,
     send_dataset_fixtures,
@@ -135,7 +136,7 @@ if __name__ == "__main__":
     parser.add_argument("--export_path")
     parser.add_argument("--host", type=str, required=False)
     parser.add_argument("--port", type=int, required=False)
-    parser.add_argument("--read-only", type=bool, default=False)
+    parser.add_argument("--read-only", action="store_true", required=False)  # Default is False
     parser.add_argument("--no-internet", action="store_true")
     parser.add_argument("--umap_params", type=str, required=False, default=DEFAULT_UMAP_PARAMS_STR)
     parser.add_argument("--debug", action="store_true")
@@ -144,6 +145,43 @@ if __name__ == "__main__":
     parser.add_argument("--no-ui", action="store_true")
     subparsers = parser.add_subparsers(dest="command", required=True)
     serve_parser = subparsers.add_parser("serve")
+    serve_parser.add_argument(
+        "--with-fixture",
+        type=str,
+        required=False,
+        default="",
+        help=("Name of an inference fixture. Example: 'fixture1'"),
+    )
+    serve_parser.add_argument(
+        "--with-trace-fixtures",
+        type=str,
+        required=False,
+        default="",
+        help=(
+            "Comma separated list of tracing fixture names (spaces are ignored). "
+            "Example: 'fixture1, fixture2'"
+        ),
+    )
+    serve_parser.add_argument(
+        "--with-projects",
+        type=str,
+        required=False,
+        default="",
+        help=(
+            "Comma separated list of project names (spaces are ignored). "
+            "Example: 'project1, project2'"
+        ),
+    )
+    serve_parser.add_argument(
+        "--force-fixture-ingestion",
+        action="store_true",  # default is False
+        required=False,
+        help=(
+            "Whether or not to check the database age before adding the fixtures. "
+            "Default is False, i.e., fixtures will only be added if the "
+            "database is new."
+        ),
+    )
     datasets_parser = subparsers.add_parser("datasets")
     datasets_parser.add_argument("--primary", type=str, required=True)
     datasets_parser.add_argument("--reference", type=str, required=False)
@@ -151,12 +189,14 @@ if __name__ == "__main__":
     datasets_parser.add_argument("--trace", type=str, required=False)
     fixture_parser = subparsers.add_parser("fixture")
     fixture_parser.add_argument("fixture", type=str, choices=[fixture.name for fixture in FIXTURES])
-    fixture_parser.add_argument("--primary-only", type=bool)
+    fixture_parser.add_argument("--primary-only", action="store_true")  # Default is False
     trace_fixture_parser = subparsers.add_parser("trace-fixture")
     trace_fixture_parser.add_argument(
         "fixture", type=str, choices=[fixture.name for fixture in TRACES_FIXTURES]
     )
-    trace_fixture_parser.add_argument("--simulate-streaming", type=bool)
+    trace_fixture_parser.add_argument(
+        "--simulate-streaming", action="store_true"
+    )  # Default is False
     demo_parser = subparsers.add_parser("demo")
     demo_parser.add_argument("fixture", type=str, choices=[fixture.name for fixture in FIXTURES])
     demo_parser.add_argument(
@@ -164,10 +204,12 @@ if __name__ == "__main__":
     )
     demo_parser.add_argument("--simulate-streaming", action="store_true")
     args = parser.parse_args()
+
     db_connection_str = (
         args.database_url if args.database_url else get_env_database_connection_str()
     )
     export_path = Path(args.export_path) if args.export_path else EXPORT_DIR
+    force_fixture_ingestion = False
     if args.command == "datasets":
         primary_inferences_name = args.primary
         reference_inferences_name = args.reference
@@ -202,7 +244,26 @@ if __name__ == "__main__":
         )
         trace_dataset_name = args.trace_fixture
         simulate_streaming = args.simulate_streaming
