huggingface-hub 0.31.0rc0__py3-none-any.whl → 1.1.3__py3-none-any.whl

huggingface_hub/cli/upload.py

@@ -0,0 +1,294 @@
+# coding=utf-8
+# Copyright 2023-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.
+"""Contains command to upload a repo or file with the CLI.
+
+Usage:
+    # Upload file (implicit)
+    hf upload my-cool-model ./my-cool-model.safetensors
+
+    # Upload file (explicit)
+    hf upload my-cool-model ./my-cool-model.safetensors  model.safetensors
+
+    # Upload directory (implicit). If `my-cool-model/` is a directory it will be uploaded, otherwise an exception is raised.
+    hf upload my-cool-model
+
+    # Upload directory (explicit)
+    hf upload my-cool-model ./models/my-cool-model .
+
+    # Upload filtered directory (example: tensorboard logs except for the last run)
+    hf upload my-cool-model ./model/training /logs --include "*.tfevents.*" --exclude "*20230905*"
+
+    # Upload with wildcard
+    hf upload my-cool-model "./model/training/*.safetensors"
+
+    # Upload private dataset
+    hf upload Wauplin/my-cool-dataset ./data . --repo-type=dataset --private
+
+    # Upload with token
+    hf upload Wauplin/my-cool-model --token=hf_****
+
+    # Sync local Space with Hub (upload new files, delete removed files)
+    hf upload Wauplin/space-example --repo-type=space --exclude="/logs/*" --delete="*" --commit-message="Sync local Space with Hub"
+
+    # Schedule commits every 30 minutes
+    hf upload Wauplin/my-cool-model --every=30
+"""
+
+import os
+import time
+import warnings
+from typing import Annotated, Optional
+
+import typer
+
+from huggingface_hub import logging
+from huggingface_hub._commit_scheduler import CommitScheduler
+from huggingface_hub.errors import RevisionNotFoundError
+from huggingface_hub.utils import disable_progress_bars, enable_progress_bars
+
+from ._cli_utils import PrivateOpt, RepoIdArg, RepoType, RepoTypeOpt, RevisionOpt, TokenOpt, get_hf_api
+
+
+logger = logging.get_logger(__name__)
+
+
+def upload(
+    repo_id: RepoIdArg,
+    local_path: Annotated[
+        Optional[str],
+        typer.Argument(
+            help="Local path to the file or folder to upload. Wildcard patterns are supported. Defaults to current directory.",
+        ),
+    ] = None,
+    path_in_repo: Annotated[
+        Optional[str],
+        typer.Argument(
+            help="Path of the file or folder in the repo. Defaults to the relative path of the file or folder.",
+        ),
+    ] = None,
+    repo_type: RepoTypeOpt = RepoType.model,
+    revision: RevisionOpt = None,
+    private: PrivateOpt = False,
+    include: Annotated[
+        Optional[list[str]],
+        typer.Option(
+            help="Glob patterns to match files to upload.",
+        ),
+    ] = None,
+    exclude: Annotated[
+        Optional[list[str]],
+        typer.Option(
+            help="Glob patterns to exclude from files to upload.",
+        ),
+    ] = None,
+    delete: Annotated[
+        Optional[list[str]],
+        typer.Option(
+            help="Glob patterns for file to be deleted from the repo while committing.",
+        ),
+    ] = None,
+    commit_message: Annotated[
+        Optional[str],
+        typer.Option(
+            help="The summary / title / first line of the generated commit.",
+        ),
+    ] = None,
+    commit_description: Annotated[
+        Optional[str],
+        typer.Option(
+            help="The description of the generated commit.",
+        ),
+    ] = None,
+    create_pr: Annotated[
+        bool,
+        typer.Option(
+            help="Whether to upload content as a new Pull Request.",
+        ),
+    ] = False,
+    every: Annotated[
+        Optional[float],
+        typer.Option(
+            help="f set, a background job is scheduled to create commits every `every` minutes.",
+        ),
+    ] = None,
+    token: TokenOpt = None,
+    quiet: Annotated[
+        bool,
+        typer.Option(
+            help="Disable progress bars and warnings; print only the returned path.",
+        ),
+    ] = False,
+) -> None:
+    """Upload a file or a folder to the Hub. Recommended for single-commit uploads."""
+
+    if every is not None and every <= 0:
+        raise typer.BadParameter("--every must be a positive value", param_hint="every")
+
+    repo_type_str = repo_type.value
+
+    api = get_hf_api(token=token)
+
+    # Resolve local_path and path_in_repo based on implicit/explicit rules
+    resolved_local_path, resolved_path_in_repo, resolved_include = _resolve_upload_paths(
+        repo_id=repo_id, local_path=local_path, path_in_repo=path_in_repo, include=include
+    )
+
+    def run_upload() -> str:
+        if os.path.isfile(resolved_local_path):
+            if resolved_include is not None and len(resolved_include) > 0 and isinstance(resolved_include, list):
+                warnings.warn("Ignoring --include since a single file is uploaded.")
+            if exclude is not None and len(exclude) > 0:
+                warnings.warn("Ignoring --exclude since a single file is uploaded.")
+            if delete is not None and len(delete) > 0:
+                warnings.warn("Ignoring --delete since a single file is uploaded.")
+
+        # Schedule commits if `every` is set
+        if every is not None:
+            if os.path.isfile(resolved_local_path):
+                # If file => watch entire folder + use allow_patterns
+                folder_path = os.path.dirname(resolved_local_path)
+                pi = (
+                    resolved_path_in_repo[: -len(resolved_local_path)]
+                    if resolved_path_in_repo.endswith(resolved_local_path)
