snowflake-ml-python 1.7.5__py3-none-any.whl → 1.8.0__py3-none-any.whl

snowflake/ml/registry/_manager/model_manager.py

@@ -4,7 +4,7 @@ from typing import Any, Dict, List, Optional, Tuple, Union
 import pandas as pd
 from absl.logging import logging
 
-from snowflake.ml._internal import telemetry
+from snowflake.ml._internal import platform_capabilities, telemetry
 from snowflake.ml._internal.exceptions import error_codes, exceptions
 from snowflake.ml._internal.human_readable_id import hrid_generator
 from snowflake.ml._internal.utils import sql_identifier
@@ -13,7 +13,7 @@ from snowflake.ml.model._client.model import model_impl, model_version_impl
 from snowflake.ml.model._client.ops import metadata_ops, model_ops, service_ops
 from snowflake.ml.model._model_composer import model_composer
 from snowflake.ml.model._packager.model_meta import model_meta
-from snowflake.snowpark import session
+from snowflake.snowpark import exceptions as snowpark_exceptions, session
 
 logger = logging.getLogger(__name__)
 
@@ -46,6 +46,7 @@ class ModelManager:
         metrics: Optional[Dict[str, Any]] = None,
         conda_dependencies: Optional[List[str]] = None,
         pip_requirements: Optional[List[str]] = None,
+        artifact_repository_map: Optional[Dict[str, str]] = None,
         target_platforms: Optional[List[model_types.SupportedTargetPlatformType]] = None,
         python_version: Optional[str] = None,
         signatures: Optional[Dict[str, model_signature.ModelSignature]] = None,
@@ -127,6 +128,7 @@ class ModelManager:
             metrics=metrics,
             conda_dependencies=conda_dependencies,
             pip_requirements=pip_requirements,
+            artifact_repository_map=artifact_repository_map,
             target_platforms=target_platforms,
             python_version=python_version,
             signatures=signatures,
@@ -149,6 +151,7 @@ class ModelManager:
         metrics: Optional[Dict[str, Any]] = None,
         conda_dependencies: Optional[List[str]] = None,
         pip_requirements: Optional[List[str]] = None,
+        artifact_repository_map: Optional[Dict[str, str]] = None,
         target_platforms: Optional[List[model_types.SupportedTargetPlatformType]] = None,
         python_version: Optional[str] = None,
         signatures: Optional[Dict[str, model_signature.ModelSignature]] = None,
@@ -163,11 +166,42 @@ class ModelManager:
         database_name_id, schema_name_id, model_name_id = sql_identifier.parse_fully_qualified_name(model_name)
         version_name_id = sql_identifier.SqlIdentifier(version_name)
 
-        stage_path = self._model_ops.prepare_model_stage_path(
-            database_name=database_name_id,
-            schema_name=schema_name_id,
-            statement_params=statement_params,
-        )
+        use_live_commit = platform_capabilities.PlatformCapabilities.get_instance().is_live_commit_enabled()
+        if use_live_commit:
+            logger.info("Using live commit model version")
+        else:
+            logger.info("Using non-live commit model version")
+
+        if use_live_commit:
+            # This step creates the live model version, and the files can be written directly to the stage
+            # after this.
+            try:
+                self._model_ops.add_or_create_live_version(
+                    database_name=database_name_id,
+                    schema_name=schema_name_id,
+                    model_name=model_name_id,
+                    version_name=version_name_id,
+                    statement_params=statement_params,
+                )
+            except (AssertionError, snowpark_exceptions.SnowparkSQLException) as e:
+                logger.info(f"Failed to create live model version: {e}, falling back to regular model version creation")
+                use_live_commit = False
+
+        if use_live_commit:
+            # using model version's stage path to write files directly to the stage
+            stage_path = self._model_ops.get_model_version_stage_path(
+                database_name=database_name_id,
+                schema_name=schema_name_id,
+                model_name=model_name_id,
+                version_name=version_name_id,
+            )
+        else:
+            # using a temp path to write files and then upload to the model version's stage
+            stage_path = self._model_ops.prepare_model_temp_stage_path(
+                database_name=database_name_id,
+                schema_name=schema_name_id,
+                statement_params=statement_params,
+            )
 
         platforms = None
         # User specified target platforms are defaulted to None and will not show up in the generated manifest.
@@ -175,6 +209,18 @@ class ModelManager:
             # Convert any string target platforms to TargetPlatform objects
             platforms = [model_types.TargetPlatform(platform) for platform in target_platforms]
 
