oracle-ads 2.12.5__tar.gz → 2.12.7__tar.gz

{oracle_ads-2.12.5 → oracle_ads-2.12.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: oracle_ads
-Version: 2.12.5
+Version: 2.12.7
 Summary: Oracle Accelerated Data Science SDK
 Keywords: Oracle Cloud Infrastructure,OCI,Machine Learning,ML,Artificial Intelligence,AI,Data Science,Cloud,Oracle
 Author: Oracle Data Science
@@ -37,7 +37,7 @@ Requires-Dist: pydantic>=2.6.3
 Requires-Dist: oracle_ads[opctl] ; extra == "anomaly"
 Requires-Dist: autots ; extra == "anomaly"
 Requires-Dist: oracledb ; extra == "anomaly"
-Requires-Dist: report-creator==1.0.9 ; extra == "anomaly"
+Requires-Dist: report-creator==1.0.28 ; extra == "anomaly"
 Requires-Dist: rrcf==0.4.4 ; extra == "anomaly"
 Requires-Dist: scikit-learn ; extra == "anomaly"
 Requires-Dist: salesforce-merlion[all]==2.0.4 ; extra == "anomaly"
@@ -77,7 +77,7 @@ Requires-Dist: sktime ; extra == "forecast"
 Requires-Dist: statsmodels ; extra == "forecast"
 Requires-Dist: plotly ; extra == "forecast"
 Requires-Dist: oracledb ; extra == "forecast"
-Requires-Dist: report-creator==1.0.9 ; extra == "forecast"
+Requires-Dist: report-creator==1.0.28 ; extra == "forecast"
 Requires-Dist: geopandas<1.0.0 ; extra == "geo"
 Requires-Dist: fiona<=1.9.6 ; extra == "geo"
 Requires-Dist: oracle_ads[viz] ; extra == "geo"
@@ -109,7 +109,7 @@ Requires-Dist: py-cpuinfo ; extra == "opctl"
 Requires-Dist: rich ; extra == "opctl"
 Requires-Dist: fire ; extra == "opctl"
 Requires-Dist: cachetools ; extra == "opctl"
-Requires-Dist: huggingface_hub==0.23.4 ; extra == "opctl"
+Requires-Dist: huggingface_hub==0.26.2 ; extra == "opctl"
 Requires-Dist: optuna==2.9.0 ; extra == "optuna"
 Requires-Dist: oracle_ads[viz] ; extra == "optuna"
 Requires-Dist: aiohttp ; extra == "pii"
@@ -121,15 +121,16 @@ Requires-Dist: scrubadub==2.0.1 ; extra == "pii"
 Requires-Dist: scrubadub_spacy ; extra == "pii"
 Requires-Dist: spacy-transformers==1.2.5 ; extra == "pii"
 Requires-Dist: spacy==3.6.1 ; extra == "pii"
-Requires-Dist: report-creator==1.0.9 ; extra == "pii"
+Requires-Dist: report-creator==1.0.28 ; extra == "pii"
 Requires-Dist: oracle_ads[opctl] ; extra == "recommender"
 Requires-Dist: scikit-surprise ; extra == "recommender"
 Requires-Dist: plotly ; extra == "recommender"
-Requires-Dist: report-creator==1.0.9 ; extra == "recommender"
+Requires-Dist: report-creator==1.0.28 ; extra == "recommender"
 Requires-Dist: pyspark>=3.0.0 ; extra == "spark"
 Requires-Dist: oracle_ads[viz] ; extra == "tensorflow"
 Requires-Dist: tensorflow<=2.15.1 ; extra == "tensorflow"
 Requires-Dist: arff ; extra == "testsuite"
+Requires-Dist: autogen-agentchat~=0.2 ; extra == "testsuite"
 Requires-Dist: category_encoders==2.6.3 ; extra == "testsuite"
 Requires-Dist: cohere==4.53 ; extra == "testsuite"
 Requires-Dist: faiss-cpu ; extra == "testsuite"

{oracle_ads-2.12.5 → oracle_ads-2.12.7}/THIRD_PARTY_LICENSES.txt

@@ -24,6 +24,12 @@ autots
 * Source code: https://github.com/winedarksea/AutoTS
 * Project home: https://winedarksea.github.io/AutoTS/build/html/index.html
 
