huggingface-hub 0.29.0rc2__py3-none-any.whl → 1.1.3__py3-none-any.whl

huggingface_hub/_inference_endpoints.py

@@ -2,18 +2,17 @@ import time
 from dataclasses import dataclass, field
 from datetime import datetime
 from enum import Enum
-from typing import TYPE_CHECKING, Dict, Optional, Union
+from typing import TYPE_CHECKING, Optional, Union
 
 from huggingface_hub.errors import InferenceEndpointError, InferenceEndpointTimeoutError
 
-from .inference._client import InferenceClient
-from .inference._generated._async_client import AsyncInferenceClient
 from .utils import get_session, logging, parse_datetime
 
 
 if TYPE_CHECKING:
     from .hf_api import HfApi
-
+    from .inference._client import InferenceClient
+    from .inference._generated._async_client import AsyncInferenceClient
 
 logger = logging.get_logger(__name__)
 
@@ -63,7 +62,7 @@ class InferenceEndpoint:
             The timestamp of the last update of the Inference Endpoint.
         type ([`InferenceEndpointType`]):
             The type of the Inference Endpoint (public, protected, private).
-        raw (`Dict`):
+        raw (`dict`):
             The raw dictionary data returned from the API.
         token (`str` or `bool`, *optional*):
             Authentication token for the Inference Endpoint, if set when requesting the API. Will default to the
@@ -101,6 +100,7 @@ class InferenceEndpoint:
     namespace: str
     repository: str = field(init=False)
     status: InferenceEndpointStatus = field(init=False)
+    health_route: str = field(init=False)
     url: Optional[str] = field(init=False)
 
     # Other fields
@@ -112,7 +112,7 @@ class InferenceEndpoint:
     type: InferenceEndpointType = field(repr=False, init=False)
 
     # Raw dict from the API
-    raw: Dict = field(repr=False)
+    raw: dict = field(repr=False)
 
     # Internal fields
     _token: Union[str, bool, None] = field(repr=False, compare=False)
@@ -120,7 +120,7 @@ class InferenceEndpoint:
 
     @classmethod
     def from_raw(
-        cls, raw: Dict, namespace: str, token: Union[str, bool, None] = None, api: Optional["HfApi"] = None
+        cls, raw: dict, namespace: str, token: Union[str, bool, None] = None, api: Optional["HfApi"] = None
     ) -> "InferenceEndpoint":
         """Initialize object from raw dictionary."""
         if api is None:
@@ -138,7 +138,7 @@ class InferenceEndpoint:
         self._populate_from_raw()
 
     @property
-    def client(self) -> InferenceClient:
+    def client(self) -> "InferenceClient":
         """Returns a client to make predictions on this Inference Endpoint.
 
         Returns:
@@ -152,13 +152,15 @@ class InferenceEndpoint:
                 "Cannot create a client for this Inference Endpoint as it is not yet deployed. "
                 "Please wait for the Inference Endpoint to be deployed using `endpoint.wait()` and try again."
             )
+        from .inference._client import InferenceClient
+
         return InferenceClient(
             model=self.url,
             token=self._token,  # type: ignore[arg-type] # boolean token shouldn't be possible. In practice it's ok.
         )
 
     @property
-    def async_client(self) -> AsyncInferenceClient:
+    def async_client(self) -> "AsyncInferenceClient":
         """Returns a client to make predictions on this Inference Endpoint.
 
