arize-phoenix 3.13.1__tar.gz → 3.14.1__tar.gz

{arize_phoenix-3.13.1 → arize_phoenix-3.14.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: arize-phoenix
-Version: 3.13.1
+Version: 3.14.1
 Summary: AI Observability and Evaluation
 Project-URL: Documentation, https://docs.arize.com/phoenix/
 Project-URL: Issues, https://github.com/Arize-ai/phoenix/issues
@@ -18,7 +18,7 @@ Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Requires-Python: <3.13,>=3.8
 Requires-Dist: ddsketch
-Requires-Dist: hdbscan<1.0.0,>=0.8.33
+Requires-Dist: hdbscan>=0.8.33
 Requires-Dist: jinja2
 Requires-Dist: numpy
 Requires-Dist: openinference-instrumentation-langchain>=0.1.12
@@ -33,13 +33,14 @@ Requires-Dist: protobuf<5.0,>=3.20
 Requires-Dist: psutil
 Requires-Dist: pyarrow
 Requires-Dist: requests
-Requires-Dist: scikit-learn<1.3.0
+Requires-Dist: scikit-learn
 Requires-Dist: scipy
 Requires-Dist: sortedcontainers
 Requires-Dist: starlette
 Requires-Dist: strawberry-graphql==0.208.2
 Requires-Dist: tqdm
-Requires-Dist: typing-extensions<5,>=4.5
+Requires-Dist: typing-extensions>=4.5; python_version < '3.12'
+Requires-Dist: typing-extensions>=4.6; python_version >= '3.12'
 Requires-Dist: umap-learn
 Requires-Dist: uvicorn
 Requires-Dist: wrapt

{arize_phoenix-3.13.1 → arize_phoenix-3.14.1}/pyproject.toml

@@ -22,18 +22,20 @@ classifiers = [
   "Programming Language :: Python :: 3.12",
 ]
 dependencies = [
-  "scikit-learn<1.3.0",
+  "scikit-learn",
   "numpy",
   "pandas",
   "jinja2",
   "umap-learn",
-  "hdbscan>=0.8.33, <1.0.0",
+  "hdbscan>=0.8.33",
   "starlette",
   "uvicorn",
   "psutil",
   "strawberry-graphql==0.208.2",
   "pyarrow",
-  "typing-extensions>=4.5, <5",
+  "typing-extensions>=4.5; python_version<'3.12'",
+  # A minimum version of typing-extensions==4.6.0 is needed to avoid this issue on Python 3.12: https://github.com/Azure/azure-sdk-for-python/issues/33442#issuecomment-1847886784
+  "typing-extensions>=4.6; python_version>='3.12'",
   "scipy",
   "wrapt",
   "sortedcontainers",
@@ -123,10 +125,12 @@ dependencies = [
   "protobuf==3.20",  # version minimum (for tests)
   "responses",
   "tiktoken",
-  "typing-extensions<4.6.0",  # for Colab
+  "typing-extensions==4.5.0; python_version<'3.12'",
+  "typing-extensions==4.6.0; python_version>='3.12'",
   "httpx", # For OpenAI testing
   "respx", # For OpenAI testing
   "nest-asyncio", # for executor testing
+  "pyfakefs", # for experimental storage implementations
 ]
 
 [tool.hatch.envs.type]
@@ -166,7 +170,7 @@ tests = "pytest {args}"
 coverage = "pytest --cov-report=term-missing --cov-config=pyproject.toml --cov=src/phoenix --cov=tests {args}"
 
 [[tool.hatch.envs.test.matrix]]
-python = ["3.8", "3.11"]
+python = ["3.8", "3.12"]
 
 [tool.pytest.ini_options]
 asyncio_mode = "auto"

{arize_phoenix-3.13.1 → arize_phoenix-3.14.1}/src/phoenix/config.py

@@ -56,6 +56,13 @@ def get_working_dir() -> Path:
     return Path.home().resolve() / ".phoenix"
 