+autogen
+* Copyright (c) 2024 Microsoft Corporation.
+* License: MIT License
+* Source code: https://github.com/microsoft/autogen
+* Project home: microsoft.github.io/autogen/
+
 bokeh
 * Copyright Copyright (c) 2012 - 2021, Anaconda, Inc., and Bokeh Contributors
 * License: BSD 3-Clause "New" or "Revised" License

{oracle_ads-2.12.5 → oracle_ads-2.12.7}/ads/aqua/common/decorator.py

@@ -69,6 +69,16 @@ def handle_exceptions(func):
                 reason=error.message,
                 service_payload=error.args[0] if error.args else None,
                 exc_info=sys.exc_info(),
+                aqua_api_details=dict(
+                    # __qualname__ gives information of class and name of api
+                    aqua_api_name=func.__qualname__,
+                    oci_api_name=getattr(
+                        error, "operation_name", "Unknown OCI Operation"
+                    ),
+                    service_endpoint=getattr(
+                        error, "request_endpoint", "Unknown Request Endpoint"
+                    )
+                )
             )
         except (
             ClientError,

{oracle_ads-2.12.5 → oracle_ads-2.12.7}/ads/aqua/common/utils.py

@@ -788,13 +788,14 @@ def get_ocid_substring(ocid: str, key_len: int) -> str:
     return ocid[-key_len:] if ocid and len(ocid) > key_len else ""
 
 
-def upload_folder(os_path: str, local_dir: str, model_name: str) -> str:
+def upload_folder(os_path: str, local_dir: str, model_name: str, exclude_pattern: str = None) -> str:
     """Upload the local folder to the object storage
 
     Args:
         os_path (str): object storage URI with prefix. This is the path to upload
         local_dir (str): Local directory where the object is downloaded
         model_name (str): Name of the huggingface model
+        exclude_pattern (optional, str): The matching pattern of files to be excluded from uploading.
     Retuns:
         str: Object name inside the bucket
     """
@@ -804,6 +805,8 @@ def upload_folder(os_path: str, local_dir: str, model_name: str) -> str:
     auth_state = AuthState()
     object_path = os_details.filepath.rstrip("/") + "/" + model_name + "/"
     command = f"oci os object bulk-upload --src-dir {local_dir} --prefix {object_path} -bn {os_details.bucket} -ns {os_details.namespace} --auth {auth_state.oci_iam_type} --profile {auth_state.oci_key_profile} --no-overwrite"
+    if exclude_pattern:
+        command += f" --exclude {exclude_pattern}"
     try:
         logger.info(f"Running: {command}")
         subprocess.check_call(shlex.split(command))

{oracle_ads-2.12.5 → oracle_ads-2.12.7}/ads/aqua/constants.py

@@ -35,6 +35,7 @@ AQUA_MODEL_ARTIFACT_CONFIG = "config.json"
 AQUA_MODEL_ARTIFACT_CONFIG_MODEL_NAME = "_name_or_path"
 AQUA_MODEL_ARTIFACT_CONFIG_MODEL_TYPE = "model_type"
 AQUA_MODEL_ARTIFACT_FILE = "model_file"
+HF_METADATA_FOLDER = ".cache/"
 HF_LOGIN_DEFAULT_TIMEOUT = 2
 
 TRAINING_METRICS_FINAL = "training_metrics_final"

{oracle_ads-2.12.5 → oracle_ads-2.12.7}/ads/aqua/evaluation/entities.py

@@ -9,11 +9,12 @@ aqua.evaluation.entities
 This module contains dataclasses for aqua evaluation.
 """
 
+from typing import Any, Dict, List, Optional
+
 from pydantic import Field
-from typing import Any, Dict, List, Optional, Union
 
