oracle-ads 2.13.1rc0__py3-none-any.whl → 2.13.2__py3-none-any.whl

ads/aqua/ui.py

@@ -1,12 +1,9 @@
 #!/usr/bin/env python
-# Copyright (c) 2024 Oracle and/or its affiliates.
+# Copyright (c) 2024, 2025 Oracle and/or its affiliates.
 # Licensed under the Universal Permissive License v 1.0 as shown at https://oss.oracle.com/licenses/upl/
 import concurrent.futures
-from dataclasses import dataclass, field, fields
 from datetime import datetime, timedelta
-from enum import Enum
 from threading import Lock
-from typing import Dict, List, Optional
 
 from cachetools import TTLCache
 from oci.exceptions import ServiceError
@@ -14,210 +11,18 @@ from oci.identity.models import Compartment
 
 from ads.aqua import logger
 from ads.aqua.app import AquaApp
-from ads.aqua.common.entities import ContainerSpec
 from ads.aqua.common.enums import Tags
 from ads.aqua.common.errors import AquaResourceAccessError, AquaValueError
 from ads.aqua.common.utils import get_container_config, sanitize_response
+from ads.aqua.config.container_config import AquaContainerConfig
 from ads.aqua.constants import PRIVATE_ENDPOINT_TYPE
 from ads.common import oci_client as oc
 from ads.common.auth import default_signer
 from ads.common.object_storage_details import ObjectStorageDetails
-from ads.common.serializer import DataClassSerializable
-from ads.config import (
-    COMPARTMENT_OCID,
-    DATA_SCIENCE_SERVICE_NAME,
-    TENANCY_OCID,
-)
+from ads.config import COMPARTMENT_OCID, DATA_SCIENCE_SERVICE_NAME, TENANCY_OCID
 from ads.telemetry import telemetry
 
