arize-phoenix 5.7.0__tar.gz → 5.8.0__tar.gz

{arize_phoenix-5.7.0 → arize_phoenix-5.8.0}/PKG-INFO

@@ -1,14 +1,12 @@
 Metadata-Version: 2.3
 Name: arize-phoenix
-Version: 5.7.0
+Version: 5.8.0
 Summary: AI Observability and Evaluation
 Project-URL: Documentation, https://docs.arize.com/phoenix/
 Project-URL: Issues, https://github.com/Arize-ai/phoenix/issues
 Project-URL: Source, https://github.com/Arize-ai/phoenix
 Author-email: Arize AI <phoenix-devs@arize.com>
 License: Elastic-2.0
-License-File: IP_NOTICE
-License-File: LICENSE
 Keywords: Explainability,Monitoring,Observability
 Classifier: Programming Language :: Python
 Classifier: Programming Language :: Python :: 3.9
@@ -131,7 +129,7 @@ Description-Content-Type: text/markdown
     <a href="https://docs.arize.com/phoenix/">
         <img src="https://img.shields.io/static/v1?message=Docs&logo=data:image/png;base64,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&labelColor=grey&color=blue&logoColor=white&label=%20"/>
     </a>
-    <a target="_blank" href="https://join.slack.com/t/arize-ai/shared_invite/zt-1px8dcmlf-fmThhDFD_V_48oU7ALan4Q">
+    <a target="_blank" href="https://join.slack.com/t/arize-ai/shared_invite/zt-22vj03k4k-MlrNEwv5WeswapTs0kNCBw">
         <img src="https://img.shields.io/static/v1?message=Community&logo=slack&labelColor=grey&color=blue&logoColor=white&label=%20"/>
     </a>
     <a target="_blank" href="https://twitter.com/ArizePhoenix">
@@ -237,7 +235,7 @@ For details about tracing integrations and example applications, see the [OpenIn
 
 Join our community to connect with thousands of AI builders.
 
-- 🌍 Join our [Slack community](https://join.slack.com/t/arize-ai/shared_invite/zt-1px8dcmlf-fmThhDFD_V_48oU7ALan4Q).
+- 🌍 Join our [Slack community](https://join.slack.com/t/arize-ai/shared_invite/zt-22vj03k4k-MlrNEwv5WeswapTs0kNCBw).
 - 📚 Read our [documentation](https://docs.arize.com/phoenix).
 - 💡 Ask questions and provide feedback in the _#phoenix-support_ channel.
 - 🌟 Leave a star on our [GitHub](https://github.com/Arize-ai/phoenix).

{arize_phoenix-5.7.0 → arize_phoenix-5.8.0}/README.md

@@ -7,7 +7,7 @@
     <a href="https://docs.arize.com/phoenix/">
         <img src="https://img.shields.io/static/v1?message=Docs&logo=data:image/png;base64,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&labelColor=grey&color=blue&logoColor=white&label=%20"/>
     </a>
-    <a target="_blank" href="https://join.slack.com/t/arize-ai/shared_invite/zt-1px8dcmlf-fmThhDFD_V_48oU7ALan4Q">
+    <a target="_blank" href="https://join.slack.com/t/arize-ai/shared_invite/zt-22vj03k4k-MlrNEwv5WeswapTs0kNCBw">
         <img src="https://img.shields.io/static/v1?message=Community&logo=slack&labelColor=grey&color=blue&logoColor=white&label=%20"/>
     </a>
     <a target="_blank" href="https://twitter.com/ArizePhoenix">
@@ -113,7 +113,7 @@ For details about tracing integrations and example applications, see the [OpenIn
 
 Join our community to connect with thousands of AI builders.
 
-- 🌍 Join our [Slack community](https://join.slack.com/t/arize-ai/shared_invite/zt-1px8dcmlf-fmThhDFD_V_48oU7ALan4Q).
+- 🌍 Join our [Slack community](https://join.slack.com/t/arize-ai/shared_invite/zt-22vj03k4k-MlrNEwv5WeswapTs0kNCBw).
 - 📚 Read our [documentation](https://docs.arize.com/phoenix).
 - 💡 Ask questions and provide feedback in the _#phoenix-support_ channel.
 - 🌟 Leave a star on our [GitHub](https://github.com/Arize-ai/phoenix).