         Returns:
@@ -172,6 +174,8 @@ class InferenceEndpoint:
                 "Cannot create a client for this Inference Endpoint as it is not yet deployed. "
                 "Please wait for the Inference Endpoint to be deployed using `endpoint.wait()` and try again."
             )
+        from .inference._generated._async_client import AsyncInferenceClient
+
         return AsyncInferenceClient(
             model=self.url,
             token=self._token,  # type: ignore[arg-type] # boolean token shouldn't be possible. In practice it's ok.
@@ -207,16 +211,22 @@ class InferenceEndpoint:
 
         start = time.time()
         while True:
-            if self.url is not None:
-                # Means the URL is provisioned => check if the endpoint is reachable
-                response = get_session().get(self.url, headers=self._api._build_hf_headers(token=self._token))
-                if response.status_code == 200:
-                    logger.info("Inference Endpoint is ready to be used.")
-                    return self
             if self.status == InferenceEndpointStatus.FAILED:
                 raise InferenceEndpointError(
                     f"Inference Endpoint {self.name} failed to deploy. Please check the logs for more information."
                 )
+            if self.status == InferenceEndpointStatus.UPDATE_FAILED:
+                raise InferenceEndpointError(
+                    f"Inference Endpoint {self.name} failed to update. Please check the logs for more information."
+                )
+            if self.status == InferenceEndpointStatus.RUNNING and self.url is not None:
+                # Verify the endpoint is actually reachable
+                _health_url = f"{self.url.rstrip('/')}/{self.health_route.lstrip('/')}"
+                response = get_session().get(_health_url, headers=self._api._build_hf_headers(token=self._token))
+                if response.status_code == 200:
+                    logger.info("Inference Endpoint is ready to be used.")
+                    return self
+
             if timeout is not None:
                 if time.time() - start > timeout:
                     raise InferenceEndpointTimeoutError("Timeout while waiting for Inference Endpoint to be deployed.")
@@ -250,8 +260,8 @@ class InferenceEndpoint:
         framework: Optional[str] = None,
         revision: Optional[str] = None,
         task: Optional[str] = None,
-        custom_image: Optional[Dict] = None,
-        secrets: Optional[Dict[str, str]] = None,
+        custom_image: Optional[dict] = None,
+        secrets: Optional[dict[str, str]] = None,
     ) -> "InferenceEndpoint":
         """Update the Inference Endpoint.
 
@@ -283,10 +293,10 @@ class InferenceEndpoint:
                 The specific model revision to deploy on the Inference Endpoint (e.g. `"6c0e6080953db56375760c0471a8c5f2929baf11"`).
             task (`str`, *optional*):
                 The task on which to deploy the model (e.g. `"text-classification"`).
-            custom_image (`Dict`, *optional*):
+            custom_image (`dict`, *optional*):
                 A custom Docker image to use for the Inference Endpoint. This is useful if you want to deploy an
                 Inference Endpoint running on the `text-generation-inference` (TGI) framework (see examples).
-            secrets (`Dict[str, str]`, *optional*):
+            secrets (`dict[str, str]`, *optional*):
                 Secret values to inject in the container environment.
         Returns:
             [`InferenceEndpoint`]: the same Inference Endpoint, mutated in place with the latest data.
@@ -319,7 +329,7 @@ class InferenceEndpoint:
         """Pause the Inference Endpoint.
 
         A paused Inference Endpoint will not be charged. It can be resumed at any time using [`InferenceEndpoint.resume`].
-        This is different than scaling the Inference Endpoint to zero with [`InferenceEndpoint.scale_to_zero`], which
+        This is different from scaling the Inference Endpoint to zero with [`InferenceEndpoint.scale_to_zero`], which
         would be automatically restarted when a request is made to it.
 
         This is an alias for [`HfApi.pause_inference_endpoint`]. The current object is mutated in place with the
@@ -357,8 +367,8 @@ class InferenceEndpoint:
     def scale_to_zero(self) -> "InferenceEndpoint":
         """Scale Inference Endpoint to zero.
 
-        An Inference Endpoint scaled to zero will not be charged. It will be resume on the next request to it, with a
-        cold start delay. This is different than pausing the Inference Endpoint with [`InferenceEndpoint.pause`], which
+        An Inference Endpoint scaled to zero will not be charged. It will be resumed on the next request to it, with a
+        cold start delay. This is different from pausing the Inference Endpoint with [`InferenceEndpoint.pause`], which
         would require a manual resume with [`InferenceEndpoint.resume`].
 