 
-class ModelFormat(Enum):
-    GGUF = "GGUF"
-    SAFETENSORS = "SAFETENSORS"
-    UNKNOWN = "UNKNOWN"
-
-    def to_dict(self):
-        return self.value
-
-
-# todo: the container config spec information is shared across ui and deployment modules, move them
-#   within ads.aqua.common.entities. In that case, check for circular imports due to usage of get_container_config.
-
-
-@dataclass(repr=False)
-class AquaContainerEvaluationConfig(DataClassSerializable):
-    """
-    Represents the evaluation configuration for the container.
-    """
-
-    inference_max_threads: Optional[int] = None
-    inference_rps: Optional[int] = None
-    inference_timeout: Optional[int] = None
-    inference_retries: Optional[int] = None
-    inference_backoff_factor: Optional[int] = None
-    inference_delay: Optional[int] = None
-
-    @classmethod
-    def from_config(cls, config: dict) -> "AquaContainerEvaluationConfig":
-        return cls(
-            inference_max_threads=config.get("inference_max_threads"),
-            inference_rps=config.get("inference_rps"),
-            inference_timeout=config.get("inference_timeout"),
-            inference_retries=config.get("inference_retries"),
-            inference_backoff_factor=config.get("inference_backoff_factor"),
-            inference_delay=config.get("inference_delay"),
-        )
-
-    def to_filtered_dict(self):
-        return {
-            field.name: getattr(self, field.name)
-            for field in fields(self)
-            if getattr(self, field.name) is not None
-        }
-
-
-@dataclass(repr=False)
-class AquaContainerConfigSpec(DataClassSerializable):
-    cli_param: str = None
-    server_port: str = None
-    health_check_port: str = None
-    env_vars: List[dict] = None
-    restricted_params: List[str] = None
-
-
-@dataclass(repr=False)
-class AquaContainerConfigItem(DataClassSerializable):
-    """Represents an item of the AQUA container configuration."""
-
-    class Platform(Enum):
-        ARM_CPU = "ARM_CPU"
-        NVIDIA_GPU = "NVIDIA_GPU"
-
-        def to_dict(self):
-            return self.value
-
-        def __repr__(self):
-            return repr(self.value)
-
-    name: str = None
-    version: str = None
-    display_name: str = None
-    family: str = None
-    platforms: List[Platform] = None
-    model_formats: List[ModelFormat] = None
-    spec: AquaContainerConfigSpec = field(default_factory=AquaContainerConfigSpec)
-
-
-@dataclass(repr=False)
-class AquaContainerConfig(DataClassSerializable):
-    """
-    Represents a configuration with AQUA containers to be returned to the client.
-    """
-
-    inference: Dict[str, AquaContainerConfigItem] = field(default_factory=dict)
-    finetune: Dict[str, AquaContainerConfigItem] = field(default_factory=dict)
-    evaluate: Dict[str, AquaContainerConfigItem] = field(default_factory=dict)
-
-    def to_dict(self):
-        return {
-            "inference": list(self.inference.values()),
-            "finetune": list(self.finetune.values()),
-            "evaluate": list(self.evaluate.values()),
-        }
-
-    @classmethod
-    def from_container_index_json(
-        cls,
-        config: Optional[Dict] = None,
-        enable_spec: Optional[bool] = False,
-    ) -> "AquaContainerConfig":
-        """
-        Create an AquaContainerConfig instance from a container index JSON.
-
-        Parameters
-        ----------
-        config : Dict
-            The container index JSON.
-        enable_spec: bool
-            flag to check if container specification details should be fetched.
-
-        Returns
-        -------
-        AquaContainerConfig
-            The container configuration instance.
-        """
-        if not config:
-            config = get_container_config()
-        inference_items = {}
-        finetune_items = {}
-        evaluate_items = {}
-
-        # extract inference containers
-        for container_type, containers in config.items():
-            if isinstance(containers, list):
-                for container in containers:
-                    platforms = [
-                        AquaContainerConfigItem.Platform[platform]
-                        for platform in container.get("platforms", [])
-                    ]
-                    model_formats = [
-                        ModelFormat[model_format]
-                        for model_format in container.get("modelFormats", [])
-                    ]
-                    container_spec = (
-                        config.get(ContainerSpec.CONTAINER_SPEC, {}).get(
-                            container_type, {}
-                        )
-                        if enable_spec
-                        else None
-                    )
-                    container_item = AquaContainerConfigItem(
-                        name=container.get("name", ""),
-                        version=container.get("version", ""),
-                        display_name=container.get(
-                            "displayName", container.get("version", "")
-                        ),
-                        family=container_type,
-                        platforms=platforms,
-                        model_formats=model_formats,
-                        spec=(
-                            AquaContainerConfigSpec(
-                                cli_param=container_spec.get(
-                                    ContainerSpec.CLI_PARM, ""
-                                ),
-                                server_port=container_spec.get(
-                                    ContainerSpec.SERVER_PORT, ""
-                                ),
-                                health_check_port=container_spec.get(
-                                    ContainerSpec.HEALTH_CHECK_PORT, ""
-                                ),
-                                env_vars=container_spec.get(ContainerSpec.ENV_VARS, []),
-                                restricted_params=container_spec.get(
-                                    ContainerSpec.RESTRICTED_PARAMS, []
-                                ),
-                            )
-                            if container_spec
-                            else None
-                        ),
-                    )
-                    if container.get("type") == "inference":
-                        inference_items[container_type] = container_item
-                    elif (
-                        container.get("type") == "fine-tune"
-                        or container_type == "odsc-llm-fine-tuning"
-                    ):
-                        finetune_items[container_type] = container_item
-                    elif (
-                        container.get("type") == "evaluate"
-                        or container_type == "odsc-llm-evaluate"
-                    ):
-                        evaluate_items[container_type] = container_item
-
-        return AquaContainerConfig(
-            inference=inference_items, finetune=finetune_items, evaluate=evaluate_items
-        )
-
-
 class AquaUIApp(AquaApp):
     """Contains APIs for supporting Aqua UI.
 
