zenml-nightly 0.73.0.dev20250130__tar.gz → 0.73.0.dev20250131__tar.gz

{zenml_nightly-0.73.0.dev20250130 → zenml_nightly-0.73.0.dev20250131}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: zenml-nightly
-Version: 0.73.0.dev20250130
+Version: 0.73.0.dev20250131
 Summary: ZenML: Write production-ready ML code.
 License: Apache-2.0
 Keywords: machine learning,production,pipeline,mlops,devops

{zenml_nightly-0.73.0.dev20250130 → zenml_nightly-0.73.0.dev20250131}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "zenml-nightly"
-version = "0.73.0.dev20250130"
+version = "0.73.0.dev20250131"
 packages = [{ include = "zenml", from = "src" }]
 description = "ZenML: Write production-ready ML code."
 authors = ["ZenML GmbH <info@zenml.io>"]

zenml_nightly-0.73.0.dev20250131/src/zenml/VERSION

@@ -0,0 +1 @@
+0.73.0.dev20250131

{zenml_nightly-0.73.0.dev20250130 → zenml_nightly-0.73.0.dev20250131}/src/zenml/cli/code_repository.py

@@ -162,6 +162,32 @@ def register_code_repository(
         cli_utils.declare(f"Successfully registered code repository `{name}`.")
 
 
+@code_repository.command("describe", help="Describe a code repository.")
+@click.argument(
+    "name_id_or_prefix",
+    type=str,
+    required=True,
+)
+def describe_code_repository(name_id_or_prefix: str) -> None:
+    """Describe a code repository.
+
+    Args:
+        name_id_or_prefix: Name, ID or prefix of the code repository.
+    """
+    client = Client()
+    try:
+        code_repository = client.get_code_repository(
+            name_id_or_prefix=name_id_or_prefix,
+        )
+    except KeyError as err:
+        cli_utils.error(str(err))
+    else:
+        cli_utils.print_pydantic_model(
+            title=f"Code repository '{code_repository.name}'",
+            model=code_repository,
+        )
+
+
 @code_repository.command("list", help="List all connected code repositories.")
 @list_options(CodeRepositoryFilter)
 def list_code_repositories(**kwargs: Any) -> None:

{zenml_nightly-0.73.0.dev20250130 → zenml_nightly-0.73.0.dev20250131}/src/zenml/client.py

@@ -34,7 +34,7 @@ from typing import (
     Union,
     cast,
 )
-from uuid import UUID, uuid4
+from uuid import UUID
 
 from pydantic import ConfigDict, SecretStr
 
@@ -4980,12 +4980,7 @@ class Client(metaclass=ClientMetaClass):
             )
         )
         try:
-            # This does a login to verify the credentials
-            code_repo_class(id=uuid4(), config=config)
-
-            # Explicitly access the config for pydantic validation, in case
-            # the login for some reason did not do that.
-            _ = code_repo_class.config
+            code_repo_class.validate_config(config)
         except Exception as e:
             raise RuntimeError(
                 "Failed to validate code repository config."

{zenml_nightly-0.73.0.dev20250130 → zenml_nightly-0.73.0.dev20250131}/src/zenml/code_repositories/base_code_repository.py

@@ -15,7 +15,7 @@
 
 from abc import ABC, abstractmethod
 from typing import TYPE_CHECKING, Any, Dict, Optional, Set, Type
-from uuid import UUID
+from uuid import UUID, uuid4
 
 from zenml.config.secret_reference_mixin import SecretReferenceMixin
 from zenml.logger import get_logger
@@ -44,15 +44,18 @@ class BaseCodeRepository(ABC):
     def __init__(
         self,
         id: UUID,
+        name: str,
         config: Dict[str, Any],
     ) -> None:
         """Initializes a code repository.
 
         Args:
             id: The ID of the code repository.
+            name: The name of the code repository.
             config: The config of the code repository.
         """
         self._id = id
+        self._name = name
         self._config = config
         self.login()
 
@@ -80,7 +83,23 @@ class BaseCodeRepository(ABC):
                 source=model.source, expected_class=BaseCodeRepository
             )
         )
-        return class_(id=model.id, config=model.config)
+        return class_(id=model.id, name=model.name, config=model.config)
+
+    @classmethod
+    def validate_config(cls, config: Dict[str, Any]) -> None:
+        """Validate the code repository config.
+
+        This method should check that the config/credentials are valid and
+        the configured repository exists.
+
+        Args:
+            config: The configuration.
+        """
+        # The initialization calls the login to verify the credentials
+        code_repo = cls(id=uuid4(), name="", config=config)
+
+        # Explicitly access the config for pydantic validation
+        _ = code_repo.config
 
     @property
     def id(self) -> UUID:
@@ -91,6 +110,15 @@ class BaseCodeRepository(ABC):
         """
         return self._id
 
+    @property
+    def name(self) -> str:
+        """Name of the code repository.
+
+        Returns:
+            The name of the code repository.
+        """
+        return self._name
+
     @property
     def requirements(self) -> Set[str]:
         """Set of PyPI requirements for the repository.