         This is an alias for [`HfApi.scale_to_zero_inference_endpoint`]. The current object is mutated in place with the
@@ -392,6 +402,7 @@ class InferenceEndpoint:
         self.repository = self.raw["model"]["repository"]
         self.status = self.raw["status"]["state"]
         self.url = self.raw["status"].get("url")
+        self.health_route = self.raw["healthRoute"]
 
         # Other fields
         self.framework = self.raw["model"]["framework"]

huggingface_hub/_jobs_api.py

@@ -0,0 +1,301 @@
+# coding=utf-8
+# Copyright 2025-present, the HuggingFace Inc. team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+from dataclasses import dataclass
+from datetime import datetime
+from enum import Enum
+from typing import Any, Optional, Union
+
+from huggingface_hub import constants
+from huggingface_hub._space_api import SpaceHardware
+from huggingface_hub.utils._datetime import parse_datetime
+
+
+class JobStage(str, Enum):
+    """
+    Enumeration of possible stage of a Job on the Hub.
+
+    Value can be compared to a string:
+    ```py
+    assert JobStage.COMPLETED == "COMPLETED"
+    ```
+    Possible values are: `COMPLETED`, `CANCELED`, `ERROR`, `DELETED`, `RUNNING`.
+    Taken from https://github.com/huggingface/moon-landing/blob/main/server/job_types/JobInfo.ts#L61 (private url).
+    """
+
+    # Copied from moon-landing > server > lib > Job.ts
+    COMPLETED = "COMPLETED"
+    CANCELED = "CANCELED"
+    ERROR = "ERROR"
+    DELETED = "DELETED"
+    RUNNING = "RUNNING"
+
+
+@dataclass
+class JobStatus:
+    stage: JobStage
+    message: Optional[str]
+
+
+@dataclass
+class JobOwner:
+    id: str
+    name: str
+    type: str
+
+
+@dataclass
+class JobInfo:
+    """
+    Contains information about a Job.
+
+    Args:
+        id (`str`):
+            Job ID.
+        created_at (`datetime` or `None`):
+            When the Job was created.
+        docker_image (`str` or `None`):
+            The Docker image from Docker Hub used for the Job.
+            Can be None if space_id is present instead.
+        space_id (`str` or `None`):
+            The Docker image from Hugging Face Spaces used for the Job.
+            Can be None if docker_image is present instead.
+        command (`list[str]` or `None`):
+            Command of the Job, e.g. `["python", "-c", "print('hello world')"]`
+        arguments (`list[str]` or `None`):
+            Arguments passed to the command
+        environment (`dict[str]` or `None`):
+            Environment variables of the Job as a dictionary.
+        secrets (`dict[str]` or `None`):
+            Secret environment variables of the Job (encrypted).
+        flavor (`str` or `None`):
+            Flavor for the hardware, as in Hugging Face Spaces. See [`SpaceHardware`] for possible values.
+            E.g. `"cpu-basic"`.
+        status: (`JobStatus` or `None`):
+            Status of the Job, e.g. `JobStatus(stage="RUNNING", message=None)`
+            See [`JobStage`] for possible stage values.
+        owner: (`JobOwner` or `None`):
+            Owner of the Job, e.g. `JobOwner(id="5e9ecfc04957053f60648a3e", name="lhoestq", type="user")`
+
+    Example:
+
+    ```python
+    >>> from huggingface_hub import run_job
+    >>> job = run_job(
+    ...     image="python:3.12",
+    ...     command=["python", "-c", "print('Hello from the cloud!')"]
+    ... )
+    >>> job
+    JobInfo(id='687fb701029421ae5549d998', created_at=datetime.datetime(2025, 7, 22, 16, 6, 25, 79000, tzinfo=datetime.timezone.utc), docker_image='python:3.12', space_id=None, command=['python', '-c', "print('Hello from the cloud!')"], arguments=[], environment={}, secrets={}, flavor='cpu-basic', status=JobStatus(stage='RUNNING', message=None), owner=JobOwner(id='5e9ecfc04957053f60648a3e', name='lhoestq', type='user'), endpoint='https://huggingface.co', url='https://huggingface.co/jobs/lhoestq/687fb701029421ae5549d998')
