arize-phoenix 3.0.0__py3-none-any.whl → 3.0.2__py3-none-any.whl

{arize_phoenix-3.0.0.dist-info → arize_phoenix-3.0.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: arize-phoenix
-Version: 3.0.0
+Version: 3.0.2
 Summary: ML Observability in your notebook
 Project-URL: Documentation, https://docs.arize.com/phoenix/
 Project-URL: Issues, https://github.com/Arize-ai/phoenix/issues
@@ -50,14 +50,14 @@ Requires-Dist: hatch; extra == 'dev'
 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>=0.9.14; extra == 'dev'
+Requires-Dist: llama-index<0.10.0; extra == 'dev'
 Requires-Dist: nbqa; extra == 'dev'
 Requires-Dist: pandas-stubs<=2.0.2.230605; extra == 'dev'
 Requires-Dist: pre-commit; extra == 'dev'
-Requires-Dist: pytest; extra == 'dev'
 Requires-Dist: pytest-asyncio; extra == 'dev'
 Requires-Dist: pytest-cov; extra == 'dev'
 Requires-Dist: pytest-lazy-fixture; extra == 'dev'
+Requires-Dist: pytest==7.4.4; extra == 'dev'
 Requires-Dist: ruff==0.1.5; extra == 'dev'
 Requires-Dist: strawberry-graphql[debug-server]==0.208.2; extra == 'dev'
 Provides-Extra: experimental
@@ -223,9 +223,8 @@ session = px.launch_app()
 
 from phoenix.trace.langchain import OpenInferenceTracer, LangChainInstrumentor
 
-# If no exporter is specified, the tracer will export to the locally running Phoenix server
-tracer = OpenInferenceTracer()
-LangChainInstrumentor(tracer).instrument()
+# By default, the traces will be exported to the locally running Phoenix server.
+LangChainInstrumentor().instrument()
 