{arize_phoenix-5.7.0 → arize_phoenix-5.8.0}/pyproject.toml

@@ -367,7 +367,6 @@ exclude = [
   "scripts/",
   "sdist/",
   "src/phoenix/evals/",
-  "tests/",
   "tutorials/",
   "js/",
 ]

{arize_phoenix-5.7.0 → arize_phoenix-5.8.0}/src/phoenix/config.py

@@ -146,9 +146,10 @@ ENV_PHOENIX_SMTP_VALIDATE_CERTS = "PHOENIX_SMTP_VALIDATE_CERTS"
 Whether to validate SMTP server certificates. Defaults to true.
 """
 
-# FastAPI and GQL extension settings
+# API extension settings
 ENV_PHOENIX_FASTAPI_MIDDLEWARE_PATHS = "PHOENIX_FASTAPI_MIDDLEWARE_PATHS"
 ENV_PHOENIX_GQL_EXTENSION_PATHS = "PHOENIX_GQL_EXTENSION_PATHS"
+ENV_PHOENIX_GRPC_INTERCEPTOR_PATHS = "PHOENIX_GRPC_INTERCEPTOR_PATHS"
 
 
 def server_instrumentation_is_enabled() -> bool:
@@ -656,7 +657,7 @@ def get_env_db_logging_level() -> int:
 
 
 def get_env_fastapi_middleware_paths() -> list[tuple[str, str]]:
-    env_value = os.getenv("PHOENIX_FASTAPI_MIDDLEWARE_PATHS", "")
+    env_value = os.getenv(ENV_PHOENIX_FASTAPI_MIDDLEWARE_PATHS, "")
     paths = []
     for entry in env_value.split(","):
         entry = entry.strip()
@@ -671,7 +672,7 @@ def get_env_fastapi_middleware_paths() -> list[tuple[str, str]]:
 
 
 def get_env_gql_extension_paths() -> list[tuple[str, str]]:
-    env_value = os.getenv("PHOENIX_GQL_EXTENSION_PATHS", "")
+    env_value = os.getenv(ENV_PHOENIX_GQL_EXTENSION_PATHS, "")
     paths = []
     for entry in env_value.split(","):
         entry = entry.strip()
@@ -685,6 +686,21 @@ def get_env_gql_extension_paths() -> list[tuple[str, str]]:
     return paths
 
 
+def get_env_grpc_interceptor_paths() -> list[tuple[str, str]]:
+    env_value = os.getenv(ENV_PHOENIX_GRPC_INTERCEPTOR_PATHS, "")
+    paths = []
+    for entry in env_value.split(","):
+        entry = entry.strip()
+        if entry:
+            if ":" not in entry:
+                raise ValueError(
+                    f"Invalid interceptor entry '{entry}'. Expected format 'file_path:ClassName'."
+                )
+            file_path, object_name = entry.split(":", 1)
+            paths.append((file_path.strip(), object_name.strip()))
+    return paths
+
+
 def _get_logging_level(env_var: str, default_level: int) -> int:
     logging_level = os.getenv(env_var)
     if not logging_level:

{arize_phoenix-5.7.0 → arize_phoenix-5.8.0}/src/phoenix/server/api/helpers/playground_clients.py

@@ -1,13 +1,12 @@
+import asyncio
 import importlib.util
+import inspect
+import json
+import time
 from abc import ABC, abstractmethod
 from collections.abc import AsyncIterator, Callable, Iterator
-from typing import (
-    TYPE_CHECKING,
-    Any,
-    Mapping,
-    Optional,
-    Union,
-)
+from functools import wraps
+from typing import TYPE_CHECKING, Any, Hashable, Mapping, Optional, Union
 
 from openinference.instrumentation import safe_json_dumps
 from openinference.semconv.trace import SpanAttributes
@@ -15,10 +14,14 @@ from strawberry import UNSET
 from strawberry.scalars import JSON as JSONScalarType
 from typing_extensions import TypeAlias, assert_never
 
