arize 8.0.0a18__tar.gz → 8.0.0a19__tar.gz

{arize-8.0.0a18 → arize-8.0.0a19}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: arize
-Version: 8.0.0a18
+Version: 8.0.0a19
 Summary: A helper library to interact with Arize AI APIs
 Project-URL: Homepage, https://arize.com
 Project-URL: Documentation, https://docs.arize.com/arize
@@ -672,7 +672,7 @@ The default behavior of Arize's logs is: enabled, `INFO` level, and not structur
 
 ## Caching
 
-When downloading big segments of data from Arize, such as a `Dataset` with all of its examples, the SDK will cache the file in `parquet` format under `~/.arize/datasets/dataset_<updated_at_timestamp>.parquet`.
+When downloading big segments of data from Arize, such as a `Dataset` with all of its examples, the SDK will cache the file in `parquet` format under `~/.arize/cache/datasets/dataset_<updated_at_timestamp>.parquet`.
 
 ### In Code
 
@@ -700,7 +700,7 @@ os.environ["ARIZE_DIRECTORY"] = "~/.arize"
 
 ### Clean the cache
 
-To clean the cache you can directly `rm` the files or directory.
+To clean the cache you can directly `rm` the files or directory. In addition, the client has the option to help with that as well using `client.clear_cache()`, which will delete the `cache/` directory inside the _arize directory_ (defaults to `~/.arize`).
 
 # Community
 

{arize-8.0.0a18 → arize-8.0.0a19}/README.md

@@ -608,7 +608,7 @@ The default behavior of Arize's logs is: enabled, `INFO` level, and not structur
 
 ## Caching
 
-When downloading big segments of data from Arize, such as a `Dataset` with all of its examples, the SDK will cache the file in `parquet` format under `~/.arize/datasets/dataset_<updated_at_timestamp>.parquet`.
+When downloading big segments of data from Arize, such as a `Dataset` with all of its examples, the SDK will cache the file in `parquet` format under `~/.arize/cache/datasets/dataset_<updated_at_timestamp>.parquet`.
 
 ### In Code
 
@@ -636,7 +636,7 @@ os.environ["ARIZE_DIRECTORY"] = "~/.arize"
 
 ### Clean the cache
 
-To clean the cache you can directly `rm` the files or directory.
+To clean the cache you can directly `rm` the files or directory. In addition, the client has the option to help with that as well using `client.clear_cache()`, which will delete the `cache/` directory inside the _arize directory_ (defaults to `~/.arize`).
 
 # Community
 

{arize-8.0.0a18 → arize-8.0.0a19}/src/arize/client.py

@@ -1,5 +1,8 @@
 from __future__ import annotations
 
+import logging
+import shutil
+from pathlib import Path
 from typing import TYPE_CHECKING
 
 from arize._lazy import LazySubclientsMixin
@@ -11,17 +14,16 @@ if TYPE_CHECKING:
     from arize.models.client import MLModelsClient
     from arize.spans.client import SpansClient
 
+logger = logging.getLogger(__name__)
 
+
+# TODO(Kiko): test caching in colab environment
 # TODO(Kiko): models need to follow resource first pattern
 # -  models.DatasetsList200Response
 # -  models.DatasetsListExamples200Response
 # -  models.ExperimentsList200Response
 # -  models.ExperimentsRunsList200Response
-# TODO(Kiko): Root client should have option to clear caches
-# TODO(Kiko): Document caching behavior
-# TODO(Kiko): Force keyword arguments
 # TODO(Kiko): Protobuf versioning is too old
-# TODO(Kiko): Make sure the client has same options as SDKConfiguration
 # TODO(Kiko): It does not make any sense to require space ID in run_experiment, dataset ID should suffice
 # TODO(Kiko): Should probably wrap every single method of gen clients so that we can add nice docstrings
 # TODO(Kiko): Add flight max_chunksize opt to write_table. In config?
@@ -33,6 +35,8 @@ if TYPE_CHECKING:
 # TODO(Kiko): Go over **every file** and do not import anything at runtime, use `if TYPE_CHECKING`
 # with `from __future__ import annotations` (must include for Python < 3.11)
 # TODO(Kiko): Go over docstrings
+# TODO(Kiko): Go headers on each logging call
+# TODO(Kiko): Make sure the client has same options as SDKConfiguration
 class ArizeClient(LazySubclientsMixin):
     """
     Root client for the Arize SDK. All parameters are optional. If not provided, they will be read
@@ -114,9 +118,12 @@ class ArizeClient(LazySubclientsMixin):
 