+                    else resolved_path_in_repo
+                )
+                allow_patterns = [resolved_local_path]
+                ignore_patterns: Optional[list[str]] = []
+            else:
+                folder_path = resolved_local_path
+                pi = resolved_path_in_repo
+                allow_patterns = (
+                    resolved_include or []
+                    if isinstance(resolved_include, list)
+                    else [resolved_include]
+                    if isinstance(resolved_include, str)
+                    else []
+                )
+                ignore_patterns = exclude or []
+                if delete is not None and len(delete) > 0:
+                    warnings.warn("Ignoring --delete when uploading with scheduled commits.")
+
+            scheduler = CommitScheduler(
+                folder_path=folder_path,
+                repo_id=repo_id,
+                repo_type=repo_type_str,
+                revision=revision,
+                allow_patterns=allow_patterns,
+                ignore_patterns=ignore_patterns,
+                path_in_repo=pi,
+                private=private,
+                every=every,
+                hf_api=api,
+            )
+            print(f"Scheduling commits every {every} minutes to {scheduler.repo_id}.")
+            try:
+                while True:
+                    time.sleep(100)
+            except KeyboardInterrupt:
+                scheduler.stop()
+                return "Stopped scheduled commits."
+
+        # Otherwise, create repo and proceed with the upload
+        if not os.path.isfile(resolved_local_path) and not os.path.isdir(resolved_local_path):
+            raise FileNotFoundError(f"No such file or directory: '{resolved_local_path}'.")
+        created = api.create_repo(
+            repo_id=repo_id,
+            repo_type=repo_type_str,
+            exist_ok=True,
+            private=private,
+            space_sdk="gradio" if repo_type_str == "space" else None,
+            # ^ We don't want it to fail when uploading to a Space => let's set Gradio by default.
+            # ^ I'd rather not add CLI args to set it explicitly as we already have `hf repo create` for that.
+        ).repo_id
+
+        # Check if branch already exists and if not, create it
+        if revision is not None and not create_pr:
+            try:
+                api.repo_info(repo_id=created, repo_type=repo_type_str, revision=revision)
+            except RevisionNotFoundError:
+                logger.info(f"Branch '{revision}' not found. Creating it...")
+                api.create_branch(repo_id=created, repo_type=repo_type_str, branch=revision, exist_ok=True)
+                # ^ `exist_ok=True` to avoid race concurrency issues
+
+        # File-based upload
+        if os.path.isfile(resolved_local_path):
+            return api.upload_file(
+                path_or_fileobj=resolved_local_path,
+                path_in_repo=resolved_path_in_repo,
+                repo_id=created,
+                repo_type=repo_type_str,
+                revision=revision,
+                commit_message=commit_message,
+                commit_description=commit_description,
+                create_pr=create_pr,
+            )
+
+        # Folder-based upload
+        return api.upload_folder(
+            folder_path=resolved_local_path,
+            path_in_repo=resolved_path_in_repo,
+            repo_id=created,
+            repo_type=repo_type_str,
+            revision=revision,
+            commit_message=commit_message,
+            commit_description=commit_description,
+            create_pr=create_pr,
+            allow_patterns=(
+                resolved_include
+                if isinstance(resolved_include, list)
+                else [resolved_include]
+                if isinstance(resolved_include, str)
+                else None
+            ),
+            ignore_patterns=exclude,
+            delete_patterns=delete,
+        )
+
+    if quiet:
+        disable_progress_bars()
+        with warnings.catch_warnings():
+            warnings.simplefilter("ignore")
+            print(run_upload())
+        enable_progress_bars()
+    else:
+        print(run_upload())
+        logging.set_verbosity_warning()
+
+
+def _resolve_upload_paths(
+    *, repo_id: str, local_path: Optional[str], path_in_repo: Optional[str], include: Optional[list[str]]
+) -> tuple[str, str, Optional[list[str]]]:
+    repo_name = repo_id.split("/")[-1]
+    resolved_include = include
+
+    if local_path is not None and any(c in local_path for c in ["*", "?", "["]):
+        if include is not None:
+            raise ValueError("Cannot set --include when local_path contains a wildcard.")
+        if path_in_repo is not None and path_in_repo != ".":
+            raise ValueError("Cannot set path_in_repo when local_path contains a wildcard.")
+        return ".", local_path, ["."]  # will be adjusted below; placeholder for type
+
+    if local_path is None and os.path.isfile(repo_name):
+        return repo_name, repo_name, resolved_include
+    if local_path is None and os.path.isdir(repo_name):
+        return repo_name, ".", resolved_include
+    if local_path is None:
+        raise ValueError(f"'{repo_name}' is not a local file or folder. Please set local_path explicitly.")
+
+    if path_in_repo is None and os.path.isfile(local_path):
+        return local_path, os.path.basename(local_path), resolved_include
+    if path_in_repo is None:
+        return local_path, ".", resolved_include
+    return local_path, path_in_repo, resolved_include