-from phoenix.server.api.helpers.playground_registry import (
-    PROVIDER_DEFAULT,
-    register_llm_client,
+from phoenix.evals.models.rate_limiters import (
+    AsyncCallable,
+    GenericType,
+    ParameterSpec,
+    RateLimiter,
+    RateLimitError,
 )
+from phoenix.server.api.helpers.playground_registry import PROVIDER_DEFAULT, register_llm_client
 from phoenix.server.api.input_types.GenerativeModelInput import GenerativeModelInput
 from phoenix.server.api.input_types.InvocationParameters import (
     BoundedFloatInvocationParameter,
@@ -42,16 +45,99 @@ from phoenix.server.api.types.GenerativeProvider import GenerativeProviderKey
 if TYPE_CHECKING:
     from anthropic.types import MessageParam
     from openai.types import CompletionUsage
-    from openai.types.chat import (
-        ChatCompletionMessageParam,
-        ChatCompletionMessageToolCallParam,
-    )
+    from openai.types.chat import ChatCompletionMessageParam, ChatCompletionMessageToolCallParam
 
 DependencyName: TypeAlias = str
 SetSpanAttributesFn: TypeAlias = Callable[[Mapping[str, Any]], None]
 ChatCompletionChunk: TypeAlias = Union[TextChunk, ToolCallChunk]
 
 
+class KeyedSingleton:
+    _instances: dict[Hashable, "KeyedSingleton"] = {}
+
+    def __new__(cls, *args: Any, **kwargs: Any) -> "KeyedSingleton":
+        if "singleton_key" in kwargs:
+            singleton_key = kwargs.pop("singleton_key")
+        elif args:
+            singleton_key = args[0]
+            args = args[1:]
+        else:
+            raise ValueError("singleton_key must be provided")
+
+        instance_key = (cls, singleton_key)
+        if instance_key not in cls._instances:
+            instance = super().__new__(cls)
+            cls._instances[instance_key] = instance
+        return cls._instances[instance_key]
+
+
+class PlaygroundRateLimiter(RateLimiter, KeyedSingleton):
+    """
+    A rate rate limiter class that will be instantiated once per `singleton_key`.
+    """
+
+    def __init__(self, singleton_key: Hashable, rate_limit_error: Optional[type[BaseException]]):
+        super().__init__(
+            rate_limit_error=rate_limit_error,
+            max_rate_limit_retries=3,
+            initial_per_second_request_rate=2.0,
+            maximum_per_second_request_rate=10.0,
+            enforcement_window_minutes=1,
+            rate_reduction_factor=0.5,
+            rate_increase_factor=0.01,
+            cooldown_seconds=5,
+            verbose=False,
+        )
+
+    # TODO: update the rate limiter class in phoenix.evals to support decorated sync functions
+    def _alimit(
+        self, fn: Callable[ParameterSpec, GenericType]
+    ) -> AsyncCallable[ParameterSpec, GenericType]:
+        @wraps(fn)
+        async def wrapper(*args: Any, **kwargs: Any) -> GenericType:
+            self._initialize_async_primitives()
+            assert self._rate_limit_handling_lock is not None and isinstance(
+                self._rate_limit_handling_lock, asyncio.Lock
+            )
+            assert self._rate_limit_handling is not None and isinstance(
+                self._rate_limit_handling, asyncio.Event
+            )
+            try:
+                try:
+                    await asyncio.wait_for(self._rate_limit_handling.wait(), 120)
+                except asyncio.TimeoutError:
+                    self._rate_limit_handling.set()  # Set the event as a failsafe
+                await self._throttler.async_wait_until_ready()
+                request_start_time = time.time()
+                if inspect.iscoroutinefunction(fn):
+                    return await fn(*args, **kwargs)  # type: ignore
+                else:
+                    return fn(*args, **kwargs)
+            except self._rate_limit_error:
+                async with self._rate_limit_handling_lock:
+                    self._rate_limit_handling.clear()  # prevent new requests from starting
+                    self._throttler.on_rate_limit_error(request_start_time, verbose=self._verbose)
+                    try:
+                        for _attempt in range(self._max_rate_limit_retries):
+                            try:
+                                request_start_time = time.time()
+                                await self._throttler.async_wait_until_ready()
+                                if inspect.iscoroutinefunction(fn):
+                                    return await fn(*args, **kwargs)  # type: ignore
+                                else:
+                                    return fn(*args, **kwargs)
+                            except self._rate_limit_error:
+                                self._throttler.on_rate_limit_error(
+                                    request_start_time, verbose=self._verbose
+                                )
+                                continue
+                    finally:
+                        self._rate_limit_handling.set()  # allow new requests to start
+            raise RateLimitError(f"Exceeded max ({self._max_rate_limit_retries}) retries")
+
+        return wrapper
+
+
 class PlaygroundStreamingClient(ABC):
     def __init__(
         self,
@@ -150,10 +236,12 @@ class OpenAIStreamingClient(PlaygroundStreamingClient):
         api_key: Optional[str] = None,
     ) -> None:
         from openai import AsyncOpenAI
+        from openai import RateLimitError as OpenAIRateLimitError
 