@@ -512,7 +317,8 @@ class AquaUIApp(AquaApp):
 
         Returns
         -------
-            str has json representation of `oci.data_science.models.ModelDeploymentShapeSummary`."""
+        str has json representation of `oci.data_science.models.ModelDeploymentShapeSummary`.
+        """
         compartment_id = kwargs.pop("compartment_id", COMPARTMENT_OCID)
         logger.info(
             f"Loading model deployment shape summary from compartment: {compartment_id}"

ads/common/auth.py

@@ -88,13 +88,15 @@ def set_auth(
     auth: Optional[str] = AuthType.API_KEY,
     oci_config_location: Optional[str] = DEFAULT_LOCATION,
     profile: Optional[str] = DEFAULT_PROFILE,
-    config: Optional[Dict] = {"region": os.environ["OCI_RESOURCE_REGION"]}
-    if os.environ.get("OCI_RESOURCE_REGION")
-    else {},
+    config: Optional[Dict] = (
+        {"region": os.environ["OCI_RESOURCE_REGION"]}
+        if os.environ.get("OCI_RESOURCE_REGION")
+        else {}
+    ),
     signer: Optional[Any] = None,
     signer_callable: Optional[Callable] = None,
-    signer_kwargs: Optional[Dict] = {},
-    client_kwargs: Optional[Dict] = {},
+    signer_kwargs: Optional[Dict] = None,
+    client_kwargs: Optional[Dict] = None,
 ) -> None:
     """
     Sets the default authentication type.
@@ -195,6 +197,9 @@ def set_auth(
     >>> # instance principals authentication dictionary created based on callable with kwargs parameters:
     >>> ads.set_auth(signer_callable=signer_callable, signer_kwargs=signer_kwargs)
     """
+    signer_kwargs = signer_kwargs or {}
+    client_kwargs = client_kwargs or {}
+
     auth_state = AuthState()
 
     valid_auth_keys = AuthFactory.classes.keys()
@@ -258,9 +263,11 @@ def api_keys(
     """
     signer_args = dict(
         oci_config=oci_config if isinstance(oci_config, Dict) else {},
-        oci_config_location=oci_config
-        if isinstance(oci_config, str)
-        else os.path.expanduser(DEFAULT_LOCATION),
+        oci_config_location=(
+            oci_config
+            if isinstance(oci_config, str)
+            else os.path.expanduser(DEFAULT_LOCATION)
+        ),
         oci_key_profile=profile,
         client_kwargs=client_kwargs,
     )
@@ -334,9 +341,11 @@ def security_token(
     """
     signer_args = dict(
         oci_config=oci_config if isinstance(oci_config, Dict) else {},
-        oci_config_location=oci_config
-        if isinstance(oci_config, str)
-        else os.path.expanduser(DEFAULT_LOCATION),
+        oci_config_location=(
+            oci_config
+            if isinstance(oci_config, str)
+            else os.path.expanduser(DEFAULT_LOCATION)
+        ),
         oci_key_profile=profile,
         client_kwargs=client_kwargs,
     )

ads/common/utils.py

@@ -1,10 +1,8 @@
 #!/usr/bin/env python
-# -*- coding: utf-8; -*-
 
 # Copyright (c) 2020, 2024 Oracle and/or its affiliates.
 # Licensed under the Universal Permissive License v 1.0 as shown at https://oss.oracle.com/licenses/upl/
 
-from __future__ import absolute_import, print_function
 
 import collections
 import contextlib
@@ -23,9 +21,8 @@ import tempfile
 from datetime import datetime
 from enum import Enum
 from io import DEFAULT_BUFFER_SIZE
-from pathlib import Path
 from textwrap import fill
-from typing import Dict, Optional, Union
+from typing import Any, Dict, Optional, Tuple, Union
 from urllib import request
 from urllib.parse import urlparse
 
@@ -66,6 +63,8 @@ MIN_RATIO_FOR_DOWN_SAMPLING = 1 / 20
 # Maximum distinct values by cardinality will be used for plotting
 MAX_DISPLAY_VALUES = 10
 