     def __init__(
         self,
+        *,
         api_key: str | None = None,
         api_host: str | None = None,
         api_scheme: str | None = None,
+        otlp_host: str | None = None,
+        otlp_scheme: str | None = None,
         flight_server_host: str | None = None,
         flight_server_port: int | None = None,
         flight_scheme: str | None = None,
@@ -124,6 +131,9 @@ class ArizeClient(LazySubclientsMixin):
         request_verify: bool | None = None,
         stream_max_workers: int | None = None,
         stream_max_queue_bound: int | None = None,
+        max_http_payload_size_mb: float | None = None,
+        arize_direcory: str | None = None,
+        enable_caching: bool | None = None,
     ):
         cfg_kwargs: dict = {}
         if api_key is not None:
@@ -132,6 +142,10 @@ class ArizeClient(LazySubclientsMixin):
             cfg_kwargs["api_host"] = api_host
         if api_scheme is not None:
             cfg_kwargs["api_scheme"] = api_scheme
+        if otlp_host is not None:
+            cfg_kwargs["otlp_host"] = otlp_host
+        if otlp_scheme is not None:
+            cfg_kwargs["otlp_scheme"] = otlp_scheme
         if flight_server_host is not None:
             cfg_kwargs["flight_server_host"] = flight_server_host
         if flight_server_port is not None:
@@ -146,6 +160,12 @@ class ArizeClient(LazySubclientsMixin):
             cfg_kwargs["stream_max_workers"] = stream_max_workers
         if stream_max_queue_bound is not None:
             cfg_kwargs["stream_max_queue_bound"] = stream_max_queue_bound
+        if max_http_payload_size_mb is not None:
+            cfg_kwargs["max_http_payload_size_mb"] = max_http_payload_size_mb
+        if arize_direcory is not None:
+            cfg_kwargs["arize_direcory"] = arize_direcory
+        if enable_caching is not None:
+            cfg_kwargs["enable_caching"] = enable_caching
 
         # Only the explicitly provided fields are passed; the rest use
         # SDKConfiguration’s default factories / defaults.
@@ -182,3 +202,17 @@ class ArizeClient(LazySubclientsMixin):
         lines.append("  }")
         lines.append(")")
         return "\n".join(lines)
+
+    def clear_cache(self) -> None:
+        p = Path(self.sdk_config.cache_dir)
+
+        if not p.exists():
+            logger.warning(
+                f"Cache directory does not exist at {p}, nothing to clear"
+            )
+        if not p.is_dir():
+            logger.error(f"Cache path is not a directory: {p}")
+            raise NotADirectoryError(f"Not a directory: {p}")
+
+        logger.info(f"Clearing cache directory at {p}")
+        shutil.rmtree(p)

{arize-8.0.0a18 → arize-8.0.0a19}/src/arize/config.py

@@ -3,6 +3,7 @@ import os
 import sys
 import threading
 from dataclasses import dataclass, field, fields
+from pathlib import Path
 from typing import Any, Dict
 
 from arize.constants.config import (
@@ -145,6 +146,8 @@ class SDKConfiguration:
     api_key: str = field(default_factory=_api_key_factory)
     api_host: str = field(default_factory=_api_host_factory)
     api_scheme: str = field(default_factory=_api_scheme_factory)
+    otlp_host: str = field(default_factory=_otlp_host_factory)
+    otlp_scheme: str = field(default_factory=_otlp_scheme_factory)
     flight_server_host: str = field(default_factory=_flight_host_factory)
     flight_server_port: int = field(default_factory=_flight_port_factory)
     flight_scheme: str = field(default_factory=_flight_scheme_factory)
@@ -154,8 +157,6 @@ class SDKConfiguration:
     stream_max_queue_bound: int = field(
         default_factory=_stream_max_queue_bound_factory
     )
-    otlp_host: str = field(default_factory=_otlp_host_factory)
-    otlp_scheme: str = field(default_factory=_otlp_scheme_factory)
     max_http_payload_size_mb: float = field(
         default_factory=_max_http_payload_size_mb_factory
     )
@@ -174,6 +175,10 @@ class SDKConfiguration:
         if not self.api_key:
             raise MissingAPIKeyError()
 