-
+    elif args.command == "serve":
+        # We use sets to avoid duplicates
+        tracing_fixture_names = set()
+        if args.with_fixture:
+            primary_inferences, reference_inferences, corpus_inferences = get_inferences(
+                str(args.with_fixture),
+                args.no_internet,
+            )
+        if args.with_trace_fixtures:
+            tracing_fixture_names.update(
+                [name.strip() for name in args.with_trace_fixtures.split(",")]
+            )
+        if args.with_projects:
+            project_names = [name.strip() for name in args.with_projects.split(",")]
+            tracing_fixture_names.update(
+                fixture.name
+                for name in project_names
+                for fixture in get_trace_fixtures_by_project_name(name)
+            )
+        force_fixture_ingestion = args.force_fixture_ingestion
     host: Optional[str] = args.host or get_env_host()
     display_host = host or "localhost"
     # If the host is "::", the convention is to bind to all interfaces. However, uvicorn
@@ -260,6 +321,9 @@ if __name__ == "__main__":
     engine = create_engine_and_run_migrations(db_connection_str)
     instrumentation_cleanups = instrument_engine_if_enabled(engine)
     factory = DbSessionFactory(db=_db(engine), dialect=engine.dialect.name)
+    corpus_model = (
+        None if corpus_inferences is None else create_model_from_inferences(corpus_inferences)
+    )
     # Print information about the server
     msg = _WELCOME_MESSAGE.format(
         version=version("arize-phoenix"),
@@ -278,9 +342,7 @@ if __name__ == "__main__":
         model=model,
         authentication_enabled=authentication_enabled,
         umap_params=umap_params,
-        corpus=None
-        if corpus_inferences is None
-        else create_model_from_inferences(corpus_inferences),
+        corpus=corpus_model,
         debug=args.debug,
         dev=args.dev,
         serve_ui=not args.no_ui,
@@ -291,6 +353,8 @@ if __name__ == "__main__":
         startup_callbacks=[lambda: print(msg)],
         shutdown_callbacks=instrumentation_cleanups,
         secret=secret,
+        tracing_fixture_names=tracing_fixture_names,
+        force_fixture_ingestion=force_fixture_ingestion,
     )
     server = Server(config=Config(app, host=host, port=port, root_path=host_root_path))  # type: ignore
     Thread(target=_write_pid_file_when_ready, args=(server,), daemon=True).start()

phoenix/server/static/.vite/manifest.json

@@ -1,21 +1,21 @@
 {
-  "_components-1Ahruijo.js": {
-    "file": "assets/components-1Ahruijo.js",
+  "_components-BYH03rjA.js": {
+    "file": "assets/components-BYH03rjA.js",
     "name": "components",
     "imports": [
       "_vendor-aSQri0vz.js",
       "_vendor-arizeai-CsdcB1NH.js",
-      "_pages-CFS6mPnW.js",
+      "_pages-DnbxgoTK.js",
       "_vendor-three-DwGkEfCM.js",
       "_vendor-codemirror-CYHkhs7D.js"
     ]
   },
-  "_pages-CFS6mPnW.js": {
-    "file": "assets/pages-CFS6mPnW.js",
+  "_pages-DnbxgoTK.js": {
+    "file": "assets/pages-DnbxgoTK.js",
     "name": "pages",
     "imports": [
       "_vendor-aSQri0vz.js",
-      "_components-1Ahruijo.js",
+      "_components-BYH03rjA.js",
       "_vendor-arizeai-CsdcB1NH.js",
       "_vendor-recharts-B0sannek.js",
       "_vendor-codemirror-CYHkhs7D.js"
@@ -61,15 +61,15 @@
     "name": "vendor-three"
   },
   "index.tsx": {
-    "file": "assets/index-BEE_RWJx.js",
+    "file": "assets/index-fqdjNpYm.js",
     "name": "index",
     "src": "index.tsx",
     "isEntry": true,
     "imports": [
       "_vendor-aSQri0vz.js",
       "_vendor-arizeai-CsdcB1NH.js",
-      "_pages-CFS6mPnW.js",
-      "_components-1Ahruijo.js",
+      "_pages-DnbxgoTK.js",
+      "_components-BYH03rjA.js",
       "_vendor-three-DwGkEfCM.js",
       "_vendor-recharts-B0sannek.js",
       "_vendor-codemirror-CYHkhs7D.js"