-from ads.aqua.data import AquaResourceIdentifier
 from ads.aqua.config.utils.serializer import Serializable
+from ads.aqua.data import AquaResourceIdentifier
 
 
 class CreateAquaEvaluationDetails(Serializable):
@@ -82,11 +83,13 @@ class CreateAquaEvaluationDetails(Serializable):
     ocpus: Optional[float] = None
     log_group_id: Optional[str] = None
     log_id: Optional[str] = None
-    metrics: Optional[List[str]] = None
+    metrics: Optional[List[Dict[str, Any]]] = None
     force_overwrite: Optional[bool] = False
 
     class Config:
         extra = "ignore"
+        protected_namespaces = ()
+
 
 class AquaEvalReport(Serializable):
     evaluation_id: str = ""
@@ -95,6 +98,7 @@ class AquaEvalReport(Serializable):
     class Config:
         extra = "ignore"
 
+
 class AquaEvalParams(Serializable):
     shape: str = ""
     dataset_path: str = ""
@@ -103,6 +107,7 @@ class AquaEvalParams(Serializable):
     class Config:
         extra = "allow"
 
+
 class AquaEvalMetric(Serializable):
     key: str
     name: str
@@ -111,6 +116,7 @@ class AquaEvalMetric(Serializable):
     class Config:
         extra = "ignore"
 
+
 class AquaEvalMetricSummary(Serializable):
     metric: str = ""
     score: str = ""
@@ -119,6 +125,7 @@ class AquaEvalMetricSummary(Serializable):
     class Config:
         extra = "ignore"
 
+
 class AquaEvalMetrics(Serializable):
     id: str
     report: str
@@ -128,17 +135,19 @@ class AquaEvalMetrics(Serializable):
     class Config:
         extra = "ignore"
 
+
 class AquaEvaluationCommands(Serializable):
     evaluation_id: str
     evaluation_target_id: str
     input_data: Dict[str, Any]
-    metrics: List[str]
+    metrics: List[Dict[str, Any]]
     output_dir: str
     params: Dict[str, Any]
 
     class Config:
         extra = "ignore"
 
+
 class AquaEvaluationSummary(Serializable):
     """Represents a summary of Aqua evalution."""
 
@@ -157,6 +166,7 @@ class AquaEvaluationSummary(Serializable):
     class Config:
         extra = "ignore"
 
+
 class AquaEvaluationDetail(AquaEvaluationSummary):
     """Represents a details of Aqua evalution."""
 

{oracle_ads-2.12.5 → oracle_ads-2.12.7}/ads/aqua/evaluation/evaluation.py

@@ -159,7 +159,8 @@ class AquaEvaluationApp(AquaApp):
                 create_aqua_evaluation_details = CreateAquaEvaluationDetails(**kwargs)
             except Exception as ex:
                 custom_errors = {
-                    ".".join(map(str, e["loc"])): e["msg"] for e in json.loads(ex.json())
+                    ".".join(map(str, e["loc"])): e["msg"]
+                    for e in json.loads(ex.json())
                 }
                 raise AquaValueError(
                     f"Invalid create evaluation parameters. Error details: {custom_errors}."
@@ -619,11 +620,6 @@ class AquaEvaluationApp(AquaApp):
             evaluation_id=evaluation_id,
             evaluation_target_id=evaluation_source_id,
             input_data={
-                "columns": {
-                    "prompt": "prompt",
-                    "completion": "completion",
-                    "category": "category",
-                },
                 "format": Path(dataset_path).suffix,
                 "url": dataset_path,
             },

{oracle_ads-2.12.5 → oracle_ads-2.12.7}/ads/aqua/extension/aqua_ws_msg_handler.py

@@ -78,10 +78,12 @@ class AquaWSMsgHandler:
                 logger.warning(reply["message"])
         # telemetry may not be present if there is an error while initializing
         if hasattr(self, "telemetry"):
+            aqua_api_details = kwargs.get("aqua_api_details", {})
             self.telemetry.record_event_async(
                 category="aqua/error",
                 action=str(status_code),
                 value=reason,
+                **aqua_api_details
             )
         response = AquaWsError(
             status=status_code,

{oracle_ads-2.12.5 → oracle_ads-2.12.7}/ads/aqua/extension/base_handler.py

@@ -98,10 +98,12 @@ class AquaAPIhandler(APIHandler):
 
         # telemetry may not be present if there is an error while initializing
         if hasattr(self, "telemetry"):
+            aqua_api_details = kwargs.get("aqua_api_details", {})
             self.telemetry.record_event_async(
                 category="aqua/error",
                 action=str(status_code),
                 value=reason,
+                **aqua_api_details
             )
 
         self.finish(json.dumps(reply))