+    @property
+    def cache_dir(self) -> str:
+        return str(Path(self.arize_direcory) / "cache")
+
     @property
     def api_url(self) -> str:
         return _endpoint(self.api_scheme, self.api_host)

{arize-8.0.0a18 → arize-8.0.0a19}/src/arize/datasets/client.py

@@ -25,7 +25,7 @@ logger = logging.getLogger(__name__)
 
 
 class DatasetsClient:
-    def __init__(self, sdk_config: SDKConfiguration):
+    def __init__(self, *, sdk_config: SDKConfiguration):
         self._sdk_config = sdk_config
 
         # Import at runtime so it’s still lazy and extras-gated by the parent
@@ -45,6 +45,7 @@ class DatasetsClient:
 
     def _create_dataset(
         self,
+        *,
         name: str,
         space_id: str,
         examples: List[Dict[str, Any]] | pd.DataFrame,
@@ -154,6 +155,7 @@ class DatasetsClient:
 
     def _list_examples(
         self,
+        *,
         dataset_id: str,
         dataset_version_id: str = "",
         limit: int = 100,
@@ -176,7 +178,7 @@ class DatasetsClient:
         # try to load dataset from cache
         if self._sdk_config.enable_caching:
             dataset_df = load_cached_resource(
-                cache_dir=self._sdk_config.arize_direcory,
+                cache_dir=self._sdk_config.cache_dir,
                 resource="dataset",
                 resource_id=dataset_id,
                 resource_updated_at=dataset_updated_at,
@@ -213,7 +215,7 @@ class DatasetsClient:
 
         # cache dataset for future use
         cache_resource(
-            cache_dir=self._sdk_config.arize_direcory,
+            cache_dir=self._sdk_config.cache_dir,
             resource="dataset",
             resource_id=dataset_id,
             resource_updated_at=dataset_updated_at,

{arize-8.0.0a18 → arize-8.0.0a19}/src/arize/experiments/client.py

@@ -50,7 +50,7 @@ logger = logging.getLogger(__name__)
 
 
 class ExperimentsClient:
-    def __init__(self, sdk_config: SDKConfiguration):
+    def __init__(self, *, sdk_config: SDKConfiguration):
         self._sdk_config = sdk_config
         from arize._generated import api_client as gen
 
@@ -72,6 +72,7 @@ class ExperimentsClient:
 
     def _run_experiment(
         self,
+        *,
         name: str,
         dataset_id: str,
         task: ExperimentTask,
@@ -158,7 +159,7 @@ class ExperimentsClient:
             # try to load dataset from cache
             if self._sdk_config.enable_caching:
                 dataset_df = load_cached_resource(
-                    cache_dir=self._sdk_config.arize_direcory,
+                    cache_dir=self._sdk_config.cache_dir,
                     resource="dataset",
                     resource_id=dataset_id,
                     resource_updated_at=dataset_updated_at,
@@ -189,7 +190,7 @@ class ExperimentsClient:
 
             # cache dataset for future use
             cache_resource(
-                cache_dir=self._sdk_config.arize_direcory,
+                cache_dir=self._sdk_config.cache_dir,
                 resource="dataset",
                 resource_id=dataset_id,
                 resource_updated_at=dataset_updated_at,
@@ -268,6 +269,7 @@ class ExperimentsClient:
 
     def _create_experiment(
         self,
+        *,
         name: str,
         dataset_id: str,
         experiment_runs: List[Dict[str, Any]] | pd.DataFrame,
@@ -377,6 +379,7 @@ class ExperimentsClient:
 
     def _list_runs(
         self,
+        *,
         experiment_id: str,
         limit: int = 100,
         all: bool = False,
@@ -400,7 +403,7 @@ class ExperimentsClient:
         # try to load dataset from cache
         if self._sdk_config.enable_caching:
             experiment_df = load_cached_resource(
-                cache_dir=self._sdk_config.arize_direcory,
+                cache_dir=self._sdk_config.cache_dir,
                 resource="experiment",
                 resource_id=experiment_id,
                 resource_updated_at=experiment_updated_at,
@@ -436,7 +439,7 @@ class ExperimentsClient:
 
         # cache dataset for future use
         cache_resource(
-            cache_dir=self._sdk_config.arize_direcory,
+            cache_dir=self._sdk_config.cache_dir,
             resource="dataset",
             resource_id=experiment_id,
             resource_updated_at=experiment_updated_at,
@@ -558,9 +561,3 @@ def _get_tracer_resource(
         trace.set_tracer_provider(tracer_provider)
 
     return tracer_provider.get_tracer(__name__), resource
-
-
-def _dataset_cache_key(dataset_id: str, dataset_updated_at: str | None) -> str:
-    # include updated_at if present to produce a new key when dataset changes
-    key_src = f"{dataset_id}:{dataset_updated_at or ''}"
-    return hashlib.sha256(key_src.encode("utf-8")).hexdigest()

{arize-8.0.0a18 → arize-8.0.0a19}/src/arize/models/client.py

@@ -96,7 +96,7 @@ _MIMIC_EXTRA = "mimic-explainer"
 
 
 class MLModelsClient:
-    def __init__(self, sdk_config: SDKConfiguration):
+    def __init__(self, *, sdk_config: SDKConfiguration):
         self._sdk_config = sdk_config
 
         # internal cache for the futures session
@@ -104,6 +104,7 @@ class MLModelsClient:
 
     def log_stream(
         self,
+        *,
         space_id: str,
         model_name: str,
         model_type: ModelTypes,
@@ -455,6 +456,7 @@ class MLModelsClient:
     # TODO(Kiko): Handle sync argument
     def log_batch(
         self,
+        *,
         space_id: str,
         model_name: str,
         model_type: ModelTypes,
@@ -666,6 +668,7 @@ class MLModelsClient:
 
     def export_to_df(
         self,
+        *,
         space_id: str,
         model_name: str,
         environment: Environments,
@@ -711,6 +714,7 @@ class MLModelsClient:
 
     def export_to_parquet(
         self,
+        *,
         space_id: str,
         model_name: str,
         environment: Environments,

{arize-8.0.0a18 → arize-8.0.0a19}/src/arize/spans/client.py

@@ -47,11 +47,12 @@ logger = logging.getLogger(__name__)
 
 
 class SpansClient:
-    def __init__(self, sdk_config: SDKConfiguration):
+    def __init__(self, *, sdk_config: SDKConfiguration):
         self._sdk_config = sdk_config
 
     def log(
         self,
+        *,
         space_id: str,
         project_name: str,
         dataframe: pd.DataFrame,
@@ -253,6 +254,7 @@ class SpansClient:
 
     def update_evaluations(
         self,
+        *,
         space_id: str,
         project_name: str,
         dataframe: pd.DataFrame,
@@ -430,6 +432,7 @@ class SpansClient:
 
     def update_annotations(
         self,
+        *,
         space_id: str,
         project_name: str,
         dataframe: pd.DataFrame,
@@ -645,6 +648,7 @@ class SpansClient:
 
     def update_metadata(
         self,
+        *,
         space_id: str,
         project_name: str,
         dataframe: pd.DataFrame,
@@ -977,6 +981,7 @@ class SpansClient:
 
     def export_to_df(
         self,
+        *,
         space_id: str,
         project_name: str,
         start_time: datetime,
@@ -1011,6 +1016,7 @@ class SpansClient:
 
     def export_to_parquet(
         self,
+        *,
         space_id: str,
         project_name: str,
         start_time: datetime,

arize-8.0.0a19/src/arize/version.py

@@ -0,0 +1 @@
+__version__ = "8.0.0a19"

arize-8.0.0a18/src/arize/version.py

@@ -1 +0,0 @@
-__version__ = "8.0.0a18"