+UNKNOWN = ""
+
 # par link of the index json file.
 PAR_LINK = "https://objectstorage.us-ashburn-1.oraclecloud.com/p/WyjtfVIG0uda-P3-2FmAfwaLlXYQZbvPZmfX1qg0-sbkwEQO6jpwabGr2hMDBmBp/n/ociodscdev/b/service-conda-packs/o/service_pack/index.json"
 
@@ -85,6 +84,7 @@ mpl.rcParams["axes.prop_cycle"] = cycler(
     color=["teal", "blueviolet", "forestgreen", "peru", "y", "dodgerblue", "r"]
 )
 
+
 # sqlalchemy engines
 _engines = {}
 
@@ -152,6 +152,22 @@ def oci_key_location():
     )
 
 
+def text_sanitizer(content):
+    if isinstance(content, str):
+        return (
+            content.replace("“", '"')
+            .replace("”", '"')
+            .replace("’", "'")
+            .replace("‘", "'")
+            .replace("—", "-")
+            .encode("utf-8", "ignore")
+            .decode("utf-8", "ignore")
+        )
+    if isinstance(content, dict):
+        return json.dumps(content)
+    return str(content)
+
+
 @deprecated(
     "2.5.10",
     details="Deprecated, use: from ads.common.auth import AuthState; AuthState().oci_config_path",
@@ -215,6 +231,37 @@ def random_valid_ocid(prefix="ocid1.dataflowapplication.oc1.iad"):
     return f"{left}.{fake}"
 
 
+def parse_bool(value: Any) -> bool:
+    """
+    Converts a value to boolean. For strings, it interprets 'true', '1', or 'yes'
+    (case insensitive) as True; everything else as False.
+
+    Parameters
+    ----------
+    value : Any
+        The value to convert to boolean.
+
+    Returns
+    -------
+    bool
+        The boolean interpretation of the value.
+    """
+    if isinstance(value, bool):
+        return value
+    if isinstance(value, str):
+        return value.strip().lower() in ("true", "1", "yes")
+    return bool(value)
+
+
+def read_file(file_path: str, **kwargs) -> str:
+    try:
+        with fsspec.open(file_path, "r", **kwargs.get("auth", {})) as f:
+            return f.read()
+    except Exception as e:
+        logger.debug(f"Failed to read file {file_path}. {e}")
+        return UNKNOWN
+
+
 def get_dataframe_styles(max_width=75):
     """Styles used for dataframe, example usage:
 
@@ -501,13 +548,13 @@ def print_user_message(
     if is_documentation_mode() and is_notebook():
         if display_type.lower() == "tip":
             if "\n" in msg:
-                t = "<b>{}:</b>".format(title.upper().strip()) if title else ""
+                t = f"<b>{title.upper().strip()}:</b>" if title else ""
 
                 user_message = "{}{}".format(
                     t,
                     "".join(
                         [
-                            "<br>&nbsp;&nbsp;+&nbsp;{}".format(x.strip())
+                            f"<br>&nbsp;&nbsp;+&nbsp;{x.strip()}"
                             for x in msg.strip().split("\n")
                         ]
                     ),
@@ -646,7 +693,7 @@ def ellipsis_strings(raw, n=24):
         else:
             n2 = int(n) // 2 - 3
             n1 = n - n2 - 3
-            result.append("{0}...{1}".format(s[:n1], s[-n2:]))
+            result.append(f"{s[:n1]}...{s[-n2:]}")
 
     return result
 
@@ -942,9 +989,9 @@ def generate_requirement_file(
     with open(os.path.join(file_path, file_name), "w") as req_file:
         for lib in requirements:
             if requirements[lib]:
-                req_file.write("{}=={}\n".format(lib, requirements[lib]))
+                req_file.write(f"{lib}=={requirements[lib]}\n")
             else:
-                req_file.write("{}\n".format(lib))
+                req_file.write(f"{lib}\n")
 