+        if artifact_repository_map:
+            for channel, artifact_repository_name in artifact_repository_map.items():
+                db_id, schema_id, repo_id = sql_identifier.parse_fully_qualified_name(artifact_repository_name)
+
+                artifact_repository_map[channel] = sql_identifier.get_fully_qualified_name(
+                    db_id,
+                    schema_id,
+                    repo_id,
+                    self._database_name,
+                    self._schema_name,
+                )
+
         logger.info("Start packaging and uploading your model. It might take some time based on the size of the model.")
 
         mc = model_composer.ModelComposer(
@@ -187,6 +233,7 @@ class ModelManager:
             sample_input_data=sample_input_data,
             conda_dependencies=conda_dependencies,
             pip_requirements=pip_requirements,
+            artifact_repository_map=artifact_repository_map,
             target_platforms=platforms,
             python_version=python_version,
             user_files=user_files,
@@ -211,6 +258,7 @@ class ModelManager:
             model_name=model_name_id,
             version_name=version_name_id,
             statement_params=statement_params,
+            use_live_commit=use_live_commit,
         )
 
         mv = model_version_impl.ModelVersion._ref(

snowflake/ml/registry/registry.py

@@ -108,6 +108,7 @@ class Registry:
         metrics: Optional[Dict[str, Any]] = None,
         conda_dependencies: Optional[List[str]] = None,
         pip_requirements: Optional[List[str]] = None,
+        artifact_repository_map: Optional[Dict[str, str]] = None,
         target_platforms: Optional[List[model_types.SupportedTargetPlatformType]] = None,
         python_version: Optional[str] = None,
         signatures: Optional[Dict[str, model_signature.ModelSignature]] = None,
@@ -140,6 +141,13 @@ class Registry:
                 See https://docs.snowflake.com/en/developer-guide/snowflake-ml/model-registry/container for more.
                 Models with pip requirements specified will not be executable in Snowflake Warehouse where all
                 dependencies must be retrieved from Snowflake Anaconda Channel.
+            artifact_repository_map: Specifies a mapping of package channels or platforms to custom artifact
+                repositories. Defaults to None. Currently, the mapping applies only to warehouse execution.
+                Note : This feature is currently in Private Preview; please contact your Snowflake account team
+                to enable it.
+                Format: {channel_name: artifact_repository_name}, where:
+                   - channel_name: The name of the Conda package channel (e.g., 'condaforge') or 'pip' for pip packages.
+                   - artifact_repository_name: The name or URL of the repository to fetch packages from.
             target_platforms: List of target platforms to run the model. The only acceptable inputs are a combination of
                 {"WAREHOUSE", "SNOWPARK_CONTAINER_SERVICES"}. Defaults to None.
             python_version: Python version in which the model is run. Defaults to None.
@@ -210,6 +218,7 @@ class Registry:
             "metrics",
             "conda_dependencies",
             "pip_requirements",
+            "artifact_repository_map",
             "target_platforms",
             "python_version",
             "signatures",
@@ -225,6 +234,7 @@ class Registry:
         metrics: Optional[Dict[str, Any]] = None,
         conda_dependencies: Optional[List[str]] = None,
         pip_requirements: Optional[List[str]] = None,
+        artifact_repository_map: Optional[Dict[str, str]] = None,
         target_platforms: Optional[List[model_types.SupportedTargetPlatformType]] = None,
         python_version: Optional[str] = None,
         signatures: Optional[Dict[str, model_signature.ModelSignature]] = None,
@@ -259,6 +269,13 @@ class Registry:
                 See https://docs.snowflake.com/en/developer-guide/snowflake-ml/model-registry/container for more.
                 Models with pip requirements specified will not be executable in Snowflake Warehouse where all
                 dependencies must be retrieved from Snowflake Anaconda Channel.
+            artifact_repository_map: Specifies a mapping of package channels or platforms to custom artifact
+                repositories. Defaults to None. Currently, the mapping applies only to warehouse execution.
+                Note : This feature is currently in Private Preview; please contact your Snowflake account team to
+                enable it.
+                Format: {channel_name: artifact_repository_name}, where:
+                   - channel_name: The name of the Conda package channel (e.g., 'condaforge') or 'pip' for pip packages.
+                   - artifact_repository_name: The name or URL of the repository to fetch packages from.
             target_platforms: List of target platforms to run the model. The only acceptable inputs are a combination of
                 {"WAREHOUSE", "SNOWPARK_CONTAINER_SERVICES"}. Defaults to None.
             python_version: Python version in which the model is run. Defaults to None.
@@ -345,6 +362,7 @@ class Registry:
             metrics=metrics,
             conda_dependencies=conda_dependencies,
             pip_requirements=pip_requirements,
+            artifact_repository_map=artifact_repository_map,
             target_platforms=target_platforms,
             python_version=python_version,
             signatures=signatures,

snowflake/ml/version.py

@@ -1 +1 @@
-VERSION="1.7.5"
+VERSION="1.8.0"