huggingface_hub/cli/upload_large_folder.py

@@ -0,0 +1,117 @@
+# coding=utf-8
+# Copyright 2023-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.
+"""Contains command to upload a large folder with the CLI."""
+
+import os
+from typing import Annotated, Optional
+
+import typer
+
+from huggingface_hub import logging
+from huggingface_hub.utils import ANSI, disable_progress_bars
+
+from ._cli_utils import PrivateOpt, RepoIdArg, RepoType, RepoTypeOpt, RevisionOpt, TokenOpt, get_hf_api
+
+
+logger = logging.get_logger(__name__)
+
+
+def upload_large_folder(
+    repo_id: RepoIdArg,
+    local_path: Annotated[
+        str,
+        typer.Argument(
+            help="Local path to the folder to upload.",
+        ),
+    ],
+    repo_type: RepoTypeOpt = RepoType.model,
+    revision: RevisionOpt = None,
+    private: PrivateOpt = False,
+    include: Annotated[
+        Optional[list[str]],
+        typer.Option(
+            help="Glob patterns to match files to upload.",
+        ),
+    ] = None,
+    exclude: Annotated[
+        Optional[list[str]],
+        typer.Option(
+            help="Glob patterns to exclude from files to upload.",
+        ),
+    ] = None,
+    token: TokenOpt = None,
+    num_workers: Annotated[
+        Optional[int],
+        typer.Option(
+            help="Number of workers to use to hash, upload and commit files.",
+        ),
+    ] = None,
+    no_report: Annotated[
+        bool,
+        typer.Option(
+            help="Whether to disable regular status report.",
+        ),
+    ] = False,
+    no_bars: Annotated[
+        bool,
+        typer.Option(
+            help="Whether to disable progress bars.",
+        ),
+    ] = False,
+) -> None:
+    """Upload a large folder to the Hub. Recommended for resumable uploads."""
+    if not os.path.isdir(local_path):
+        raise typer.BadParameter("Large upload is only supported for folders.", param_hint="local_path")
+
+    print(
+        ANSI.yellow(
+            "You are about to upload a large folder to the Hub using `hf upload-large-folder`. "
+            "This is a new feature so feedback is very welcome!\n"
+            "\n"
+            "A few things to keep in mind:\n"
+            "  - Repository limits still apply: https://huggingface.co/docs/hub/repositories-recommendations\n"
+            "  - Do not start several processes in parallel.\n"
+            "  - You can interrupt and resume the process at any time. "
+            "The script will pick up where it left off except for partially uploaded files that would have to be entirely reuploaded.\n"
+            "  - Do not upload the same folder to several repositories. If you need to do so, you must delete the `./.cache/huggingface/` folder first.\n"
+            "\n"
+            f"Some temporary metadata will be stored under `{local_path}/.cache/huggingface`.\n"
+            "  - You must not modify those files manually.\n"
+            "  - You must not delete the `./.cache/huggingface/` folder while a process is running.\n"
+            "  - You can delete the `./.cache/huggingface/` folder to reinitialize the upload state when process is not running. Files will have to be hashed and preuploaded again, except for already committed files.\n"
+            "\n"
+            "If the process output is too verbose, you can disable the progress bars with `--no-bars`. "
+            "You can also entirely disable the status report with `--no-report`.\n"
+            "\n"
+            "For more details, run `hf upload-large-folder --help` or check the documentation at "
+            "https://huggingface.co/docs/huggingface_hub/guides/upload#upload-a-large-folder."
+        )
+    )
+
+    if no_bars:
+        disable_progress_bars()
+
+    api = get_hf_api(token=token)
+    api.upload_large_folder(
+        repo_id=repo_id,
+        folder_path=local_path,
+        repo_type=repo_type.value,
+        revision=revision,
+        private=private,
+        allow_patterns=include,
+        ignore_patterns=exclude,
+        num_workers=num_workers,
+        print_report=not no_report,
+    )