 # Initialize your LangChain application
 from langchain.chains import RetrievalQA

{arize_phoenix-3.0.0.dist-info → arize_phoenix-3.0.2.dist-info}/RECORD

@@ -1,10 +1,10 @@
 phoenix/__init__.py,sha256=MSamw7OMCIcCtrJdvJYMvQmC6O0uyehfPAXXELsllZM,1451
-phoenix/config.py,sha256=RbQw8AkVyI4SSo5CD520AjUNcwkDNOGZA6_ErE48R7A,3454
+phoenix/config.py,sha256=RT4UR5VH8JUAKQhfmlkpxq2kmBSA8G8MODrC3GXWyaI,3507
 phoenix/datetime_utils.py,sha256=D955QLrkgrrSdUM6NyqbCeAu2SMsjhR5rHVQEsVUdng,2773
 phoenix/exceptions.py,sha256=X5k9ipUDfwSCwZB-H5zFJLas86Gf9tAx0W4l5TZxp5k,108
 phoenix/py.typed,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
 phoenix/services.py,sha256=f6AeyKTuOpy9RCcTCjVH3gx5nYZhbTMFOuv1WSUOB5o,4992
-phoenix/version.py,sha256=EPmgXOdWKks5S__ZMH7Nu6xpAeVrZpfxaFy4pykuyeI,22
+phoenix/version.py,sha256=YKXrr5J7dV2n7ZhaRv0tigylRDtfOuvJC9Y4ouFZpzo,22
 phoenix/core/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 phoenix/core/embedding_dimension.py,sha256=zKGbcvwOXgLf-yrJBpQyKtd-LEOPRKHnUToyAU8Owis,87
 phoenix/core/evals.py,sha256=gJyqQzpud5YjtoY8h4pgXvHDsdubGfqmEewLuZHPPmQ,10224
@@ -23,7 +23,7 @@ phoenix/experimental/evals/__init__.py,sha256=q96YKLMt2GJD9zL8sjugvWx1INfw40Wa7E
 phoenix/experimental/evals/evaluators.py,sha256=r7fXrS-l4gn58SUhLAZSfY3P8lxysouSVJwHddrZJ_Q,15956
 phoenix/experimental/evals/retrievals.py,sha256=o3fqrsYbYZjyGj_jWkN_9VQVyXjLkDKDw5Ws7l8bwdI,3828
 phoenix/experimental/evals/functions/__init__.py,sha256=NNd0-_cmIopdV7vm3rspjfgM726qoQJ4DPq_vqbnaxQ,180
-phoenix/experimental/evals/functions/classify.py,sha256=6yCajPT9i98b4_2qYn9ZxGhdI3CLhfUSrEyUUcqQqmQ,19517
+phoenix/experimental/evals/functions/classify.py,sha256=wqsLtDVPckt4OYuyr1YzaIokqnhzfYf_bMhR5ythWww,18961
 phoenix/experimental/evals/functions/executor.py,sha256=bM7PI2rcPukQQzZ2rWqN_-Kfo_a935YJj0bh1Red8Ps,13406
 phoenix/experimental/evals/functions/generate.py,sha256=8LnnPAjBM9yxitdkaGZ67OabuDTOWBF3fvinJ_uCFRg,5584
 phoenix/experimental/evals/functions/processing.py,sha256=F4xtLsulLV4a8CkuLldRddsCim75dSTIShEJUYN6I6w,1823
@@ -141,7 +141,7 @@ phoenix/trace/errors.py,sha256=DbXSJnNErV7305tKv7pUWLD6jcVHJ6EBdSu4mZJ6IM4,112
 phoenix/trace/evaluation_conventions.py,sha256=t8jydM3U0-T5YpiQKRJ3tWdWGlHtzKyttYdw-ddvPOk,1048
 phoenix/trace/exporter.py,sha256=O-9Arn-S_B9Me-jy4Qa84y6lvxKNoa8pczrsamPl3_Q,4871
 phoenix/trace/fixtures.py,sha256=LokNedhbGYxpzXznteO4m5QehvNYjzvoh231-CMJQeY,7113
-phoenix/trace/otel.py,sha256=9oum5RPCsEZvKg41mEy8aKDcXHBwtR-P9eeqEXp-ts4,14642
+phoenix/trace/otel.py,sha256=lJoTvkCSFiPT4dFRLMZZgZfsOoTSATctW9lQk9J30ZI,14821
 phoenix/trace/schemas.py,sha256=QDBlM94faceTnloY_ZVo2BHWWnE5UGymefC0jmfj4ew,6011
 phoenix/trace/semantic_conventions.py,sha256=u6NG85ZhbreriZr8cqJaddldM_jUcew7JilszY7JUk8,4652
 phoenix/trace/span_evaluations.py,sha256=FvY9YRnKuYIzCa-H9P5SuDaI2DeqGnVCjNgclC2v3HA,12909
@@ -169,8 +169,8 @@ phoenix/trace/v1/evaluation_pb2.pyi,sha256=cCbbx06gwQmaH14s3J1X25TtaARh-k1abbxQd
 phoenix/utilities/__init__.py,sha256=3TVirVnjIGyaCFuJCqeZO4tjlzQ_chZgYM0itIwsEpE,656
 phoenix/utilities/error_handling.py,sha256=7b5rpGFj9EWZ8yrZK1IHvxB89suWk3lggDayUQcvZds,1946
 phoenix/utilities/logging.py,sha256=lDXd6EGaamBNcQxL4vP1au9-i_SXe0OraUDiJOcszSw,222
-arize_phoenix-3.0.0.dist-info/METADATA,sha256=wC0glS_s7OdY7E0MQnvztpSmPYm8p-Hp_gSih5pPsAo,28597
-arize_phoenix-3.0.0.dist-info/WHEEL,sha256=TJPnKdtrSue7xZ_AVGkp9YXcvDrobsjBds1du3Nx6dc,87
-arize_phoenix-3.0.0.dist-info/licenses/IP_NOTICE,sha256=JBqyyCYYxGDfzQ0TtsQgjts41IJoa-hiwDrBjCb9gHM,469
-arize_phoenix-3.0.0.dist-info/licenses/LICENSE,sha256=HFkW9REuMOkvKRACuwLPT0hRydHb3zNg-fdFt94td18,3794
-arize_phoenix-3.0.0.dist-info/RECORD,,
+arize_phoenix-3.0.2.dist-info/METADATA,sha256=80fEGOl8VUcS4B1STEVmDITkrkYytJtE6f9d7VmkyhI,28555
+arize_phoenix-3.0.2.dist-info/WHEEL,sha256=TJPnKdtrSue7xZ_AVGkp9YXcvDrobsjBds1du3Nx6dc,87
+arize_phoenix-3.0.2.dist-info/licenses/IP_NOTICE,sha256=JBqyyCYYxGDfzQ0TtsQgjts41IJoa-hiwDrBjCb9gHM,469
+arize_phoenix-3.0.2.dist-info/licenses/LICENSE,sha256=HFkW9REuMOkvKRACuwLPT0hRydHb3zNg-fdFt94td18,3794
+arize_phoenix-3.0.2.dist-info/RECORD,,