{snowflake_ml_python-1.7.5.dist-info → snowflake_ml_python-1.8.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
-Metadata-Version: 2.2
+Metadata-Version: 2.4
 Name: snowflake-ml-python
-Version: 1.7.5
+Version: 1.8.0
 Summary: The machine learning client library that is used for interacting with Snowflake to build machine learning solutions.
 Author-email: "Snowflake, Inc" <support@snowflake.com>
 License: 
@@ -251,7 +251,7 @@ Requires-Dist: retrying<2,>=1.3.3
 Requires-Dist: s3fs<2026,>=2024.6.1
 Requires-Dist: scikit-learn<1.6,>=1.4
 Requires-Dist: scipy<2,>=1.9
-Requires-Dist: snowflake-connector-python[pandas]<4,>=3.5.0
+Requires-Dist: snowflake-connector-python[pandas]<4,>=3.12.0
 Requires-Dist: snowflake-snowpark-python!=1.26.0,<2,>=1.17.0
 Requires-Dist: sqlparse<1,>=0.4
 Requires-Dist: typing-extensions<5,>=4.1.0
@@ -261,14 +261,14 @@ Requires-Dist: catboost<2,>=1.2.0; extra == "all"
 Requires-Dist: keras<4,>=2.0.0; extra == "all"
 Requires-Dist: lightgbm<5,>=4.1.0; extra == "all"
 Requires-Dist: mlflow<3,>=2.16.0; extra == "all"
-Requires-Dist: sentence-transformers<3,>=2.7.0; extra == "all"
+Requires-Dist: sentence-transformers<4,>=2.7.0; extra == "all"
 Requires-Dist: sentencepiece<0.2.0,>=0.1.95; extra == "all"
 Requires-Dist: shap<1,>=0.46.0; extra == "all"
 Requires-Dist: tensorflow<3,>=2.17.0; extra == "all"
 Requires-Dist: tokenizers<1,>=0.15.1; extra == "all"
 Requires-Dist: torch<3,>=2.0.1; extra == "all"
 Requires-Dist: torchdata<1,>=0.4; extra == "all"
-Requires-Dist: transformers<5,>=4.37.2; extra == "all"
+Requires-Dist: transformers<5,>=4.39.3; extra == "all"
 Provides-Extra: catboost
 Requires-Dist: catboost<2,>=1.2.0; extra == "catboost"
 Provides-Extra: keras
@@ -287,11 +287,12 @@ Provides-Extra: torch
 Requires-Dist: torch<3,>=2.0.1; extra == "torch"
 Requires-Dist: torchdata<1,>=0.4; extra == "torch"
 Provides-Extra: transformers
-Requires-Dist: sentence-transformers<3,>=2.7.0; extra == "transformers"
+Requires-Dist: sentence-transformers<4,>=2.7.0; extra == "transformers"
 Requires-Dist: sentencepiece<0.2.0,>=0.1.95; extra == "transformers"
 Requires-Dist: tokenizers<1,>=0.15.1; extra == "transformers"
 Requires-Dist: torch<3,>=2.0.1; extra == "transformers"
-Requires-Dist: transformers<5,>=4.37.2; extra == "transformers"
+Requires-Dist: transformers<5,>=4.39.3; extra == "transformers"
+Dynamic: license-file
 
 # Snowpark ML
 
@@ -401,10 +402,287 @@ NOTE: Version 1.7.0 is used as example here. Please choose the the latest versio
 