{zenml_nightly-0.73.0.dev20250130 → zenml_nightly-0.73.0.dev20250131}/src/zenml/code_repositories/git/local_git_repository_context.py

@@ -14,11 +14,14 @@
 """Implementation of the Local git repository context."""
 
 from typing import TYPE_CHECKING, Callable, Optional, cast
-from uuid import UUID
 
 from zenml.code_repositories import (
     LocalRepositoryContext,
 )
+from zenml.constants import (
+    ENV_ZENML_CODE_REPOSITORY_IGNORE_UNTRACKED_FILES,
+    handle_bool_env_var,
+)
 from zenml.logger import get_logger
 
 if TYPE_CHECKING:
@@ -26,6 +29,8 @@ if TYPE_CHECKING:
     from git.remote import Remote
     from git.repo.base import Repo
 
+    from zenml.code_repositories import BaseCodeRepository
+
 logger = get_logger(__name__)
 
 
@@ -33,16 +38,19 @@ class LocalGitRepositoryContext(LocalRepositoryContext):
     """Local git repository context."""
 
     def __init__(
-        self, code_repository_id: UUID, git_repo: "Repo", remote_name: str
+        self,
+        code_repository: "BaseCodeRepository",
+        git_repo: "Repo",
+        remote_name: str,
     ):
         """Initializes a local git repository context.
 
         Args:
-            code_repository_id: The ID of the code repository.
+            code_repository: The code repository.
             git_repo: The git repo.
             remote_name: Name of the remote.
         """
-        super().__init__(code_repository_id=code_repository_id)
+        super().__init__(code_repository=code_repository)
         self._git_repo = git_repo
         self._remote = git_repo.remote(name=remote_name)
 
@@ -50,14 +58,14 @@ class LocalGitRepositoryContext(LocalRepositoryContext):
     def at(
         cls,
         path: str,
-        code_repository_id: UUID,
+        code_repository: "BaseCodeRepository",
         remote_url_validation_callback: Callable[[str], bool],
     ) -> Optional["LocalGitRepositoryContext"]:
         """Returns a local git repository at the given path.
 
         Args:
             path: The path to the local git repository.
-            code_repository_id: The ID of the code repository.
+            code_repository: The code repository.
             remote_url_validation_callback: A callback that validates the
                 remote URL of the git repository.
 
@@ -70,11 +78,13 @@ class LocalGitRepositoryContext(LocalRepositoryContext):
             from git.exc import InvalidGitRepositoryError
             from git.repo.base import Repo
         except ImportError:
+            logger.debug("Failed to import git library.")
             return None
 
         try:
             git_repo = Repo(path=path, search_parent_directories=True)
         except InvalidGitRepositoryError:
+            logger.debug("No git repository exists at path %s.", path)
             return None
 
         remote_name = None
@@ -87,7 +97,7 @@ class LocalGitRepositoryContext(LocalRepositoryContext):
             return None
 
         return LocalGitRepositoryContext(
-            code_repository_id=code_repository_id,
+            code_repository=code_repository,
             git_repo=git_repo,
             remote_name=remote_name,
         )
@@ -124,13 +134,19 @@ class LocalGitRepositoryContext(LocalRepositoryContext):
     def is_dirty(self) -> bool:
         """Whether the git repo is dirty.
 
-        A repository counts as dirty if it has any untracked or uncommitted
-        changes.
+        By default, a repository counts as dirty if it has any untracked or
+        uncommitted changes. Users can use an environment variable to ignore
+        untracked files.
 