huggingface_hub/community.py

@@ -7,7 +7,7 @@ for more information on Pull Requests, Discussions, and the community tab.
 
 from dataclasses import dataclass
 from datetime import datetime
-from typing import List, Literal, Optional, Union
+from typing import Literal, Optional, TypedDict, Union
 
 from . import constants
 from .utils import parse_datetime
@@ -116,7 +116,7 @@ class DiscussionWithDetails(Discussion):
             The `datetime` of creation of the Discussion / Pull Request.
         events (`list` of [`DiscussionEvent`])
             The list of [`DiscussionEvents`] in this Discussion or Pull Request.
-        conflicting_files (`Union[List[str], bool, None]`, *optional*):
+        conflicting_files (`Union[list[str], bool, None]`, *optional*):
             A list of conflicting files if this is a Pull Request.
             `None` if `self.is_pull_request` is `False`.
             `True` if there are conflicting files but the list can't be retrieved.
@@ -136,13 +136,21 @@ class DiscussionWithDetails(Discussion):
             (property) URL of the discussion on the Hub.
     """
 
-    events: List["DiscussionEvent"]
-    conflicting_files: Union[List[str], bool, None]
+    events: list["DiscussionEvent"]
+    conflicting_files: Union[list[str], bool, None]
     target_branch: Optional[str]
     merge_commit_oid: Optional[str]
     diff: Optional[str]
 
 
+class DiscussionEventArgs(TypedDict):
+    id: str
+    type: str
+    created_at: datetime
+    author: str
+    _event: dict
+
+
 @dataclass
 class DiscussionEvent:
     """
@@ -222,7 +230,7 @@ class DiscussionComment(DiscussionEvent):
         return self._event["data"]["latest"].get("author", {}).get("name", "deleted")
 
     @property
-    def edit_history(self) -> List[dict]:
+    def edit_history(self) -> list[dict]:
         """The edit history of the comment"""
         return self._event["data"]["history"]
 
@@ -319,13 +327,13 @@ def deserialize_event(event: dict) -> DiscussionEvent:
     event_type: str = event["type"]
     created_at = parse_datetime(event["createdAt"])
 
-    common_args = dict(
-        id=event_id,
-        type=event_type,
-        created_at=created_at,
-        author=event.get("author", {}).get("name", "deleted"),
-        _event=event,
-    )
+    common_args: DiscussionEventArgs = {
+        "id": event_id,
+        "type": event_type,
+        "created_at": created_at,
+        "author": event.get("author", {}).get("name", "deleted"),
+        "_event": event,
+    }
 
     if event_type == "comment":
         return DiscussionComment(

huggingface_hub/constants.py

@@ -1,7 +1,7 @@
 import os
 import re
 import typing
-from typing import Literal, Optional, Tuple
+from typing import Literal, Optional
 