 
 def _get_feature_type_and_dtype(column):
@@ -966,7 +1013,7 @@ def to_dataframe(
         pd.Series,
         np.ndarray,
         pd.DataFrame,
-    ]
+    ],
 ):
     """
     Convert to pandas DataFrame.
@@ -1391,7 +1438,7 @@ def remove_file(file_path: str, auth: Optional[Dict] = None) -> None:
     fs = fsspec.filesystem(scheme, **auth)
     try:
         fs.rm(file_path)
-    except FileNotFoundError as e:
+    except FileNotFoundError:
         raise FileNotFoundError(f"`{file_path}` not found.")
     except Exception as e:
         raise e
@@ -1786,3 +1833,36 @@ def get_log_links(
         console_link_url = f"https://cloud.oracle.com/logging/log-groups/{log_group_id}?region={region}"
 
     return console_link_url
+
+
+def parse_content_disposition(header: str) -> Tuple[str, Dict[str, str]]:
+    """
+    Parses a Content-Disposition header into its main disposition and a dictionary of parameters.
+
+    For example:
+        'attachment; filename="example.txt"'
+    will be parsed into:
+        ('attachment', {'filename': 'example.txt'})
+
+    Parameters
+    ----------
+    header (str): The Content-Disposition header string.
+
+    Returns
+    -------
+    Tuple[str, Dict[str, str]]: A tuple containing the disposition and a dictionary of parameters.
+    """
+    if not header:
+        return "", {}
+
+    parts = header.split(";")
+    # The first part is the main disposition (e.g., "attachment").
+    disposition = parts[0].strip().lower()
+    params: Dict[str, str] = {}
+
+    # Process each subsequent part to extract key-value pairs.
+    for part in parts[1:]:
+        if "=" in part:
+            key, value = part.split("=", 1)
+            params[key.strip().lower()] = value.strip().strip('"')
+    return disposition, params

ads/config.py

@@ -80,6 +80,9 @@ AQUA_TELEMETRY_BUCKET_NS = os.environ.get("AQUA_TELEMETRY_BUCKET_NS", CONDA_BUCK
 DEBUG_TELEMETRY = os.environ.get("DEBUG_TELEMETRY", None)
 AQUA_SERVICE_NAME = "aqua"
 DATA_SCIENCE_SERVICE_NAME = "data-science"
+USER = "USER"
+SERVICE = "SERVICE"
+
 
 
 THREADED_DEFAULT_TIMEOUT = 5

ads/llm/__init__.py

@@ -16,6 +16,7 @@ try:
         OCIDataScienceEmbedding,
     )
     from ads.llm.langchain.plugins.llms.oci_data_science_model_deployment_endpoint import (
+        OCIModelDeploymentLLM,
         OCIModelDeploymentTGI,
         OCIModelDeploymentVLLM,
     )

ads/llm/langchain/plugins/llms/oci_data_science_model_deployment_endpoint.py

@@ -1,6 +1,6 @@
 #!/usr/bin/env python
 
-# Copyright (c) 2024 Oracle and/or its affiliates.
+# Copyright (c) 2024, 2025 Oracle and/or its affiliates.
 # Licensed under the Universal Permissive License v 1.0 as shown at https://oss.oracle.com/licenses/upl/
 
 
@@ -433,23 +433,6 @@ class OCIModelDeploymentLLM(BaseLLM, BaseOCIModelDeployment):
     model: str = DEFAULT_MODEL_NAME
     """The name of the model."""
 