         super().__init__(model=model, api_key=api_key)
         self.client = AsyncOpenAI(api_key=api_key)
         self.model_name = model.name
+        self.rate_limiter = PlaygroundRateLimiter(model.provider_key, OpenAIRateLimitError)
 
     @classmethod
     def dependencies(cls) -> list[DependencyName]:
@@ -174,19 +262,16 @@ class OpenAIStreamingClient(PlaygroundStreamingClient):
                 invocation_name="max_tokens",
                 canonical_name=CanonicalParameterName.MAX_COMPLETION_TOKENS,
                 label="Max Tokens",
-                default_value=UNSET,
             ),
             BoundedFloatInvocationParameter(
                 invocation_name="frequency_penalty",
                 label="Frequency Penalty",
-                default_value=UNSET,
                 min_value=-2.0,
                 max_value=2.0,
             ),
             BoundedFloatInvocationParameter(
                 invocation_name="presence_penalty",
                 label="Presence Penalty",
-                default_value=UNSET,
                 min_value=-2.0,
                 max_value=2.0,
             ),
@@ -194,13 +279,11 @@ class OpenAIStreamingClient(PlaygroundStreamingClient):
                 invocation_name="stop",
                 canonical_name=CanonicalParameterName.STOP_SEQUENCES,
                 label="Stop Sequences",
-                default_value=UNSET,
             ),
             BoundedFloatInvocationParameter(
                 invocation_name="top_p",
                 canonical_name=CanonicalParameterName.TOP_P,
                 label="Top P",
-                default_value=UNSET,
                 min_value=0.0,
                 max_value=1.0,
             ),
@@ -208,20 +291,16 @@ class OpenAIStreamingClient(PlaygroundStreamingClient):
                 invocation_name="seed",
                 canonical_name=CanonicalParameterName.RANDOM_SEED,
                 label="Seed",
-                default_value=UNSET,
             ),
             JSONInvocationParameter(
                 invocation_name="tool_choice",
                 label="Tool Choice",
                 canonical_name=CanonicalParameterName.TOOL_CHOICE,
-                default_value=UNSET,
-                hidden=True,
             ),
             JSONInvocationParameter(
                 invocation_name="response_format",
                 label="Response Format",
                 canonical_name=CanonicalParameterName.RESPONSE_FORMAT,
-                default_value=UNSET,
             ),
         ]
 
@@ -240,7 +319,8 @@ class OpenAIStreamingClient(PlaygroundStreamingClient):
         openai_messages = [self.to_openai_chat_completion_param(*message) for message in messages]
         tool_call_ids: dict[int, str] = {}
         token_usage: Optional["CompletionUsage"] = None
-        async for chunk in await self.client.chat.completions.create(
+        throttled_create = self.rate_limiter.alimit(self.client.chat.completions.create)
+        async for chunk in await throttled_create(
             messages=openai_messages,
             model=self.model_name,
             stream=True,
@@ -370,20 +450,16 @@ class OpenAIO1StreamingClient(OpenAIStreamingClient):
                 invocation_name="max_completion_tokens",
                 canonical_name=CanonicalParameterName.MAX_COMPLETION_TOKENS,
                 label="Max Completion Tokens",
-                default_value=UNSET,
             ),
             IntInvocationParameter(
                 invocation_name="seed",
                 canonical_name=CanonicalParameterName.RANDOM_SEED,
                 label="Seed",
-                default_value=UNSET,
             ),
             JSONInvocationParameter(
                 invocation_name="tool_choice",
                 label="Tool Choice",
                 canonical_name=CanonicalParameterName.TOOL_CHOICE,
-                default_value=UNSET,
-                hidden=True,
             ),
         ]
 
@@ -409,7 +485,8 @@ class OpenAIO1StreamingClient(OpenAIStreamingClient):
 
         tool_call_ids: dict[int, str] = {}
 
-        response = await self.client.chat.completions.create(
+        throttled_create = self.rate_limiter.alimit(self.client.chat.completions.create)
+        response = await throttled_create(
             messages=openai_messages,
             model=self.model_name,
             tools=tools or NOT_GIVEN,
@@ -544,6 +621,7 @@ class AnthropicStreamingClient(PlaygroundStreamingClient):
         super().__init__(model=model, api_key=api_key)
         self.client = anthropic.AsyncAnthropic(api_key=api_key)
         self.model_name = model.name
+        self.rate_limiter = PlaygroundRateLimiter(model.provider_key, anthropic.RateLimitError)
 
     @classmethod
     def dependencies(cls) -> list[DependencyName]:
@@ -556,14 +634,12 @@ class AnthropicStreamingClient(PlaygroundStreamingClient):
                 invocation_name="max_tokens",
                 canonical_name=CanonicalParameterName.MAX_COMPLETION_TOKENS,
                 label="Max Tokens",
-                default_value=UNSET,
                 required=True,
             ),
             BoundedFloatInvocationParameter(
                 invocation_name="temperature",
                 canonical_name=CanonicalParameterName.TEMPERATURE,
                 label="Temperature",
-                default_value=UNSET,
                 min_value=0.0,
                 max_value=1.0,
             ),
@@ -571,13 +647,11 @@ class AnthropicStreamingClient(PlaygroundStreamingClient):
                 invocation_name="stop_sequences",
                 canonical_name=CanonicalParameterName.STOP_SEQUENCES,
                 label="Stop Sequences",
-                default_value=UNSET,
             ),
             BoundedFloatInvocationParameter(
                 invocation_name="top_p",
                 canonical_name=CanonicalParameterName.TOP_P,
                 label="Top P",
-                default_value=UNSET,
                 min_value=0.0,
                 max_value=1.0,
             ),
@@ -585,8 +659,6 @@ class AnthropicStreamingClient(PlaygroundStreamingClient):
                 invocation_name="tool_choice",
                 label="Tool Choice",
                 canonical_name=CanonicalParameterName.TOOL_CHOICE,
-                default_value=UNSET,
-                hidden=True,
             ),
         ]
 
@@ -608,9 +680,11 @@ class AnthropicStreamingClient(PlaygroundStreamingClient):
             "model": self.model_name,
             "system": system_prompt,
             "max_tokens": 1024,
+            "tools": tools,
             **invocation_parameters,
         }
-        async with self.client.messages.stream(**anthropic_params) as stream:
+        throttled_stream = self.rate_limiter._alimit(self.client.messages.stream)
+        async with await throttled_stream(**anthropic_params) as stream:
             async for event in stream:
                 if isinstance(event, anthropic_types.RawMessageStartEvent):
                     self._attributes.update(
@@ -622,6 +696,18 @@ class AnthropicStreamingClient(PlaygroundStreamingClient):
                     self._attributes.update(
                         {LLM_TOKEN_COUNT_COMPLETION: event.message.usage.output_tokens}
                     )
+                elif (
+                    isinstance(event, anthropic_streaming.ContentBlockStopEvent)
+                    and event.content_block.type == "tool_use"
+                ):
+                    tool_call_chunk = ToolCallChunk(
+                        id=event.content_block.id,
+                        function=FunctionCallChunk(
+                            name=event.content_block.name,
+                            arguments=json.dumps(event.content_block.input),
+                        ),
+                    )
+                    yield tool_call_chunk
                 elif isinstance(
                     event,
                     (
@@ -629,6 +715,7 @@ class AnthropicStreamingClient(PlaygroundStreamingClient):
                         anthropic_types.RawContentBlockDeltaEvent,
                         anthropic_types.RawMessageDeltaEvent,
                         anthropic_streaming.ContentBlockStopEvent,
+                        anthropic_streaming.InputJsonEvent,
                     ),
                 ):
                     # event types emitted by the stream that don't contain useful information