+    >>> job.id
+    '687fb701029421ae5549d998'
+    >>> job.url
+    'https://huggingface.co/jobs/lhoestq/687fb701029421ae5549d998'
+    >>> job.status.stage
+    'RUNNING'
+    ```
+    """
+
+    id: str
+    created_at: Optional[datetime]
+    docker_image: Optional[str]
+    space_id: Optional[str]
+    command: Optional[list[str]]
+    arguments: Optional[list[str]]
+    environment: Optional[dict[str, Any]]
+    secrets: Optional[dict[str, Any]]
+    flavor: Optional[SpaceHardware]
+    status: JobStatus
+    owner: JobOwner
+
+    # Inferred fields
+    endpoint: str
+    url: str
+
+    def __init__(self, **kwargs) -> None:
+        self.id = kwargs["id"]
+        created_at = kwargs.get("createdAt") or kwargs.get("created_at")
+        self.created_at = parse_datetime(created_at) if created_at else None
+        self.docker_image = kwargs.get("dockerImage") or kwargs.get("docker_image")
+        self.space_id = kwargs.get("spaceId") or kwargs.get("space_id")
+        owner = kwargs.get("owner", {})
+        self.owner = JobOwner(id=owner["id"], name=owner["name"], type=owner["type"])
+        self.command = kwargs.get("command")
+        self.arguments = kwargs.get("arguments")
+        self.environment = kwargs.get("environment")
+        self.secrets = kwargs.get("secrets")
+        self.flavor = kwargs.get("flavor")
+        status = kwargs.get("status", {})
+        self.status = JobStatus(stage=status["stage"], message=status.get("message"))
+
+        # Inferred fields
+        self.endpoint = kwargs.get("endpoint", constants.ENDPOINT)
+        self.url = f"{self.endpoint}/jobs/{self.owner.name}/{self.id}"
+
+
+@dataclass
+class JobSpec:
+    docker_image: Optional[str]
+    space_id: Optional[str]
+    command: Optional[list[str]]
+    arguments: Optional[list[str]]
+    environment: Optional[dict[str, Any]]
+    secrets: Optional[dict[str, Any]]
+    flavor: Optional[SpaceHardware]
+    timeout: Optional[int]
+    tags: Optional[list[str]]
+    arch: Optional[str]
+
+    def __init__(self, **kwargs) -> None:
+        self.docker_image = kwargs.get("dockerImage") or kwargs.get("docker_image")
+        self.space_id = kwargs.get("spaceId") or kwargs.get("space_id")
+        self.command = kwargs.get("command")
+        self.arguments = kwargs.get("arguments")
+        self.environment = kwargs.get("environment")
+        self.secrets = kwargs.get("secrets")
+        self.flavor = kwargs.get("flavor")
+        self.timeout = kwargs.get("timeout")
+        self.tags = kwargs.get("tags")
+        self.arch = kwargs.get("arch")
+
+
+@dataclass
+class LastJobInfo:
+    id: str
+    at: datetime
+
+    def __init__(self, **kwargs) -> None:
+        self.id = kwargs["id"]
+        self.at = parse_datetime(kwargs["at"])
+
+
+@dataclass
+class ScheduledJobStatus:
+    last_job: Optional[LastJobInfo]
+    next_job_run_at: Optional[datetime]
+
+    def __init__(self, **kwargs) -> None:
+        last_job = kwargs.get("lastJob") or kwargs.get("last_job")
+        self.last_job = LastJobInfo(**last_job) if last_job else None
+        next_job_run_at = kwargs.get("nextJobRunAt") or kwargs.get("next_job_run_at")
+        self.next_job_run_at = parse_datetime(str(next_job_run_at)) if next_job_run_at else None
+
+
+@dataclass
+class ScheduledJobInfo:
+    """
+    Contains information about a Job.
+
+    Args:
+        id (`str`):
+            Scheduled Job ID.
+        created_at (`datetime` or `None`):
+            When the scheduled Job was created.
+        tags (`list[str]` or `None`):