{oracle_ads-2.12.5 → oracle_ads-2.12.7}/ads/aqua/extension/model_handler.py

@@ -129,6 +129,8 @@ class AquaModelHandler(AquaAPIhandler):
             str(input_data.get("download_from_hf", "false")).lower() == "true"
         )
         inference_container_uri = input_data.get("inference_container_uri")
+        allow_patterns = input_data.get("allow_patterns")
+        ignore_patterns = input_data.get("ignore_patterns")
 
         return self.finish(
             AquaModelApp().register(
@@ -141,6 +143,8 @@ class AquaModelHandler(AquaAPIhandler):
                 project_id=project_id,
                 model_file=model_file,
                 inference_container_uri=inference_container_uri,
+                allow_patterns=allow_patterns,
+                ignore_patterns=ignore_patterns,
             )
         )
 

{oracle_ads-2.12.5 → oracle_ads-2.12.7}/ads/aqua/finetuning/constants.py

@@ -15,3 +15,6 @@ class FineTuneCustomMetadata(str, metaclass=ExtendedEnumMeta):
     SERVICE_MODEL_ARTIFACT_LOCATION = "artifact_location"
     SERVICE_MODEL_DEPLOYMENT_CONTAINER = "deployment-container"
     SERVICE_MODEL_FINE_TUNE_CONTAINER = "finetune-container"
+
+
+ENV_AQUA_FINE_TUNING_CONTAINER = "AQUA_FINE_TUNING_CONTAINER"

{oracle_ads-2.12.5 → oracle_ads-2.12.7}/ads/aqua/finetuning/finetuning.py

@@ -31,7 +31,10 @@ from ads.aqua.constants import (
     UNKNOWN_DICT,
 )
 from ads.aqua.data import AquaResourceIdentifier
-from ads.aqua.finetuning.constants import *
+from ads.aqua.finetuning.constants import (
+    ENV_AQUA_FINE_TUNING_CONTAINER,
+    FineTuneCustomMetadata,
+)
 from ads.aqua.finetuning.entities import *
 from ads.common.auth import default_signer
 from ads.common.object_storage_details import ObjectStorageDetails
@@ -310,6 +313,15 @@ class AquaFineTuningApp(AquaApp):
         except Exception:
             pass
 
+        if not is_custom_container and ENV_AQUA_FINE_TUNING_CONTAINER in os.environ:
+            ft_container = os.environ[ENV_AQUA_FINE_TUNING_CONTAINER]
+            logger.info(
+                "Using container set by environment variable %s=%s",
+                ENV_AQUA_FINE_TUNING_CONTAINER,
+                ft_container,
+            )
+            is_custom_container = True
+
         ft_parameters.batch_size = ft_parameters.batch_size or (
             ft_config.get("shape", UNKNOWN_DICT)
             .get(create_fine_tuning_details.shape_name, UNKNOWN_DICT)
@@ -559,7 +571,6 @@ class AquaFineTuningApp(AquaApp):
         Dict:
             A dict of allowed finetuning configs.
         """
-
         config = self.get_config(model_id, AQUA_MODEL_FINETUNING_CONFIG)
         if not config:
             logger.debug(

{oracle_ads-2.12.5 → oracle_ads-2.12.7}/ads/aqua/model/entities.py

@@ -289,6 +289,8 @@ class ImportModelDetails(CLIBuilderMixin):
     project_id: Optional[str] = None
     model_file: Optional[str] = None
     inference_container_uri: Optional[str] = None
+    allow_patterns: Optional[List[str]] = None
+    ignore_patterns: Optional[List[str]] = None
 
     def __post_init__(self):
         self._command = "model register"