 
 # Possible values for env variables
@@ -35,7 +35,6 @@ DEFAULT_ETAG_TIMEOUT = 10
 DEFAULT_DOWNLOAD_TIMEOUT = 10
 DEFAULT_REQUEST_TIMEOUT = 10
 DOWNLOAD_CHUNK_SIZE = 10 * 1024 * 1024
-HF_TRANSFER_CONCURRENCY = 100
 MAX_HTTP_DOWNLOAD_SIZE = 50 * 1000 * 1000 * 1000  # 50 GB
 
 # Constants for serialization
@@ -82,6 +81,17 @@ INFERENCE_ENDPOINT = os.environ.get("HF_INFERENCE_ENDPOINT", "https://api-infere
 INFERENCE_ENDPOINTS_ENDPOINT = "https://api.endpoints.huggingface.cloud/v2"
 INFERENCE_CATALOG_ENDPOINT = "https://endpoints.huggingface.co/api/catalog"
 
+# See https://api.endpoints.huggingface.cloud/#post-/v2/endpoint/-namespace-
+INFERENCE_ENDPOINT_IMAGE_KEYS = [
+    "custom",
+    "huggingface",
+    "huggingfaceNeuron",
+    "llamacpp",
+    "tei",
+    "tgi",
+    "tgiNeuron",
+]
+
 # Proxy for third-party providers
 INFERENCE_PROXY_TEMPLATE = "https://router.huggingface.co/{provider}"
 
@@ -107,9 +117,9 @@ REPO_TYPES_MAPPING = {
 }
 
 DiscussionTypeFilter = Literal["all", "discussion", "pull_request"]
-DISCUSSION_TYPES: Tuple[DiscussionTypeFilter, ...] = typing.get_args(DiscussionTypeFilter)
+DISCUSSION_TYPES: tuple[DiscussionTypeFilter, ...] = typing.get_args(DiscussionTypeFilter)
 DiscussionStatusFilter = Literal["all", "open", "closed"]
-DISCUSSION_STATUS: Tuple[DiscussionTypeFilter, ...] = typing.get_args(DiscussionStatusFilter)
+DISCUSSION_STATUS: tuple[DiscussionTypeFilter, ...] = typing.get_args(DiscussionStatusFilter)
 
 # Webhook subscription types
 WEBHOOK_DOMAIN_T = Literal["repo", "discussions"]
@@ -124,7 +134,6 @@ HF_HOME = os.path.expandvars(
         )
     )
 )
-hf_cache_home = HF_HOME  # for backward compatibility. TODO: remove this in 1.0.0
 
 default_cache_path = os.path.join(HF_HOME, "hub")
 default_assets_cache_path = os.path.join(HF_HOME, "assets")
@@ -153,6 +162,10 @@ HF_ASSETS_CACHE = os.path.expandvars(
 
 HF_HUB_OFFLINE = _is_true(os.environ.get("HF_HUB_OFFLINE") or os.environ.get("TRANSFORMERS_OFFLINE"))
 
+# File created to mark that the version check has been done.
+# Check is performed once per 24 hours at most.
+CHECK_FOR_UPDATE_DONE_PATH = os.path.join(HF_HOME, ".check_for_update_done")
+
 # If set, log level will be set to DEBUG and all requests made to the Hub will be logged
 # as curl commands for reproducibility.
 HF_DEBUG = _is_true(os.environ.get("HF_DEBUG"))
@@ -201,18 +214,18 @@ HF_HUB_DISABLE_EXPERIMENTAL_WARNING: bool = _is_true(os.environ.get("HF_HUB_DISA
 # Disable sending the cached token by default is all HTTP requests to the Hub
 HF_HUB_DISABLE_IMPLICIT_TOKEN: bool = _is_true(os.environ.get("HF_HUB_DISABLE_IMPLICIT_TOKEN"))
 
-# Enable fast-download using external dependency "hf_transfer"
-# See:
-# - https://pypi.org/project/hf-transfer/
-# - https://github.com/huggingface/hf_transfer (private)
-HF_HUB_ENABLE_HF_TRANSFER: bool = _is_true(os.environ.get("HF_HUB_ENABLE_HF_TRANSFER"))
+HF_XET_HIGH_PERFORMANCE: bool = _is_true(os.environ.get("HF_XET_HIGH_PERFORMANCE"))
 
+# hf_transfer is not used anymore. Let's warn user is case they set the env variable
+if _is_true(os.environ.get("HF_HUB_ENABLE_HF_TRANSFER")) and not HF_XET_HIGH_PERFORMANCE:
+    import warnings
 