 
+def get_storage_dir() -> Path:
+    """
+    Get the directory for storing traces.
+    """
+    return get_working_dir() / "storage"
+
+
 PHOENIX_DIR = Path(__file__).resolve().parent
 # Server config
 SERVER_DIR = PHOENIX_DIR / "server"

{arize_phoenix-3.13.1 → arize_phoenix-3.14.1}/src/phoenix/core/project.py

@@ -82,7 +82,6 @@ _ParentSpanID: TypeAlias = SpanID
 _ChildSpanID: TypeAlias = SpanID
 _ProjectName: TypeAlias = str
 
-DEFAULT_PROJECT_NAME: str = "default"
 
 EvaluationName: TypeAlias = str
 DocumentPosition: TypeAlias = int

{arize_phoenix-3.13.1 → arize_phoenix-3.14.1}/src/phoenix/core/traces.py

@@ -9,12 +9,12 @@ from typing_extensions import assert_never
 
 import phoenix.trace.v1 as pb
 from phoenix.core.project import (
-    DEFAULT_PROJECT_NAME,
     END_OF_QUEUE,
     Project,
     _ProjectName,
 )
 from phoenix.trace.schemas import Span
+from phoenix.utilities.project import DEFAULT_PROJECT_NAME
 
 _SpanItem = Tuple[Span, _ProjectName]
 _EvalItem = Tuple[pb.Evaluation, _ProjectName]

{arize_phoenix-3.13.1 → arize_phoenix-3.14.1}/src/phoenix/server/api/routers/evaluation_handler.py

@@ -15,11 +15,11 @@ from starlette.status import (
 )
 
 import phoenix.trace.v1 as pb
-from phoenix.core.project import DEFAULT_PROJECT_NAME
 from phoenix.core.traces import Traces
 from phoenix.server.api.routers.utils import table_to_bytes
 from phoenix.session.evaluation import encode_evaluations
 from phoenix.trace.span_evaluations import Evaluations
+from phoenix.utilities.project import DEFAULT_PROJECT_NAME
 
 
 class EvaluationHandler(HTTPEndpoint):

{arize_phoenix-3.13.1 → arize_phoenix-3.14.1}/src/phoenix/server/api/routers/span_handler.py

@@ -7,11 +7,11 @@ from starlette.requests import Request
 from starlette.responses import Response, StreamingResponse
 from starlette.status import HTTP_404_NOT_FOUND, HTTP_422_UNPROCESSABLE_ENTITY
 
-from phoenix.core.project import DEFAULT_PROJECT_NAME
 from phoenix.core.traces import Traces
 from phoenix.server.api.routers.utils import df_to_bytes, from_iso_format
 from phoenix.trace.dsl import SpanQuery
 from phoenix.utilities import query_spans
+from phoenix.utilities.project import DEFAULT_PROJECT_NAME
 
 
 class SpanHandler(HTTPEndpoint):

{arize_phoenix-3.13.1 → arize_phoenix-3.14.1}/src/phoenix/server/api/routers/trace_handler.py

@@ -1,25 +1,27 @@
 import asyncio
 import gzip
 import zlib
-from typing import Iterable, Optional
+from typing import Optional
 
 from google.protobuf.message import DecodeError
-from openinference.semconv.resource import ResourceAttributes
 from opentelemetry.proto.collector.trace.v1.trace_service_pb2 import (
     ExportTraceServiceRequest,
 )
-from opentelemetry.proto.common.v1.common_pb2 import KeyValue
+from opentelemetry.proto.trace.v1.trace_pb2 import TracesData
 from starlette.endpoints import HTTPEndpoint
 from starlette.requests import Request
 from starlette.responses import Response
 from starlette.status import HTTP_415_UNSUPPORTED_MEDIA_TYPE, HTTP_422_UNPROCESSABLE_ENTITY
 