 # Release History
 
-## 1.7.5
+## 1.8.0
+
+### Bug Fixes
+
+- Modeling: Fix a bug in some metrics that allowed an unsupported version of numpy to be installed
+  automatically in the stored procedure, resulting in a numpy error on execution
+- Registry: Fix a bug that leads to incorrect `Model is does not have _is_inference_api` error message when assigning
+  a supported model as a property of a CustomModel.
+- Registry: Fix a bug that inference is not working when models with more than 500 input features
+  are deployed to SPCS.
+
+### Behavior Change
+
+- Registry: With FeatureGroupSpec support, auto inferred model signature for `transformers.Pipeline` models have been
+  updated, including:
+  - Signature for fill-mask task has been changed from
+
+    ```python
+    ModelSignature(
+        inputs=[
+            FeatureSpec(name="inputs", dtype=DataType.STRING),
+        ],
+        outputs=[
+            FeatureSpec(name="outputs", dtype=DataType.STRING),
+        ],
+    )
+    ```
+
+    to
+
+    ```python
+    ModelSignature(
+        inputs=[
+            FeatureSpec(name="inputs", dtype=DataType.STRING),
+        ],
+        outputs=[
+            FeatureGroupSpec(
+                name="outputs",
+                specs=[
+                    FeatureSpec(name="sequence", dtype=DataType.STRING),
+                    FeatureSpec(name="score", dtype=DataType.DOUBLE),
+                    FeatureSpec(name="token", dtype=DataType.INT64),
+                    FeatureSpec(name="token_str", dtype=DataType.STRING),
+                ],
+                shape=(-1,),
+            ),
+        ],
+    )
+    ```
+
+  - Signature for token-classification task has been changed from
+
+    ```python
+    ModelSignature(
+        inputs=[
+            FeatureSpec(name="inputs", dtype=DataType.STRING),
+        ],
+        outputs=[
+            FeatureSpec(name="outputs", dtype=DataType.STRING),
+        ],
+    )
+    ```
+
+    to
+
+    ```python
+    ModelSignature(
+        inputs=[FeatureSpec(name="inputs", dtype=DataType.STRING)],
+        outputs=[
+            FeatureGroupSpec(
+                name="outputs",
+                specs=[
+                    FeatureSpec(name="word", dtype=DataType.STRING),
+                    FeatureSpec(name="score", dtype=DataType.DOUBLE),
+                    FeatureSpec(name="entity", dtype=DataType.STRING),
+                    FeatureSpec(name="index", dtype=DataType.INT64),
+                    FeatureSpec(name="start", dtype=DataType.INT64),
+                    FeatureSpec(name="end", dtype=DataType.INT64),
+                ],
+                shape=(-1,),
+            ),
+        ],
+    )
+    ```
+
+  - Signature for question-answering task when top_k is larger than 1 has been changed from
+
+    ```python
+    ModelSignature(
+        inputs=[
+            FeatureSpec(name="question", dtype=DataType.STRING),
+            FeatureSpec(name="context", dtype=DataType.STRING),
+        ],
+        outputs=[
+            FeatureSpec(name="outputs", dtype=DataType.STRING),
+        ],
+    )
+    ```
+
+    to
+
+    ```python
+    ModelSignature(
+        inputs=[
+            FeatureSpec(name="question", dtype=DataType.STRING),
+            FeatureSpec(name="context", dtype=DataType.STRING),
+        ],
+        outputs=[
+            FeatureGroupSpec(
+                name="answers",
+                specs=[
+                    FeatureSpec(name="score", dtype=DataType.DOUBLE),
+                    FeatureSpec(name="start", dtype=DataType.INT64),
+                    FeatureSpec(name="end", dtype=DataType.INT64),
+                    FeatureSpec(name="answer", dtype=DataType.STRING),
+                ],
+                shape=(-1,),
+            ),
+        ],
+    )
+    ```
+
+  - Signature for text-classification task when top_k is `None` has been changed from
+
+    ```python
+    ModelSignature(
+        inputs=[
+            FeatureSpec(name="text", dtype=DataType.STRING),
+            FeatureSpec(name="text_pair", dtype=DataType.STRING),
+        ],
+        outputs=[
+            FeatureSpec(name="label", dtype=DataType.STRING),
+            FeatureSpec(name="score", dtype=DataType.DOUBLE),
+        ],
+    )
+    ```
+
+    to
+
+    ```python
+    ModelSignature(
+        inputs=[
+            FeatureSpec(name="text", dtype=DataType.STRING),