         Returns:
             True if the git repo is dirty, False otherwise.
         """
-        return self.git_repo.is_dirty(untracked_files=True)
+        ignore_untracked_files = handle_bool_env_var(
+            ENV_ZENML_CODE_REPOSITORY_IGNORE_UNTRACKED_FILES, default=False
+        )
+        return self.git_repo.is_dirty(
+            untracked_files=not ignore_untracked_files
+        )
 
     @property
     def has_local_changes(self) -> bool:

{zenml_nightly-0.73.0.dev20250130 → zenml_nightly-0.73.0.dev20250131}/src/zenml/code_repositories/local_repository_context.py

@@ -14,10 +14,13 @@
 """Base class for local code repository contexts."""
 
 from abc import ABC, abstractmethod
-from uuid import UUID
+from typing import TYPE_CHECKING
 
 from zenml.logger import get_logger
 
+if TYPE_CHECKING:
+    from zenml.code_repositories import BaseCodeRepository
+
 logger = get_logger(__name__)
 
 
@@ -30,22 +33,22 @@ class LocalRepositoryContext(ABC):
     commit, and whether the repository is dirty.
     """
 
-    def __init__(self, code_repository_id: UUID) -> None:
+    def __init__(self, code_repository: "BaseCodeRepository") -> None:
         """Initializes a local repository context.
 
         Args:
-            code_repository_id: The ID of the code repository.
+            code_repository: The code repository.
         """
-        self._code_repository_id = code_repository_id
+        self._code_repository = code_repository
 
     @property
-    def code_repository_id(self) -> UUID:
-        """Returns the ID of the code repository.
+    def code_repository(self) -> "BaseCodeRepository":
+        """Returns the code repository.
 
         Returns:
-            The ID of the code repository.
+            The code repository.
         """
-        return self._code_repository_id
+        return self._code_repository
 
     @property
     @abstractmethod

{zenml_nightly-0.73.0.dev20250130 → zenml_nightly-0.73.0.dev20250131}/src/zenml/constants.py

@@ -175,6 +175,9 @@ ENV_ZENML_WHEEL_PACKAGE_NAME = "ZENML_WHEEL_PACKAGE_NAME"
 ENV_ZENML_PIPELINE_RUN_API_TOKEN_EXPIRATION = (
     "ZENML_PIPELINE_API_TOKEN_EXPIRATION"
 )
+ENV_ZENML_CODE_REPOSITORY_IGNORE_UNTRACKED_FILES = (
+    "ZENML_CODE_REPOSITORY_IGNORE_UNTRACKED_FILES"
+)
 
 # Materializer environment variables
 ENV_ZENML_MATERIALIZER_ALLOW_NON_ASCII_JSON_DUMPS = (

{zenml_nightly-0.73.0.dev20250130 → zenml_nightly-0.73.0.dev20250131}/src/zenml/integrations/github/code_repositories/github_code_repository.py

@@ -15,7 +15,8 @@
 
 import os
 import re
-from typing import List, Optional
+from typing import Any, Dict, List, Optional
+from uuid import uuid4
 
 import requests
 from github import Consts, Github, GithubException
@@ -62,6 +63,20 @@ class GitHubCodeRepositoryConfig(BaseCodeRepositoryConfig):
 class GitHubCodeRepository(BaseCodeRepository):
     """GitHub code repository."""
 
+    @classmethod
+    def validate_config(cls, config: Dict[str, Any]) -> None:
+        """Validate the code repository config.
+
+        This method should check that the config/credentials are valid and
+        the configured repository exists.
+
+        Args:
+            config: The configuration.
+        """
+        code_repo = cls(id=uuid4(), name="", config=config)
+        # Try to access the project to make sure it exists
+        _ = code_repo.github_repo
+
     @property
     def config(self) -> GitHubCodeRepositoryConfig:
         """Returns the `GitHubCodeRepositoryConfig` config.
@@ -190,7 +205,7 @@ class GitHubCodeRepository(BaseCodeRepository):
         """
         return LocalGitRepositoryContext.at(
             path=path,
-            code_repository_id=self.id,
+            code_repository=self,
             remote_url_validation_callback=self.check_remote_url,
         )
 

{zenml_nightly-0.73.0.dev20250130 → zenml_nightly-0.73.0.dev20250131}/src/zenml/integrations/gitlab/code_repositories/gitlab_code_repository.py

@@ -15,7 +15,8 @@
 
 import os
 import re
-from typing import Optional
+from typing import Any, Dict, Optional
+from uuid import uuid4
 
 from gitlab import Gitlab
 from gitlab.v4.objects import Project
@@ -63,6 +64,20 @@ class GitLabCodeRepositoryConfig(BaseCodeRepositoryConfig):
 class GitLabCodeRepository(BaseCodeRepository):
     """GitLab code repository."""
 