phoenix/config.py

@@ -12,7 +12,7 @@ ENV_PHOENIX_COLLECTOR_ENDPOINT = "PHOENIX_COLLECTOR_ENDPOINT"
 The endpoint traces and evals are sent to. This must be set if the Phoenix
 server is running on a remote instance.
 """
-ENV_WORKING_DIR = "PHOENIX_WORKING_DIR"
+ENV_PHOENIX_WORKING_DIR = "PHOENIX_WORKING_DIR"
 """
 The directory in which to save, load, and export datasets. This directory must
 be accessible by both the Phoenix server and the notebook environment.
@@ -45,7 +45,7 @@ def get_working_dir() -> Path:
     """
     Get the working directory for saving, loading, and exporting datasets.
     """
-    working_dir_str = os.getenv(ENV_WORKING_DIR)
+    working_dir_str = os.getenv(ENV_PHOENIX_WORKING_DIR)
     if working_dir_str is not None:
         return Path(working_dir_str)
     # Fall back to ~/.phoenix if PHOENIX_WORKING_DIR is not set
@@ -77,7 +77,9 @@ except Exception as e:
         f"⚠️ Failed to initialize the working directory at {WORKING_DIR} due to an error: {str(e)}"
     )
     print("⚠️ While phoenix will still run, you will not be able to save, load, or export data")
-    print("ℹ️ To change, set the `{ENV_WORKING_DIR}` environment variable before importing phoenix.")
+    print(
+        f"ℹ️ To change, set the `{ENV_PHOENIX_WORKING_DIR}` environment variable before importing phoenix."  # noqa: E501
+    )
 
 
 def get_exported_files(directory: Path) -> List[Path]:

phoenix/experimental/evals/functions/classify.py

@@ -3,6 +3,7 @@ from __future__ import annotations
 import logging
 import warnings
 from collections import defaultdict
+from itertools import product
 from typing import (
     Any,
     DefaultDict,
@@ -54,8 +55,7 @@ Label: TypeAlias = str
 Score: TypeAlias = Optional[float]
 Explanation: TypeAlias = Optional[str]
 Record: TypeAlias = Mapping[str, Any]
-EvaluatorIndex: TypeAlias = int
-RowIndex: TypeAlias = Any
+Index: TypeAlias = int
 
 # snapped_response, explanation, response
 ParsedLLMResponse: TypeAlias = Tuple[Optional[str], Optional[str], str]
@@ -343,8 +343,6 @@ def _get_contents_from_openinference_documents(documents: Iterable[Any]) -> List
 
 
 class RunEvalsPayload(NamedTuple):
-    evaluator_index: EvaluatorIndex
-    row_index: RowIndex
     evaluator: LLMEvaluator
     record: Record
 