+            The tags of the scheduled Job.
+        schedule (`str` or `None`):
+            One of "@annually", "@yearly", "@monthly", "@weekly", "@daily", "@hourly", or a
+            CRON schedule expression (e.g., '0 9 * * 1' for 9 AM every Monday).
+        suspend (`bool` or `None`):
+            Whether the scheduled job is suspended (paused).
+        concurrency (`bool` or `None`):
+            Whether multiple instances of this Job can run concurrently.
+        status (`ScheduledJobStatus` or `None`):
+            Status of the scheduled Job.
+        owner: (`JobOwner` or `None`):
+            Owner of the scheduled Job, e.g. `JobOwner(id="5e9ecfc04957053f60648a3e", name="lhoestq", type="user")`
+        job_spec: (`JobSpec` or `None`):
+            Specifications of the Job.
+
+    Example:
+
+    ```python
+    >>> from huggingface_hub import run_job
+    >>> scheduled_job = create_scheduled_job(
+    ...     image="python:3.12",
+    ...     command=["python", "-c", "print('Hello from the cloud!')"],
+    ...     schedule="@hourly",
+    ... )
+    >>> scheduled_job.id
+    '687fb701029421ae5549d999'
+    >>> scheduled_job.status.next_job_run_at
+    datetime.datetime(2025, 7, 22, 17, 6, 25, 79000, tzinfo=datetime.timezone.utc)
+    ```
+    """
+
+    id: str
+    created_at: Optional[datetime]
+    job_spec: JobSpec
+    schedule: Optional[str]
+    suspend: Optional[bool]
+    concurrency: Optional[bool]
+    status: ScheduledJobStatus
+    owner: JobOwner
+
+    def __init__(self, **kwargs) -> None:
+        self.id = kwargs["id"]
+        created_at = kwargs.get("createdAt") or kwargs.get("created_at")
+        self.created_at = parse_datetime(created_at) if created_at else None
+        self.job_spec = JobSpec(**(kwargs.get("job_spec") or kwargs.get("jobSpec", {})))
+        self.schedule = kwargs.get("schedule")
+        self.suspend = kwargs.get("suspend")
+        self.concurrency = kwargs.get("concurrency")
+        status = kwargs.get("status", {})
+        self.status = ScheduledJobStatus(
+            last_job=status.get("last_job") or status.get("lastJob"),
+            next_job_run_at=status.get("next_job_run_at") or status.get("nextJobRunAt"),
+        )
+        owner = kwargs.get("owner", {})
+        self.owner = JobOwner(id=owner["id"], name=owner["name"], type=owner["type"])
+
+
+def _create_job_spec(
+    *,
+    image: str,
+    command: list[str],
+    env: Optional[dict[str, Any]],
+    secrets: Optional[dict[str, Any]],
+    flavor: Optional[SpaceHardware],
+    timeout: Optional[Union[int, float, str]],
+) -> dict[str, Any]:
+    # prepare job spec to send to HF Jobs API
+    job_spec: dict[str, Any] = {
+        "command": command,
+        "arguments": [],
+        "environment": env or {},
+        "flavor": flavor or SpaceHardware.CPU_BASIC,
+    }
+    # secrets are optional
+    if secrets:
+        job_spec["secrets"] = secrets
+    # timeout is optional
+    if timeout:
+        time_units_factors = {"s": 1, "m": 60, "h": 3600, "d": 3600 * 24}
+        if isinstance(timeout, str) and timeout[-1] in time_units_factors:
+            job_spec["timeoutSeconds"] = int(float(timeout[:-1]) * time_units_factors[timeout[-1]])
+        else:
+            job_spec["timeoutSeconds"] = int(timeout)
+    # input is either from docker hub or from HF spaces
+    for prefix in (
+        "https://huggingface.co/spaces/",
+        "https://hf.co/spaces/",
+        "huggingface.co/spaces/",
+        "hf.co/spaces/",
+    ):
+        if image.startswith(prefix):
+            job_spec["spaceId"] = image[len(prefix) :]
+            break
+    else:
+        job_spec["dockerImage"] = image
+    return job_spec