{oracle_ads-2.12.5 → oracle_ads-2.12.7}/ads/aqua/model/model.py

@@ -40,6 +40,7 @@ from ads.aqua.constants import (
     AQUA_MODEL_ARTIFACT_CONFIG_MODEL_TYPE,
     AQUA_MODEL_ARTIFACT_FILE,
     AQUA_MODEL_TYPE_CUSTOM,
+    HF_METADATA_FOLDER,
     LICENSE_TXT,
     MODEL_BY_REFERENCE_OSS_PATH_KEY,
     README,
@@ -1274,6 +1275,8 @@ class AquaModelApp(AquaApp):
         model_name: str,
         os_path: str,
         local_dir: str = None,
+        allow_patterns: List[str] = None,
+        ignore_patterns: List[str] = None,
     ) -> str:
         """This helper function downloads the model artifact from Hugging Face to a local folder, then uploads
         to object storage location.
@@ -1283,6 +1286,12 @@ class AquaModelApp(AquaApp):
         model_name (str): The huggingface model name.
         os_path (str): The OS path where the model files are located.
         local_dir (str): The local temp dir to store the huggingface model.
+        allow_patterns (list): Model files matching at least one pattern are downloaded.
+            Example: ["*.json"] will download all .json files. ["folder/*"] will download all files under `folder`.
+            Patterns are Standard Wildcards (globbing patterns) and rules can be found here: https://docs.python.org/3/library/fnmatch.html
+        ignore_patterns (list): Model files matching any of the patterns are not downloaded.
+            Example: ["*.json"] will ignore all .json files. ["folder/*"] will ignore all files under `folder`.
+            Patterns are Standard Wildcards (globbing patterns) and rules can be found here: https://docs.python.org/3/library/fnmatch.html
 
         Returns
         -------
@@ -1293,30 +1302,19 @@ class AquaModelApp(AquaApp):
         if not local_dir:
             local_dir = os.path.join(os.path.expanduser("~"), "cached-model")
         local_dir = os.path.join(local_dir, model_name)
-        retry = 10
-        i = 0
-        huggingface_download_err_message = None
-        while i < retry:
-            try:
-                # Download to cache folder. The while loop retries when there is a network failure
-                snapshot_download(repo_id=model_name)
-            except Exception as e:
-                huggingface_download_err_message = str(e)
-                i += 1
-            else:
-                break
-        if i == retry:
-            raise Exception(
-                f"Could not download the model {model_name} from https://huggingface.co with message {huggingface_download_err_message}"
-            )
         os.makedirs(local_dir, exist_ok=True)
-        # Copy the model from the cache to destination
-        snapshot_download(repo_id=model_name, local_dir=local_dir)
-        # Upload to object storage
+        snapshot_download(
+            repo_id=model_name,
+            local_dir=local_dir,
+            allow_patterns=allow_patterns,
+            ignore_patterns=ignore_patterns,
+        )
+        # Upload to object storage and skip .cache/huggingface/ folder
         model_artifact_path = upload_folder(
             os_path=os_path,
             local_dir=local_dir,
             model_name=model_name,
+            exclude_pattern=f"{HF_METADATA_FOLDER}*"
         )
 
         return model_artifact_path
@@ -1335,6 +1333,12 @@ class AquaModelApp(AquaApp):
                 os_path (str): Object storage destination URI to store the downloaded model. Format: oci://bucket-name@namespace/prefix
                 inference_container (str): selects service defaults
                 finetuning_container (str): selects service defaults
+                allow_patterns (list): Model files matching at least one pattern are downloaded.
+                    Example: ["*.json"] will download all .json files. ["folder/*"] will download all files under `folder`.
+                    Patterns are Standard Wildcards (globbing patterns) and rules can be found here: https://docs.python.org/3/library/fnmatch.html
+                ignore_patterns (list): Model files matching any of the patterns are not downloaded.
+                    Example: ["*.json"] will ignore all .json files. ["folder/*"] will ignore all files under `folder`.
+                    Patterns are Standard Wildcards (globbing patterns) and rules can be found here: https://docs.python.org/3/library/fnmatch.html
 
         Returns:
             AquaModel:
@@ -1381,6 +1385,8 @@ class AquaModelApp(AquaApp):
                 model_name=model_name,
                 os_path=import_model_details.os_path,
                 local_dir=import_model_details.local_dir,
+                allow_patterns=import_model_details.allow_patterns,
+                ignore_patterns=import_model_details.ignore_patterns,
             ).rstrip("/")
         else:
             artifact_path = import_model_details.os_path.rstrip("/")