@@ -404,23 +402,21 @@ def run_evals(
 
     async def _arun_eval(
         payload: RunEvalsPayload,
-    ) -> Tuple[EvaluatorIndex, RowIndex, Label, Score, Explanation]:
-        label, score, explanation = await payload.evaluator.aevaluate(
+    ) -> Tuple[Label, Score, Explanation]:
+        return await payload.evaluator.aevaluate(
             payload.record,
             provide_explanation=provide_explanation,
             use_function_calling_if_available=use_function_calling_if_available,
         )
-        return payload.evaluator_index, payload.row_index, label, score, explanation
 
     def _run_eval(
         payload: RunEvalsPayload,
-    ) -> Tuple[EvaluatorIndex, RowIndex, Label, Score, Explanation]:
-        label, score, explanation = payload.evaluator.evaluate(
+    ) -> Tuple[Label, Score, Explanation]:
+        return payload.evaluator.evaluate(
             payload.record,
             provide_explanation=provide_explanation,
             use_function_calling_if_available=use_function_calling_if_available,
         )
-        return payload.evaluator_index, payload.row_index, label, score, explanation
 
     executor = get_executor_on_sync_context(
         _run_eval,
@@ -428,24 +424,20 @@ def run_evals(
         concurrency=concurrency,
         tqdm_bar_format=get_tqdm_progress_bar_formatter("run_evals"),
         exit_on_error=True,
-        fallback_return_value=(None, None),
+        fallback_return_value=(None, None, None),
     )
+
+    total_records = len(dataframe)
     payloads = [
-        RunEvalsPayload(
-            evaluator_index=evaluator_index,
-            row_index=row_index,
-            evaluator=evaluator,
-            record=row.to_dict(),
-        )
-        # use the position of the row rather than the dataframe index, which is used
-        # to ensure the output dataframe has the same row order as the input dataframe
-        for row_index, (_, row) in enumerate(dataframe.iterrows())
-        for evaluator_index, evaluator in enumerate(evaluators)
+        RunEvalsPayload(evaluator=evaluator, record=row)
+        for evaluator, (_, row) in product(evaluators, dataframe.iterrows())
     ]
-    eval_results: List[DefaultDict[RowIndex, Dict[ColumnName, Union[Label, Explanation]]]] = [
+    eval_results: List[DefaultDict[Index, Dict[ColumnName, Union[Label, Explanation]]]] = [
         defaultdict(dict) for _ in range(len(evaluators))
     ]
-    for evaluator_index, row_index, label, score, explanation in executor.run(payloads):
+    for index, (label, score, explanation) in enumerate(executor.run(payloads)):
+        evaluator_index = index // total_records
+        row_index = index % total_records
         eval_results[evaluator_index][row_index]["label"] = label
         eval_results[evaluator_index][row_index]["score"] = score
         if provide_explanation:

phoenix/trace/otel.py

@@ -19,6 +19,7 @@ from typing import (
     cast,
 )
 
+import numpy as np
 import opentelemetry.proto.trace.v1.trace_pb2 as otlp
 from opentelemetry.proto.common.v1.common_pb2 import AnyValue, ArrayValue, KeyValue
 from opentelemetry.util.types import Attributes, AttributeValue
@@ -320,7 +321,11 @@ def encode(span: Span) -> otlp.Span:
                 attributes[key] = json.dumps(value)
             else:
                 attributes.update(_flatten_mapping(value, key))
-        elif not isinstance(value, str) and isinstance(value, Sequence) and _has_mapping(value):
+        elif (
+            not isinstance(value, str)
+            and (isinstance(value, Sequence) or isinstance(value, np.ndarray))
+            and _has_mapping(value)
+        ):
             attributes.pop(key, None)
             attributes.update(_flatten_sequence(value, key))
 
@@ -413,6 +418,8 @@ def _encode_attributes(attributes: Attributes) -> Iterator[KeyValue]:
     if not attributes:
         return
     for key, value in attributes.items():
+        if isinstance(value, np.ndarray):
+            value = value.tolist()
         yield KeyValue(key=key, value=_encode_value(value))
 
 

phoenix/version.py

@@ -1 +1 @@
-__version__ = "3.0.0"
+__version__ = "3.0.2"