huggingface_hub/_local_folder.py

@@ -86,7 +86,13 @@ class LocalDownloadFilePaths:
 
     def incomplete_path(self, etag: str) -> Path:
         """Return the path where a file will be temporarily downloaded before being moved to `file_path`."""
-        return self.metadata_path.parent / f"{_short_hash(self.metadata_path.name)}.{etag}.incomplete"
+        path = self.metadata_path.parent / f"{_short_hash(self.metadata_path.name)}.{etag}.incomplete"
+        resolved_path = str(path.resolve())
+        # Some Windows versions do not allow for paths longer than 255 characters.
+        # In this case, we must specify it as an extended path by using the "\\?\" prefix.
+        if os.name == "nt" and len(resolved_path) > 255 and not resolved_path.startswith("\\\\?\\"):
+            path = Path("\\\\?\\" + resolved_path)
+        return path
 
 
 @dataclass(frozen=True)
@@ -149,6 +155,7 @@ class LocalUploadFileMetadata:
     should_ignore: Optional[bool] = None
     sha256: Optional[str] = None
     upload_mode: Optional[str] = None
+    remote_oid: Optional[str] = None
     is_uploaded: bool = False
     is_committed: bool = False
 
@@ -174,6 +181,10 @@ class LocalUploadFileMetadata:
                     f.write(self.upload_mode)
                 f.write("\n")
 
+                if self.remote_oid is not None:
+                    f.write(self.remote_oid)
+                f.write("\n")
+
                 f.write(str(int(self.is_uploaded)) + "\n")
                 f.write(str(int(self.is_committed)) + "\n")
 
@@ -195,7 +206,7 @@ def get_local_download_paths(local_dir: Path, filename: str) -> LocalDownloadFil
         [`LocalDownloadFilePaths`]: the paths to the files (file_path, lock_path, metadata_path, incomplete_path).
     """
     # filename is the path in the Hub repository (separated by '/')
-    # make sure to have a cross platform transcription
+    # make sure to have a cross-platform transcription
     sanitized_filename = os.path.join(*filename.split("/"))
     if os.name == "nt":
         if sanitized_filename.startswith("..\\") or "\\..\\" in sanitized_filename:
@@ -235,7 +246,7 @@ def get_local_upload_paths(local_dir: Path, filename: str) -> LocalUploadFilePat
         [`LocalUploadFilePaths`]: the paths to the files (file_path, lock_path, metadata_path).
     """
     # filename is the path in the Hub repository (separated by '/')
-    # make sure to have a cross platform transcription
+    # make sure to have a cross-platform transcription
     sanitized_filename = os.path.join(*filename.split("/"))
     if os.name == "nt":
         if sanitized_filename.startswith("..\\") or "\\..\\" in sanitized_filename:
@@ -346,6 +357,9 @@ def read_upload_metadata(local_dir: Path, filename: str) -> LocalUploadFileMetad
                     if upload_mode not in (None, "regular", "lfs"):
                         raise ValueError(f"Invalid upload mode in metadata {paths.path_in_repo}: {upload_mode}")
 
+                    _remote_oid = f.readline().strip()
+                    remote_oid = None if _remote_oid == "" else _remote_oid
+
                     is_uploaded = bool(int(f.readline().strip()))
                     is_committed = bool(int(f.readline().strip()))
 
@@ -355,6 +369,7 @@ def read_upload_metadata(local_dir: Path, filename: str) -> LocalUploadFileMetad
                         should_ignore=should_ignore,
                         sha256=sha256,
                         upload_mode=upload_mode,
+                        remote_oid=remote_oid,
                         is_uploaded=is_uploaded,
                         is_committed=is_committed,
                     )