+    @classmethod
+    def validate_config(cls, config: Dict[str, Any]) -> None:
+        """Validate the code repository config.
+
+        This method should check that the config/credentials are valid and
+        the configured repository exists.
+
+        Args:
+            config: The configuration.
+        """
+        code_repo = cls(id=uuid4(), name="", config=config)
+        # Try to access the project to make sure it exists
+        _ = code_repo.gitlab_project
+
     @property
     def config(self) -> GitLabCodeRepositoryConfig:
         """Returns the `GitLabCodeRepositoryConfig` config.
@@ -147,7 +162,7 @@ class GitLabCodeRepository(BaseCodeRepository):
         """
         return LocalGitRepositoryContext.at(
             path=path,
-            code_repository_id=self.id,
+            code_repository=self,
             remote_url_validation_callback=self.check_remote_url,
         )
 

{zenml_nightly-0.73.0.dev20250130 → zenml_nightly-0.73.0.dev20250131}/src/zenml/integrations/pytorch/materializers/base_pytorch_materializer.py

@@ -44,7 +44,7 @@ class BasePyTorchMaterializer(BaseMaterializer):
             # NOTE (security): The `torch.load` function uses `pickle` as
             # the default unpickler, which is NOT secure. This materializer
             # is intended for use with trusted data sources.
-            return torch.load(f)  # nosec
+            return torch.load(f, weights_only=False)  # nosec
 
     def save(self, obj: Any) -> None:
         """Uses `torch.save` to save a PyTorch object.

{zenml_nightly-0.73.0.dev20250130 → zenml_nightly-0.73.0.dev20250131}/src/zenml/pipelines/build_utils.py

@@ -517,32 +517,9 @@ def verify_local_repository_context(
                     "changes."
                 )
 
-        if local_repo_context:
-            if local_repo_context.is_dirty:
-                logger.warning(
-                    "Unable to use code repository to download code for this "
-                    "run as there are uncommitted changes."
-                )
-            elif local_repo_context.has_local_changes:
-                logger.warning(
-                    "Unable to use code repository to download code for this "
-                    "run as there are unpushed changes."
-                )
-
     code_repository = None
     if local_repo_context and not local_repo_context.has_local_changes:
-        model = Client().get_code_repository(
-            local_repo_context.code_repository_id
-        )
-        code_repository = BaseCodeRepository.from_model(model)
-
-        if will_download_from_code_repository(
-            deployment=deployment, local_repo_context=local_repo_context
-        ):
-            logger.info(
-                "Using code repository `%s` to download code for this run.",
-                model.name,
-            )
+        code_repository = local_repo_context.code_repository
 
     return code_repository
 
@@ -738,25 +715,17 @@ def should_upload_code(
     return False
 
 
-def will_download_from_code_repository(
+def allows_download_from_code_repository(
     deployment: PipelineDeploymentBase,
-    local_repo_context: "LocalRepositoryContext",
 ) -> bool:
-    """Checks whether a code repository will be used to download code.
+    """Checks whether a code repository can be used to download code.
 
     Args:
         deployment: The deployment.
-        local_repo_context: The local repository context.
 
     Returns:
-        Whether a code repository will be used to download code.
+        Whether a code repository can be used to download code.
     """
-    if not build_required(deployment=deployment):
-        return False
-
-    if local_repo_context.has_local_changes:
-        return False
-
     for step in deployment.step_configurations.values():
         docker_settings = step.config.docker_settings
 
@@ -764,3 +733,41 @@ def will_download_from_code_repository(
             return True
 
     return False
+
+
+def log_code_repository_usage(
+    deployment: PipelineDeploymentBase,
+    local_repo_context: "LocalRepositoryContext",
+) -> None:
+    """Log what the code repository can (not) be used for given a deployment.
+
+    Args:
+        deployment: The deployment.
+        local_repo_context: The local repository context.
+    """
+    if build_required(deployment) and allows_download_from_code_repository(
+        deployment
+    ):
+        if local_repo_context.is_dirty:
+            logger.warning(
+                "Unable to use code repository `%s` to download code or track "
+                "the commit hash as there are uncommitted or untracked files.",
+                local_repo_context.code_repository.name,
+            )
+        elif local_repo_context.has_local_changes:
+            logger.warning(
+                "Unable to use code repository `%s` to download code as there "
+                "are unpushed commits.",
+                local_repo_context.code_repository.name,
+            )
+        else:
+            logger.info(
+                "Using code repository `%s` to download code for this run.",
+                local_repo_context.code_repository.name,
+            )
+    elif local_repo_context.is_dirty:
+        logger.warning(
+            "Unable to use code repository `%s` to track the commit hash as "
+            "there are uncommitted or untracked files.",
+            local_repo_context.code_repository.name,
+        )

{zenml_nightly-0.73.0.dev20250130 → zenml_nightly-0.73.0.dev20250131}/src/zenml/pipelines/pipeline_definition.py

@@ -643,7 +643,6 @@ To avoid this consider setting pipeline parameters only in one place (config or
         pipeline_id = None
         if register_pipeline:
             pipeline_id = self._register().id
-
         else:
             logger.debug(f"Pipeline {self.name} is unlisted.")
 