+        ],
+        outputs=[
+            FeatureSpec(name="label", dtype=DataType.STRING),
+            FeatureSpec(name="score", dtype=DataType.DOUBLE),
+        ],
+    )
+    ```
+
+  - Signature for text-classification task when top_k is not `None` has been changed from
+
+    ```python
+    ModelSignature(
+        inputs=[
+            FeatureSpec(name="text", dtype=DataType.STRING),
+            FeatureSpec(name="text_pair", dtype=DataType.STRING),
+        ],
+        outputs=[
+            FeatureSpec(name="outputs", dtype=DataType.STRING),
+        ],
+    )
+    ```
+
+    to
+
+    ```python
+    ModelSignature(
+        inputs=[
+            FeatureSpec(name="text", dtype=DataType.STRING),
+        ],
+        outputs=[
+            FeatureGroupSpec(
+                name="labels",
+                specs=[
+                    FeatureSpec(name="label", dtype=DataType.STRING),
+                    FeatureSpec(name="score", dtype=DataType.DOUBLE),
+                ],
+                shape=(-1,),
+            ),
+        ],
+    )
+    ```
+
+  - Signature for text-generation task has been changed from
+
+    ```python
+    ModelSignature(
+        inputs=[FeatureSpec(name="inputs", dtype=DataType.STRING)],
+        outputs=[
+            FeatureSpec(name="outputs", dtype=DataType.STRING),
+        ],
+    )
+    ```
+
+    to
+
+    ```python
+    ModelSignature(
+        inputs=[
+            FeatureGroupSpec(
+                name="inputs",
+                specs=[
+                    FeatureSpec(name="role", dtype=DataType.STRING),
+                    FeatureSpec(name="content", dtype=DataType.STRING),
+                ],
+                shape=(-1,),
+            ),
+        ],
+        outputs=[
+            FeatureGroupSpec(
+                name="outputs",
+                specs=[
+                    FeatureSpec(name="generated_text", dtype=DataType.STRING),
+                ],
+                shape=(-1,),
+            )
+        ],
+    )
+    ```
+
+- Registry: PyTorch and TensorFlow models now expect a single tensor input/output by default when logging to Model
+  Registry. To use multiple tensors (previous behavior), set `options={"multiple_inputs": True}`.
+
+  Example with single tensor input:
+
+  ```python
+  import torch
+
+  class TorchModel(torch.nn.Module):
+      def __init__(self, n_input: int, n_hidden: int, n_out: int, dtype: torch.dtype = torch.float32) -> None:
+          super().__init__()
+          self.model = torch.nn.Sequential(
+              torch.nn.Linear(n_input, n_hidden, dtype=dtype),
+              torch.nn.ReLU(),
+              torch.nn.Linear(n_hidden, n_out, dtype=dtype),
+              torch.nn.Sigmoid(),
+          )
+
+      def forward(self, tensor: torch.Tensor) -> torch.Tensor:
+          return cast(torch.Tensor, self.model(tensor))
+
+  # Sample usage:
+  data_x = torch.rand(size=(batch_size, n_input))
+
+  # Log model with single tensor
+  reg.log_model(
+      model=model,
+      ...,
+      sample_input_data=data_x
+  )
+
+  # Run inference with single tensor
+  mv.run(data_x)
+  ```
+
+  For multiple tensor inputs/outputs, use:
+
+  ```python
+  reg.log_model(
+      model=model,
+      ...,
+      sample_input_data=[data_x_1, data_x_2],
+      options={"multiple_inputs": True}
+  )
+  ```
+
+- Registry: Default `enable_explainability` to False when the model can be deployed to Snowpark Container Services.
+
+### New Features
+
+- Registry: Added support to single `torch.Tensor`, `tensorflow.Tensor` and `tensorflow.Variable` as input or output
+  data.
+- Registry: Support [`xgboost.DMatrix`](https://xgboost.readthedocs.io/en/stable/python/python_api.html#xgboost.DMatrix)
+ datatype for XGBoost models.
+
+## 1.7.5 (03-06-2025)
 
 - Support Python 3.12.
-- Explainability: Support native and snowml sklearn pipeline
+- Explainability: Support native and snowflake.ml.modeling sklearn pipeline
 
 ### Bug Fixes
 
@@ -421,8 +699,6 @@ NOTE: Version 1.7.0 is used as example here. Please choose the the latest versio
   `ValueError(f"{self.entrypoint} must be a subpath of {self.source}")`
 - ML Job (PrPr): Fixed a bug in Ray cluster startup config which caused certain Runtime APIs to fail
 
-### Behavior Change
-
 ### New Features
 
 - Registry: Added support for handling Hugging Face model configurations with auto-mapping functionality.