 from phoenix.core.traces import Traces
+from phoenix.storage.spanstore import SpanStore
 from phoenix.trace.otel import decode
+from phoenix.utilities.project import get_project_name
 
 
 class TraceHandler(HTTPEndpoint):
     traces: Traces
+    store: Optional[SpanStore]
 
     async def post(self, request: Request) -> Response:
         content_type = request.headers.get("content-type")
@@ -47,17 +49,12 @@ class TraceHandler(HTTPEndpoint):
                 content="Request body is invalid ExportTraceServiceRequest",
                 status_code=HTTP_422_UNPROCESSABLE_ENTITY,
             )
+        if self.store:
+            self.store.save(TracesData(resource_spans=req.resource_spans))
         for resource_spans in req.resource_spans:
-            project_name = _get_project_name(resource_spans.resource.attributes)
+            project_name = get_project_name(resource_spans.resource.attributes)
             for scope_span in resource_spans.scope_spans:
                 for span in scope_span.spans:
                     self.traces.put(decode(span), project_name=project_name)
                     await asyncio.sleep(0)
         return Response()
-
-
-def _get_project_name(attributes: Iterable[KeyValue]) -> Optional[str]:
-    for kv in attributes:
-        if kv.key == ResourceAttributes.PROJECT_NAME and (v := kv.value.string_value):
-            return v
-    return None

{arize_phoenix-3.13.1 → arize_phoenix-3.14.1}/src/phoenix/server/api/types/Span.py

@@ -10,13 +10,14 @@ from strawberry import ID, UNSET
 from strawberry.types import Info
 
 import phoenix.trace.schemas as trace_schema
-from phoenix.core.project import DEFAULT_PROJECT_NAME, Project, WrappedSpan
+from phoenix.core.project import Project, WrappedSpan
 from phoenix.metrics.retrieval_metrics import RetrievalMetrics
 from phoenix.server.api.context import Context
 from phoenix.server.api.types.DocumentRetrievalMetrics import DocumentRetrievalMetrics
 from phoenix.server.api.types.Evaluation import DocumentEvaluation, SpanEvaluation
 from phoenix.server.api.types.MimeType import MimeType
 from phoenix.trace.schemas import ComputedAttributes, SpanID
+from phoenix.utilities.project import DEFAULT_PROJECT_NAME
 
 EMBEDDING_EMBEDDINGS = SpanAttributes.EMBEDDING_EMBEDDINGS
 EMBEDDING_VECTOR = EmbeddingAttributes.EMBEDDING_VECTOR

{arize_phoenix-3.13.1 → arize_phoenix-3.14.1}/src/phoenix/server/app.py

@@ -28,6 +28,7 @@ from phoenix.server.api.routers.evaluation_handler import EvaluationHandler
 from phoenix.server.api.routers.span_handler import SpanHandler
 from phoenix.server.api.routers.trace_handler import TraceHandler
 from phoenix.server.api.schema import schema
+from phoenix.storage.spanstore import SpanStore
 
 logger = logging.getLogger(__name__)
 
@@ -147,6 +148,7 @@ def create_app(
     umap_params: UMAPParameters,
     corpus: Optional[Model] = None,
     traces: Optional[Traces] = None,
+    span_store: Optional[SpanStore] = None,
     debug: bool = False,
     read_only: bool = False,
 ) -> Starlette:
@@ -173,7 +175,7 @@ def create_app(
                 ),
                 Route(
                     "/v1/traces",
-                    type("TraceEndpoint", (TraceHandler,), {"traces": traces}),
+                    type("TraceEndpoint", (TraceHandler,), {"traces": traces, "store": span_store}),
                 ),
                 Route(
                     "/v1/evaluations",

{arize_phoenix-3.13.1 → arize_phoenix-3.14.1}/src/phoenix/server/main.py

@@ -11,7 +11,13 @@ from typing import Iterable, Optional, Protocol, TypeVar
 import pkg_resources
 from uvicorn import Config, Server
 
-from phoenix.config import EXPORT_DIR, get_env_host, get_env_port, get_pids_path
+from phoenix.config import (
+    EXPORT_DIR,
+    get_env_host,
+    get_env_port,
+    get_pids_path,
+    get_storage_dir,
+)
 from phoenix.core.model_schema_adapter import create_model_from_datasets
 from phoenix.core.traces import Traces
 from phoenix.datasets.dataset import EMPTY_DATASET, Dataset
@@ -23,6 +29,8 @@ from phoenix.pointcloud.umap_parameters import (
     UMAPParameters,
 )
 from phoenix.server.app import create_app
+from phoenix.storage.spanstore import SpanStore
+from phoenix.storage.spanstore.text_file import TextFileSpanStoreImpl
 from phoenix.trace.fixtures import (
     TRACES_FIXTURES,
     _download_traces_fixture,
@@ -31,6 +39,7 @@ from phoenix.trace.fixtures import (
 )
 from phoenix.trace.otel import decode, encode
 from phoenix.trace.span_json_decoder import json_string_to_span
+from phoenix.utilities.project import get_project_name
 
 logger = logging.getLogger(__name__)
 
@@ -99,6 +108,15 @@ def _load_items(
         queue.put(item)
 
 
+def _load_from_store(traces: Traces, span_store: SpanStore) -> None:
+    for traces_data in span_store.load():
+        for resource_spans in traces_data.resource_spans:
+            project_name = get_project_name(resource_spans.resource.attributes)
+            for scope_span in resource_spans.scope_spans:
+                for span in scope_span.spans:
+                    traces.put(decode(span), project_name=project_name)
+
+
 DEFAULT_UMAP_PARAMS_STR = f"{DEFAULT_MIN_DIST},{DEFAULT_N_NEIGHBORS},{DEFAULT_N_SAMPLES}"
 
 if __name__ == "__main__":
@@ -124,6 +142,8 @@ if __name__ == "__main__":
     parser.add_argument("--debug", action="store_false")
     subparsers = parser.add_subparsers(dest="command", required=True)
     serve_parser = subparsers.add_parser("serve")
+    experimental_parser = subparsers.add_parser("extremely-dangerous-experimental-span-storage")
+    experimental_parser.add_argument("--storage-path", type=str, required=False)
     datasets_parser = subparsers.add_parser("datasets")
     datasets_parser.add_argument("--primary", type=str, required=True)
     datasets_parser.add_argument("--reference", type=str, required=False)
@@ -145,6 +165,7 @@ if __name__ == "__main__":
     demo_parser.add_argument("--simulate-streaming", action="store_true")
     args = parser.parse_args()
     export_path = Path(args.export_path) if args.export_path else EXPORT_DIR
+    span_store: Optional[SpanStore] = None
     if args.command == "datasets":
         primary_dataset_name = args.primary
         reference_dataset_name = args.reference
@@ -179,12 +200,19 @@ if __name__ == "__main__":
         )
         trace_dataset_name = args.trace_fixture
         simulate_streaming = args.simulate_streaming
+    elif args.command == "extremely-dangerous-experimental-span-storage":
+        span_store_path = (
+            get_storage_dir() if args.storage_path is None else Path(args.storage_path)
+        )
+        span_store = TextFileSpanStoreImpl(span_store_path)
 
     model = create_model_from_datasets(
         primary_dataset,
         reference_dataset,
     )
     traces = Traces()
+    if span_store:
+        Thread(target=_load_from_store, args=(traces, span_store), daemon=True).start()
     if trace_dataset_name is not None:
         fixture_spans = list(
             # Apply `encode` here because legacy jsonl files contains UUIDs as strings.
@@ -221,6 +249,7 @@ if __name__ == "__main__":
         corpus=None if corpus_dataset is None else create_model_from_datasets(corpus_dataset),
         debug=args.debug,
         read_only=read_only,
+        span_store=span_store,
     )
     host = args.host or get_env_host()
     port = args.port or get_env_port()