@@ -702,6 +701,10 @@ To avoid this consider setting pipeline parameters only in one place (config or
             deployment=deployment, local_repo_context=local_repo_context
         )
         can_download_from_code_repository = code_repository is not None
+        if local_repo_context:
+            build_utils.log_code_repository_usage(
+                deployment=deployment, local_repo_context=local_repo_context
+            )
 
         if prevent_build_reuse:
             logger.warning(
@@ -731,7 +734,7 @@ To avoid this consider setting pipeline parameters only in one place (config or
             code_reference = CodeReferenceRequest(
                 commit=local_repo_context.current_commit,
                 subdirectory=subdirectory.as_posix(),
-                code_repository=local_repo_context.code_repository_id,
+                code_repository=local_repo_context.code_repository.id,
             )
 
         code_path = None

{zenml_nightly-0.73.0.dev20250130 → zenml_nightly-0.73.0.dev20250131}/src/zenml/utils/code_repository_utils.py

@@ -79,7 +79,7 @@ def set_custom_local_repository(
 
     path = os.path.abspath(source_utils.get_source_root())
     _CODE_REPOSITORY_CACHE[path] = _DownloadedRepositoryContext(
-        code_repository_id=repo.id, root=root, commit=commit
+        code_repository=repo, root=root, commit=commit
     )
 
 
@@ -106,7 +106,8 @@ def find_active_code_repository(
         return _CODE_REPOSITORY_CACHE[path]
 
     local_context: Optional["LocalRepositoryContext"] = None
-    for model in depaginate(list_method=Client().list_code_repositories):
+    code_repositories = depaginate(list_method=Client().list_code_repositories)
+    for model in code_repositories:
         try:
             repo = BaseCodeRepository.from_model(model)
         except ImportError:
@@ -125,6 +126,14 @@ def find_active_code_repository(
         local_context = repo.get_local_context(path)
         if local_context:
             break
+    else:
+        if code_repositories:
+            # There are registered code repositories, but none was matching the
+            # current path -> We log the path to help in debugging issues
+            # related to the source root.
+            logger.info(
+                "No matching code repository found for path `%s`.", path
+            )
 
     _CODE_REPOSITORY_CACHE[path] = local_context
     return local_context

{zenml_nightly-0.73.0.dev20250130 → zenml_nightly-0.73.0.dev20250131}/src/zenml/utils/downloaded_repository_context.py

@@ -13,9 +13,7 @@
 #  permissions and limitations under the License.
 """Downloaded code repository."""
 
-from uuid import UUID
-
-from zenml.code_repositories import LocalRepositoryContext
+from zenml.code_repositories import BaseCodeRepository, LocalRepositoryContext
 
 
 class _DownloadedRepositoryContext(LocalRepositoryContext):
@@ -27,11 +25,11 @@ class _DownloadedRepositoryContext(LocalRepositoryContext):
 
     def __init__(
         self,
-        code_repository_id: UUID,
+        code_repository: BaseCodeRepository,
         root: str,
         commit: str,
     ):
-        super().__init__(code_repository_id=code_repository_id)
+        super().__init__(code_repository=code_repository)
         self._root = root
         self._commit = commit
 

{zenml_nightly-0.73.0.dev20250130 → zenml_nightly-0.73.0.dev20250131}/src/zenml/utils/source_utils.py

@@ -226,7 +226,7 @@ def resolve(
             subdir = PurePath(source_root).relative_to(local_repo_context.root)
 
             return CodeRepositorySource(
-                repository_id=local_repo_context.code_repository_id,
+                repository_id=local_repo_context.code_repository.id,
                 commit=local_repo_context.current_commit,
                 subdirectory=subdir.as_posix(),
                 module=module_name,
@@ -482,7 +482,7 @@ def _warn_about_potential_source_loading_issues(
             source.repository_id,
             get_source_root(),
         )
-    elif local_repo.code_repository_id != source.repository_id:
+    elif local_repo.code_repository.id != source.repository_id:
         logger.warning(
             "Potential issue when loading the source `%s`: The source "
             "references the code repository `%s` but there is a different "
@@ -492,7 +492,7 @@ def _warn_about_potential_source_loading_issues(
             "source was originally stored.",
             source.import_path,
             source.repository_id,
-            local_repo.code_repository_id,
+            local_repo.code_repository.id,
             get_source_root(),
         )
     elif local_repo.current_commit != source.commit: