arize-phoenix 11.23.1__py3-none-any.whl → 12.28.1__py3-none-any.whl

{arize_phoenix-11.23.1.dist-info → arize_phoenix-12.28.1.dist-info}/WHEEL

@@ -1,4 +1,4 @@
 Wheel-Version: 1.0
-Generator: hatchling 1.27.0
+Generator: hatchling 1.28.0
 Root-Is-Purelib: true
 Tag: py3-none-any

{arize_phoenix-11.23.1.dist-info → arize_phoenix-12.28.1.dist-info}/licenses/IP_NOTICE

@@ -1,4 +1,4 @@
-Copyright 2023 Arize AI, Inc. All Rights Reserved.
+Copyright 2025 Arize AI, Inc. All Rights Reserved.
 
 Portions of this code are patent protected by one or more of U.S. Patent Nos. 11,315,043 and 11,615,345. The patent protection includes, but is not limited to, functionality enabled by executing the portions of the code, a device or a system executing the portions of the code, a memory storing the portions of the code, etc.
 

phoenix/__generated__/classification_evaluator_configs/__init__.py

@@ -0,0 +1,20 @@
+# This file is generated. Do not edit by hand.
+
+from ._document_relevance_classification_evaluator_config import (
+    DOCUMENT_RELEVANCE_CLASSIFICATION_EVALUATOR_CONFIG,
+)
+from ._hallucination_classification_evaluator_config import (
+    HALLUCINATION_CLASSIFICATION_EVALUATOR_CONFIG,
+)
+from ._models import ClassificationEvaluatorConfig, PromptMessage
+from ._tool_selection_classification_evaluator_config import (
+    TOOL_SELECTION_CLASSIFICATION_EVALUATOR_CONFIG,
+)
+
+__all__ = [
+    "ClassificationEvaluatorConfig",
+    "PromptMessage",
+    "DOCUMENT_RELEVANCE_CLASSIFICATION_EVALUATOR_CONFIG",
+    "HALLUCINATION_CLASSIFICATION_EVALUATOR_CONFIG",
+    "TOOL_SELECTION_CLASSIFICATION_EVALUATOR_CONFIG",
+]

phoenix/__generated__/classification_evaluator_configs/_document_relevance_classification_evaluator_config.py

@@ -0,0 +1,17 @@
+# This file is generated. Do not edit by hand.
+# ruff: noqa: E501
+
+from ._models import ClassificationEvaluatorConfig, PromptMessage
+
+DOCUMENT_RELEVANCE_CLASSIFICATION_EVALUATOR_CONFIG = ClassificationEvaluatorConfig(
+    name="document_relevance",
+    description="A specialized evaluator for determining document relevance to a given question.",
+    optimization_direction="maximize",
+    messages=[
+        PromptMessage(
+            role="user",
+            content='You are comparing a document to a question and trying to determine if the document text contains information relevant to answering the question. Here is the data:\n\n<data>\n<question>\n{{input}}\n</question>\n<document_text>\n{{document_text}}\n</document_text>\n</data>\n\nCompare the question above to the document text. You must determine whether the document text contains information that can answer the question. Please focus on whether the very specific question can be answered by the information in the document text. Your response must be either "relevant" or "unrelated". "unrelated" means that the document text does not contain an answer to the question. "relevant" means the document text contains an answer to the question.',
+        )
+    ],
+    choices={"relevant": 1.0, "unrelated": 0.0},
+)

phoenix/__generated__/classification_evaluator_configs/_hallucination_classification_evaluator_config.py

@@ -0,0 +1,17 @@
+# This file is generated. Do not edit by hand.
+# ruff: noqa: E501
+
+from ._models import ClassificationEvaluatorConfig, PromptMessage
+
+HALLUCINATION_CLASSIFICATION_EVALUATOR_CONFIG = ClassificationEvaluatorConfig(
+    name="hallucination",
+    description="A specialized evaluator for detecting hallucinations in grounded LLM responses.",
+    optimization_direction="minimize",
+    messages=[
+        PromptMessage(
+            role="user",
+            content="In this task, you will be presented with a query, some context and a response. The response is generated to the question based on the context. The response may contain false information. You must use the context to determine if the response to the question contains false information, if the response is a hallucination of facts. Your objective is to determine whether the response text contains factual information and is not a hallucination. A 'hallucination' refers to a response that is not based on the context or assumes information that is not available in the context. Your response should be a single word: either 'factual' or 'hallucinated', and it should not include any other text or characters. 'hallucinated' indicates that the response provides factually inaccurate information to the query based on the context. 'factual' indicates that the response to the question is correct relative to the context, and does not contain made up information. Please read the query and context carefully before determining your response.\n\n<data>\n<query>\n{{input}}\n</query>\n<context>\n{{context}}\n</context>\n<response>\n{{output}}\n</response>\n</data>\n\nIs the response above factual or hallucinated based on the query and context?",
+        )
+    ],
+    choices={"hallucinated": 1.0, "factual": 0.0},
+)

phoenix/__generated__/classification_evaluator_configs/_models.py

@@ -0,0 +1,18 @@
+# This file is generated. Do not edit by hand.
+
+from typing import Literal
+
+from pydantic import BaseModel
+
+
+class PromptMessage(BaseModel):
+    role: Literal["user"]
+    content: str
+
+
+class ClassificationEvaluatorConfig(BaseModel):
+    name: str
+    description: str
+    optimization_direction: Literal["minimize", "maximize"]
+    messages: list[PromptMessage]
+    choices: dict[str, float]

phoenix/__generated__/classification_evaluator_configs/_tool_selection_classification_evaluator_config.py

@@ -0,0 +1,17 @@
+# This file is generated. Do not edit by hand.
+# ruff: noqa: E501
+
+from ._models import ClassificationEvaluatorConfig, PromptMessage
+
+TOOL_SELECTION_CLASSIFICATION_EVALUATOR_CONFIG = ClassificationEvaluatorConfig(
+    name="tool_selection",
+    description="For determining if the correct tool was selected for a given context. Requires conversation context, a list of available tools, and the LLM's tool selections.",
+    optimization_direction="maximize",
+    messages=[
+        PromptMessage(
+            role="user",
+            content="You are an impartial judge evaluating an LLM's tool-calling behavior, specifically whether the LLM selected the most appropriate tool or tools for the task.\nYour task: Determine whether the LLM's tool selection was correct or incorrect based on: - The conversation context - The available tools - The LLM's tool invocation(s)\nCriteria Return \"correct\" only when ALL of the following are true: - The LLM chose the best available tool for the user query OR correctly avoided tools if none were needed. - The tool name exists in the available tools list. - The tool is allowed and safe to call. - The LLM selected the correct number of tools for the task.\nReturn \"incorrect\" if ANY of the following are true: - The LLM used a hallucinated or nonexistent tool. - The LLM selected a tool when none was needed. - The LLM did not use a tool when one was required. - The LLM chose a suboptimal or irrelevant tool. - The LLM selected an unsafe or not-permitted tool. - The tool name does not appear in the available tools list.\nBefore providing your final judgment, explain your reasoning and consider: - What does the input context require? - Can this be answered without tools, or is a tool necessary? - If a tool was selected, does it exist in the available tools? - Does the selected tool's description match the user's needs? - Is the selection safe and appropriate? - Is there a better tool available that should have been chosen instead?\n<data> <context> {{input}} </context>\n<available_tools> {{available_tools}} </available_tools>\n<tool_selection> {{tool_selection}} </tool_selection> </data>\nGiven the above data, is the tool selection correct or incorrect?",
+        )
+    ],
+    choices={"correct": 1.0, "incorrect": 0.0},
+)

phoenix/__init__.py

@@ -87,8 +87,9 @@ class PhoenixTraceOpenAILoader(Loader):
             "https://arize.com/docs/phoenix/tracing/integrations-tracing/openai"
             "\n\n"
             "Example usage:\n\n"
+            "pip install openinference-instrumentation-openai\n\n"
             "```python\n"
-            "from phoenix.otel register\n"
+            "from phoenix.otel import register\n"
             "from openinference.instrumentation.openai import OpenAIInstrumentor\n\n"
             "tracer_provider = register()\n"
             "OpenAIInstrumentor().instrument(tracer_provider=tracer_provider)\n"

phoenix/auth.py

@@ -2,7 +2,7 @@ from __future__ import annotations
 
 import re
 from dataclasses import dataclass
-from datetime import datetime, timedelta
+from datetime import datetime, timedelta, timezone
 from enum import Enum, auto
 from hashlib import pbkdf2_hmac
 from typing import Any, Literal, Optional, Protocol
@@ -129,6 +129,18 @@ def set_oauth2_nonce_cookie(
     )
 
 
+def set_oauth2_code_verifier_cookie(
+    *, response: ResponseType, code_verifier: str, max_age: timedelta
+) -> ResponseType:
+    return _set_cookie(
+        response=response,
+        cookie_name=PHOENIX_OAUTH2_CODE_VERIFIER_COOKIE_NAME,
+        cookie_max_age=max_age,
+        samesite="lax",
+        value=code_verifier,
+    )
+
+
 def _set_cookie(
     *,
     response: ResponseType,
@@ -169,6 +181,11 @@ def delete_oauth2_nonce_cookie(response: ResponseType) -> ResponseType:
     return response
 
 
+def delete_oauth2_code_verifier_cookie(response: ResponseType) -> ResponseType:
+    response.delete_cookie(key=PHOENIX_OAUTH2_CODE_VERIFIER_COOKIE_NAME)
+    return response
+
+
 @dataclass(frozen=True)
 class _PasswordRequirements:
     """
@@ -270,6 +287,8 @@ PHOENIX_OAUTH2_STATE_COOKIE_NAME = "phoenix-oauth2-state"
 """The name of the cookie that stores the state used for the OAuth2 authorization code flow."""
 PHOENIX_OAUTH2_NONCE_COOKIE_NAME = "phoenix-oauth2-nonce"
 """The name of the cookie that stores the nonce used for the OAuth2 authorization code flow."""
+PHOENIX_OAUTH2_CODE_VERIFIER_COOKIE_NAME = "phoenix-oauth2-code-verifier"
+"""The name of the cookie that stores the PKCE code verifier for OAuth2."""
 DEFAULT_OAUTH2_LOGIN_EXPIRY_MINUTES = 15
 """
 The default amount of time in minutes that can elapse between the initial
@@ -312,7 +331,13 @@ class ClaimSet:
 
     @property
     def status(self) -> ClaimSetStatus:
-        if self.expiration_time and self.expiration_time.timestamp() < datetime.now().timestamp():
+        # Per JWT RFC 7519 Section 4.1.4, the expiration time identifies the time
+        # "on or after which" the JWT must not be accepted. Use <= for inclusive check.
+        # https://datatracker.ietf.org/doc/html/rfc7519#section-4.1.4
+        if (
+            self.expiration_time
+            and self.expiration_time.timestamp() <= datetime.now(timezone.utc).timestamp()
+        ):
             return ClaimSetStatus.EXPIRED
         if self.token_id is not None and self.subject is not None:
             return ClaimSetStatus.VALID