-# UNUSED
-# We don't use symlinks in local dir anymore.
-HF_HUB_LOCAL_DIR_AUTO_SYMLINK_THRESHOLD: int = (
-    _as_int(os.environ.get("HF_HUB_LOCAL_DIR_AUTO_SYMLINK_THRESHOLD")) or 5 * 1024 * 1024
-)
+    warnings.warn(
+        "The `HF_HUB_ENABLE_HF_TRANSFER` environment variable is deprecated as 'hf_transfer' is not used anymore. "
+        "Please use `HF_XET_HIGH_PERFORMANCE` instead to enable high performance transfer with Xet. "
+        "Visit https://huggingface.co/docs/huggingface_hub/package_reference/environment_variables#hfxethighperformance for more details.",
+        DeprecationWarning,
+    )
 
 # Used to override the etag timeout on a system level
 HF_HUB_ETAG_TIMEOUT: int = _as_int(os.environ.get("HF_HUB_ETAG_TIMEOUT")) or DEFAULT_ETAG_TIMEOUT
@@ -220,49 +233,20 @@ HF_HUB_ETAG_TIMEOUT: int = _as_int(os.environ.get("HF_HUB_ETAG_TIMEOUT")) or DEF
 # Used to override the get request timeout on a system level
 HF_HUB_DOWNLOAD_TIMEOUT: int = _as_int(os.environ.get("HF_HUB_DOWNLOAD_TIMEOUT")) or DEFAULT_DOWNLOAD_TIMEOUT
 
-# Allows to add information about the requester in the user-agent (eg. partner name)
+# Allows to add information about the requester in the user-agent (e.g. partner name)
 HF_HUB_USER_AGENT_ORIGIN: Optional[str] = os.environ.get("HF_HUB_USER_AGENT_ORIGIN")
 
-# List frameworks that are handled by the InferenceAPI service. Useful to scan endpoints and check which models are
-# deployed and running. Since 95% of the models are using the top 4 frameworks listed below, we scan only those by
-# default. We still keep the full list of supported frameworks in case we want to scan all of them.
-MAIN_INFERENCE_API_FRAMEWORKS = [
-    "diffusers",
-    "sentence-transformers",
-    "text-generation-inference",
-    "transformers",
-]
+# If OAuth didn't work after 2 redirects, there's likely a third-party cookie issue in the Space iframe view.
+# In this case, we redirect the user to the non-iframe view.
+OAUTH_MAX_REDIRECTS = 2
 
-ALL_INFERENCE_API_FRAMEWORKS = MAIN_INFERENCE_API_FRAMEWORKS + [
-    "adapter-transformers",
-    "allennlp",
-    "asteroid",
-    "bertopic",
-    "doctr",
-    "espnet",
-    "fairseq",
-    "fastai",
-    "fasttext",
-    "flair",
-    "k2",
-    "keras",
-    "mindspore",
-    "nemo",
-    "open_clip",
-    "paddlenlp",
-    "peft",
-    "pyannote-audio",
-    "sklearn",
-    "spacy",
-    "span-marker",
-    "speechbrain",
-    "stanza",
-    "timm",
-]
+# OAuth-related environment variables injected by the Space
+OAUTH_CLIENT_ID = os.environ.get("OAUTH_CLIENT_ID")
+OAUTH_CLIENT_SECRET = os.environ.get("OAUTH_CLIENT_SECRET")
+OAUTH_SCOPES = os.environ.get("OAUTH_SCOPES")
+OPENID_PROVIDER_URL = os.environ.get("OPENID_PROVIDER_URL")
 
 # Xet constants
-
-
 HUGGINGFACE_HEADER_X_XET_ENDPOINT = "X-Xet-Cas-Url"
 HUGGINGFACE_HEADER_X_XET_ACCESS_TOKEN = "X-Xet-Access-Token"
 HUGGINGFACE_HEADER_X_XET_EXPIRATION = "X-Xet-Token-Expiration"
@@ -272,3 +256,4 @@ HUGGINGFACE_HEADER_LINK_XET_AUTH_KEY = "xet-auth"
 
 default_xet_cache_path = os.path.join(HF_HOME, "xet")
 HF_XET_CACHE = os.getenv("HF_XET_CACHE", default_xet_cache_path)
+HF_HUB_DISABLE_XET: bool = _is_true(os.environ.get("HF_HUB_DISABLE_XET"))