-    max_tokens: int = 256
-    """Denotes the number of tokens to predict per generation."""
-
-    temperature: float = 0.2
-    """A non-negative float that tunes the degree of randomness in generation."""
-
-    k: int = -1
-    """Number of most likely tokens to consider at each step."""
-
-    p: float = 0.75
-    """Total probability mass of tokens to consider at each step."""
-
-    best_of: int = 1
-    """Generates best_of completions server-side and returns the "best"
-    (the one with the highest log probability per token).
-    """
-
     stop: Optional[List[str]] = None
     """Stop words to use when generating. Model output is cut off
     at the first occurrence of any of these substrings."""
@@ -466,14 +449,9 @@ class OCIModelDeploymentLLM(BaseLLM, BaseOCIModelDeployment):
     def _default_params(self) -> Dict[str, Any]:
         """Get the default parameters."""
         return {
-            "best_of": self.best_of,
-            "max_tokens": self.max_tokens,
             "model": self.model,
             "stop": self.stop,
             "stream": self.streaming,
-            "temperature": self.temperature,
-            "top_k": self.k,
-            "top_p": self.p,
         }
 
     @property
@@ -788,6 +766,23 @@ class OCIModelDeploymentTGI(OCIModelDeploymentLLM):
 
     """
 
+    max_tokens: int = 256
+    """Denotes the number of tokens to predict per generation."""
+
+    temperature: float = 0.2
+    """A non-negative float that tunes the degree of randomness in generation."""
+
+    k: int = -1
+    """Number of most likely tokens to consider at each step."""
+
+    p: float = 0.75
+    """Total probability mass of tokens to consider at each step."""
+
+    best_of: int = 1
+    """Generates best_of completions server-side and returns the "best"
+    (the one with the highest log probability per token).
+    """
+
     api: Literal["/generate", "/v1/completions"] = "/v1/completions"
     """Api spec."""
 
@@ -922,6 +917,20 @@ class OCIModelDeploymentVLLM(OCIModelDeploymentLLM):
 
     """
 
+    max_tokens: int = 256
+    """Denotes the number of tokens to predict per generation."""
+
+    temperature: float = 0.2
+    """A non-negative float that tunes the degree of randomness in generation."""
+
+    p: float = 0.75
+    """Total probability mass of tokens to consider at each step."""
+
+    best_of: int = 1
+    """Generates best_of completions server-side and returns the "best"
+    (the one with the highest log probability per token).
+    """
+
     n: int = 1
     """Number of output sequences to return for the given prompt."""
 

ads/model/artifact_downloader.py

@@ -170,7 +170,10 @@ class LargeArtifactDownloader(ArtifactDownloader):
         """Downloads model artifacts."""
         self.progress.update("Importing model artifacts from catalog")
 
-        if self.dsc_model._is_model_by_reference() and self.model_file_description:
+        if (
+            self.dsc_model.is_model_created_by_reference()
+            and self.model_file_description
+        ):
             self.download_from_model_file_description()
             self.progress.update()
             return

ads/model/common/utils.py

@@ -1,5 +1,4 @@
 #!/usr/bin/env python
-# -*- coding: utf-8 -*--
 
 # Copyright (c) 2022, 2023 Oracle and/or its affiliates.
 # Licensed under the Universal Permissive License v 1.0 as shown at https://oss.oracle.com/licenses/upl/
@@ -7,16 +6,29 @@
 import json
 import os
 import tempfile
-import yaml
 from typing import Any, Dict, Optional
 from zipfile import ZipFile
-from ads.common import utils
 
+import yaml
+
+from ads.common import utils
+from ads.common.extended_enum import ExtendedEnum
 
 DEPRECATE_AS_ONNX_WARNING = "This attribute `as_onnx` will be deprecated in the future. You can choose specific format by setting `model_save_serializer`."
 DEPRECATE_USE_TORCH_SCRIPT_WARNING = "This attribute `use_torch_script` will be deprecated in the future. You can choose specific format by setting `model_save_serializer`."
 
 
+class MetadataArtifactPathType(ExtendedEnum):
+    """
+    Enum for defining metadata artifact path type.
+    Can be either local path or OSS path. It can also be the content itself.
+    """
+
+    LOCAL = "local"
+    OSS = "oss"
+    CONTENT = "content"
+
+
 def _extract_locals(
     locals: Dict[str, Any], filter_out_nulls: Optional[bool] = True
 ) -> Dict[str, Any]: