huggingface-hub 0.36.0__py3-none-any.whl → 1.0.0__py3-none-any.whl

huggingface_hub/_inference_endpoints.py

@@ -2,7 +2,7 @@ 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
 
@@ -62,7 +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
@@ -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:
@@ -260,8 +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.
 
@@ -293,10 +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.

huggingface_hub/_jobs_api.py

@@ -15,7 +15,7 @@
 from dataclasses import dataclass
 from datetime import datetime
 from enum import Enum
-from typing import Any, Dict, List, Optional, Union
+from typing import Any, Optional, Union
 
 from huggingface_hub import constants
 from huggingface_hub._space_api import SpaceHardware
@@ -71,13 +71,13 @@ class JobInfo:
         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 (`list[str]` or `None`):
             Command of the Job, e.g. `["python", "-c", "print('hello world')"]`
-        arguments (`List[str]` or `None`):
+        arguments (`list[str]` or `None`):
             Arguments passed to the command
-        environment (`Dict[str]` or `None`):
+        environment (`dict[str]` or `None`):
             Environment variables of the Job as a dictionary.
-        secrets (`Dict[str]` or `None`):
+        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.
@@ -111,10 +111,10 @@ class JobInfo:
     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]]
+    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
@@ -148,13 +148,13 @@ class JobInfo:
 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]]
+    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]]
+    tags: Optional[list[str]]
     arch: Optional[str]
 
     def __init__(self, **kwargs) -> None:
@@ -202,7 +202,7 @@ class ScheduledJobInfo:
             Scheduled Job ID.
         created_at (`datetime` or `None`):
             When the scheduled Job was created.
-        tags (`List[str]` or `None`):
+        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
@@ -263,14 +263,14 @@ class ScheduledJobInfo:
 def _create_job_spec(
     *,
     image: str,
-    command: List[str],
-    env: Optional[Dict[str, Any]],
-    secrets: Optional[Dict[str, Any]],
+    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]:
+) -> dict[str, Any]:
     # prepare job spec to send to HF Jobs API
-    job_spec: Dict[str, Any] = {
+    job_spec: dict[str, Any] = {
         "command": command,
         "arguments": [],
         "environment": env or {},

huggingface_hub/_login.py

@@ -19,9 +19,11 @@ from getpass import getpass
 from pathlib import Path
 from typing import Optional
 
+import typer
+
 from . import constants
-from .commands._cli_utils import ANSI
 from .utils import (
+    ANSI,
     capture_output,
     get_token,
     is_google_colab,
@@ -41,7 +43,6 @@ from .utils._auth import (
     _save_token,
     get_stored_tokens,
 )
-from .utils._deprecation import _deprecate_arguments, _deprecate_positional_args
 
 
 logger = logging.get_logger(__name__)
@@ -55,18 +56,11 @@ _HF_LOGO_ASCII = """
 """
 
 
-@_deprecate_arguments(
-    version="1.0",
-    deprecated_args="write_permission",
-    custom_message="Fine-grained tokens added complexity to the permissions, making it irrelevant to check if a token has 'write' access.",
-)
-@_deprecate_positional_args(version="1.0")
 def login(
     token: Optional[str] = None,
     *,
     add_to_git_credential: bool = False,
-    new_session: bool = True,
-    write_permission: bool = False,
+    skip_if_logged_in: bool = False,
 ) -> None:
     """Login the machine to access the Hub.
 
@@ -96,10 +90,8 @@ def login(
             is configured, a warning will be displayed to the user. If `token` is `None`,
             the value of `add_to_git_credential` is ignored and will be prompted again
             to the end user.
-        new_session (`bool`, defaults to `True`):
-            If `True`, will request a token even if one is already saved on the machine.
-        write_permission (`bool`):
-            Ignored and deprecated argument.
+        skip_if_logged_in (`bool`, defaults to `False`):
+            If `True`, do not prompt for token if user is already logged in.
     Raises:
         [`ValueError`](https://docs.python.org/3/library/exceptions.html#ValueError)
             If an organization token is passed. Only personal account tokens are valid
@@ -119,9 +111,9 @@ def login(
             )
         _login(token, add_to_git_credential=add_to_git_credential)
     elif is_notebook():
-        notebook_login(new_session=new_session)
+        notebook_login(skip_if_logged_in=skip_if_logged_in)
     else:
-        interpreter_login(new_session=new_session)
+        interpreter_login(skip_if_logged_in=skip_if_logged_in)
 
 
 def logout(token_name: Optional[str] = None) -> None:
@@ -236,13 +228,7 @@ def auth_list() -> None:
 ###
 
 
-@_deprecate_arguments(
-    version="1.0",
-    deprecated_args="write_permission",
-    custom_message="Fine-grained tokens added complexity to the permissions, making it irrelevant to check if a token has 'write' access.",
-)
-@_deprecate_positional_args(version="1.0")
-def interpreter_login(*, new_session: bool = True, write_permission: bool = False) -> None:
+def interpreter_login(*, skip_if_logged_in: bool = False) -> None:
     """
     Displays a prompt to log in to the HF website and store the token.
 
@@ -253,17 +239,13 @@ def interpreter_login(*, new_session: bool = True, write_permission: bool = Fals
     For more details, see [`login`].
 
     Args:
-        new_session (`bool`, defaults to `True`):
-            If `True`, will request a token even if one is already saved on the machine.
-        write_permission (`bool`):
-            Ignored and deprecated argument.
+        skip_if_logged_in (`bool`, defaults to `False`):
+            If `True`, do not prompt for token if user is already logged in.
     """
-    if not new_session and get_token() is not None:
+    if not skip_if_logged_in and get_token() is not None:
         logger.info("User is already logged in.")
         return
 
-    from .commands.delete_cache import _ask_for_confirmation_no_tui
-
     print(_HF_LOGO_ASCII)
     if get_token() is not None:
         logger.info(
@@ -279,7 +261,7 @@ def interpreter_login(*, new_session: bool = True, write_permission: bool = Fals
     if os.name == "nt":
         logger.info("Token can be pasted using 'Right-Click'.")
     token = getpass("Enter your token (input will not be visible): ")
-    add_to_git_credential = _ask_for_confirmation_no_tui("Add token as git credential?")
+    add_to_git_credential = typer.confirm("Add token as git credential?")
 
     _login(token=token, add_to_git_credential=add_to_git_credential)
 
@@ -308,13 +290,7 @@ NOTEBOOK_LOGIN_TOKEN_HTML_END = """
 notebooks. </center>"""
 
 
-@_deprecate_arguments(
-    version="1.0",
-    deprecated_args="write_permission",
-    custom_message="Fine-grained tokens added complexity to the permissions, making it irrelevant to check if a token has 'write' access.",
-)
-@_deprecate_positional_args(version="1.0")
-def notebook_login(*, new_session: bool = True, write_permission: bool = False) -> None:
+def notebook_login(*, skip_if_logged_in: bool = False) -> None:
     """
     Displays a widget to log in to the HF website and store the token.
 
@@ -325,10 +301,8 @@ def notebook_login(*, new_session: bool = True, write_permission: bool = False)
     For more details, see [`login`].
 
     Args:
-        new_session (`bool`, defaults to `True`):
-            If `True`, will request a token even if one is already saved on the machine.
-        write_permission (`bool`):
-            Ignored and deprecated argument.
+        skip_if_logged_in (`bool`, defaults to `False`):
+            If `True`, do not prompt for token if user is already logged in.
     """
     try:
         import ipywidgets.widgets as widgets  # type: ignore
@@ -338,7 +312,7 @@ def notebook_login(*, new_session: bool = True, write_permission: bool = False)
             "The `notebook_login` function can only be used in a notebook (Jupyter or"
             " Colab) and you need the `ipywidgets` module: `pip install ipywidgets`."
         )
-    if not new_session and get_token() is not None:
+    if not skip_if_logged_in and get_token() is not None:
         logger.info("User is already logged in.")
         return
 

huggingface_hub/_oauth.py

@@ -6,7 +6,7 @@ import time
 import urllib.parse
 import warnings
 from dataclasses import dataclass
-from typing import TYPE_CHECKING, Dict, List, Literal, Optional, Tuple, Union
+from typing import TYPE_CHECKING, Literal, Optional, Union
 
 from . import constants
 from .hf_api import whoami
@@ -39,7 +39,7 @@ class OAuthOrgInfo:
             Whether the org has a payment method set up. Hugging Face field.
         role_in_org (`Optional[str]`, *optional*):
             The user's role in the org. Hugging Face field.
-        security_restrictions (`Optional[List[Literal["ip", "token-policy", "mfa", "sso"]]]`, *optional*):
+        security_restrictions (`Optional[list[Literal["ip", "token-policy", "mfa", "sso"]]]`, *optional*):
             Array of security restrictions that the user hasn't completed for this org. Possible values: "ip", "token-policy", "mfa", "sso". Hugging Face field.
     """
 
@@ -50,7 +50,7 @@ class OAuthOrgInfo:
     is_enterprise: bool
     can_pay: Optional[bool] = None
     role_in_org: Optional[str] = None
-    security_restrictions: Optional[List[Literal["ip", "token-policy", "mfa", "sso"]]] = None
+    security_restrictions: Optional[list[Literal["ip", "token-policy", "mfa", "sso"]]] = None
 
 
 @dataclass
@@ -79,7 +79,7 @@ class OAuthUserInfo:
             Whether the user is a pro user. Hugging Face field.
         can_pay (`Optional[bool]`, *optional*):
             Whether the user has a payment method set up. Hugging Face field.
-        orgs (`Optional[List[OrgInfo]]`, *optional*):
+        orgs (`Optional[list[OrgInfo]]`, *optional*):
             List of organizations the user is part of. Hugging Face field.
     """
 
@@ -93,7 +93,7 @@ class OAuthUserInfo:
     website: Optional[str]
     is_pro: bool
     can_pay: Optional[bool]
-    orgs: Optional[List[OAuthOrgInfo]]
+    orgs: Optional[list[OAuthOrgInfo]]
 
 
 @dataclass
@@ -306,7 +306,7 @@ def _add_oauth_routes(app: "fastapi.FastAPI", route_prefix: str) -> None:
             target_url = request.query_params.get("_target_url")
 
             # Build redirect URI with the same query params as before and bump nb_redirects count
-            query_params: Dict[str, Union[int, str]] = {"_nb_redirects": nb_redirects + 1}
+            query_params: dict[str, Union[int, str]] = {"_nb_redirects": nb_redirects + 1}
             if target_url:
                 query_params["_target_url"] = target_url
 
@@ -406,7 +406,7 @@ def _get_redirect_target(request: "fastapi.Request", default_target: str = "/")
     return request.query_params.get("_target_url", default_target)
 
 
-def _get_mocked_oauth_info() -> Dict:
+def _get_mocked_oauth_info() -> dict:
     token = get_token()
     if token is None:
         raise ValueError(
@@ -449,7 +449,7 @@ def _get_mocked_oauth_info() -> Dict:
     }
 
 
-def _get_oauth_uris(route_prefix: str = "/") -> Tuple[str, str, str]:
+def _get_oauth_uris(route_prefix: str = "/") -> tuple[str, str, str]:
     route_prefix = route_prefix.strip("/")
     if route_prefix:
         route_prefix = f"/{route_prefix}"

huggingface_hub/_snapshot_download.py

@@ -1,20 +1,21 @@
 import os
 from pathlib import Path
-from typing import Dict, Iterable, List, Literal, Optional, Type, Union
+from typing import Iterable, List, Literal, Optional, Union, overload
 
-import requests
+import httpx
 from tqdm.auto import tqdm as base_tqdm
 from tqdm.contrib.concurrent import thread_map
 
 from . import constants
 from .errors import (
+    DryRunError,
     GatedRepoError,
     HfHubHTTPError,
     LocalEntryNotFoundError,
     RepositoryNotFoundError,
     RevisionNotFoundError,
 )
-from .file_download import REGEX_COMMIT_HASH, hf_hub_download, repo_folder_name
+from .file_download import REGEX_COMMIT_HASH, DryRunFileInfo, hf_hub_download, repo_folder_name
 from .hf_api import DatasetInfo, HfApi, ModelInfo, RepoFile, SpaceInfo
 from .utils import OfflineModeIsEnabled, filter_repo_objects, logging, validate_hf_hub_args
 from .utils import tqdm as hf_tqdm
@@ -25,6 +26,81 @@ logger = logging.get_logger(__name__)
 VERY_LARGE_REPO_THRESHOLD = 50000  # After this limit, we don't consider `repo_info.siblings` to be reliable enough
 
 
+@overload
+def snapshot_download(
+    repo_id: str,
+    *,
+    repo_type: Optional[str] = None,
+    revision: Optional[str] = None,
+    cache_dir: Union[str, Path, None] = None,
+    local_dir: Union[str, Path, None] = None,
+    library_name: Optional[str] = None,
+    library_version: Optional[str] = None,
+    user_agent: Optional[Union[dict, str]] = None,
+    etag_timeout: float = constants.DEFAULT_ETAG_TIMEOUT,
+    force_download: bool = False,
+    token: Optional[Union[bool, str]] = None,
+    local_files_only: bool = False,
+    allow_patterns: Optional[Union[list[str], str]] = None,
+    ignore_patterns: Optional[Union[list[str], str]] = None,
+    max_workers: int = 8,
+    tqdm_class: Optional[type[base_tqdm]] = None,
+    headers: Optional[dict[str, str]] = None,
+    endpoint: Optional[str] = None,
+    dry_run: Literal[False] = False,
+) -> str: ...
+
+
+@overload
+def snapshot_download(
+    repo_id: str,
+    *,
+    repo_type: Optional[str] = None,
+    revision: Optional[str] = None,
+    cache_dir: Union[str, Path, None] = None,
+    local_dir: Union[str, Path, None] = None,
+    library_name: Optional[str] = None,
+    library_version: Optional[str] = None,
+    user_agent: Optional[Union[dict, str]] = None,
+    etag_timeout: float = constants.DEFAULT_ETAG_TIMEOUT,
+    force_download: bool = False,
+    token: Optional[Union[bool, str]] = None,
+    local_files_only: bool = False,
+    allow_patterns: Optional[Union[list[str], str]] = None,
+    ignore_patterns: Optional[Union[list[str], str]] = None,
+    max_workers: int = 8,
+    tqdm_class: Optional[type[base_tqdm]] = None,
+    headers: Optional[dict[str, str]] = None,
+    endpoint: Optional[str] = None,
+    dry_run: Literal[True] = True,
+) -> list[DryRunFileInfo]: ...
+
+
+@overload
+def snapshot_download(
+    repo_id: str,
+    *,
+    repo_type: Optional[str] = None,
+    revision: Optional[str] = None,
+    cache_dir: Union[str, Path, None] = None,
+    local_dir: Union[str, Path, None] = None,
+    library_name: Optional[str] = None,
+    library_version: Optional[str] = None,
+    user_agent: Optional[Union[dict, str]] = None,
+    etag_timeout: float = constants.DEFAULT_ETAG_TIMEOUT,
+    force_download: bool = False,
+    token: Optional[Union[bool, str]] = None,
+    local_files_only: bool = False,
+    allow_patterns: Optional[Union[list[str], str]] = None,
+    ignore_patterns: Optional[Union[list[str], str]] = None,
+    max_workers: int = 8,
+    tqdm_class: Optional[type[base_tqdm]] = None,
+    headers: Optional[dict[str, str]] = None,
+    endpoint: Optional[str] = None,
+    dry_run: bool = False,
+) -> Union[str, list[DryRunFileInfo]]: ...
+
+
 @validate_hf_hub_args
 def snapshot_download(
     repo_id: str,
@@ -35,22 +111,19 @@ def snapshot_download(
     local_dir: Union[str, Path, None] = None,
     library_name: Optional[str] = None,
     library_version: Optional[str] = None,
-    user_agent: Optional[Union[Dict, str]] = None,
-    proxies: Optional[Dict] = None,
+    user_agent: Optional[Union[dict, str]] = None,
     etag_timeout: float = constants.DEFAULT_ETAG_TIMEOUT,
     force_download: bool = False,
     token: Optional[Union[bool, str]] = None,
     local_files_only: bool = False,
-    allow_patterns: Optional[Union[List[str], str]] = None,
-    ignore_patterns: Optional[Union[List[str], str]] = None,
+    allow_patterns: Optional[Union[list[str], str]] = None,
+    ignore_patterns: Optional[Union[list[str], str]] = None,
     max_workers: int = 8,
-    tqdm_class: Optional[Type[base_tqdm]] = None,
-    headers: Optional[Dict[str, str]] = None,
+    tqdm_class: Optional[type[base_tqdm]] = None,
+    headers: Optional[dict[str, str]] = None,
     endpoint: Optional[str] = None,
-    # Deprecated args
-    local_dir_use_symlinks: Union[bool, Literal["auto"]] = "auto",
-    resume_download: Optional[bool] = None,
-) -> str:
+    dry_run: bool = False,
+) -> Union[str, list[DryRunFileInfo]]:
     """Download repo files.
 
     Download a whole snapshot of a repo's files at the specified revision. This is useful when you want all files from
@@ -85,12 +158,9 @@ def snapshot_download(
             The version of the library.
         user_agent (`str`, `dict`, *optional*):
             The user-agent info in the form of a dictionary or a string.
-        proxies (`dict`, *optional*):
-            Dictionary mapping protocol to the URL of the proxy passed to
-            `requests.request`.
         etag_timeout (`float`, *optional*, defaults to `10`):
             When fetching ETag, how many seconds to wait for the server to send
-            data before giving up which is passed to `requests.request`.
+            data before giving up which is passed to `httpx.request`.
         force_download (`bool`, *optional*, defaults to `False`):
             Whether the file should be downloaded even if it already exists in the local cache.
         token (`str`, `bool`, *optional*):
@@ -103,9 +173,9 @@ def snapshot_download(
         local_files_only (`bool`, *optional*, defaults to `False`):
             If `True`, avoid downloading the file and return the path to the
             local cached file if it exists.
-        allow_patterns (`List[str]` or `str`, *optional*):
+        allow_patterns (`list[str]` or `str`, *optional*):
             If provided, only files matching at least one pattern are downloaded.
-        ignore_patterns (`List[str]` or `str`, *optional*):
+        ignore_patterns (`list[str]` or `str`, *optional*):
             If provided, files matching any of the patterns are not downloaded.
         max_workers (`int`, *optional*):
             Number of concurrent threads to download files (1 thread = 1 file download).
@@ -116,9 +186,14 @@ def snapshot_download(
             Note that the `tqdm_class` is not passed to each individual download.
             Defaults to the custom HF progress bar that can be disabled by setting
             `HF_HUB_DISABLE_PROGRESS_BARS` environment variable.
+        dry_run (`bool`, *optional*, defaults to `False`):
+            If `True`, perform a dry run without actually downloading the files. Returns a list of
+            [`DryRunFileInfo`] objects containing information about what would be downloaded.
 
     Returns:
-        `str`: folder path of the repo snapshot.
+        `str` or list of [`DryRunFileInfo`]:
+            - If `dry_run=False`: Local snapshot path.
+            - If `dry_run=True`: A list of [`DryRunFileInfo`] objects containing download information.
 
     Raises:
         [`~utils.RepositoryNotFoundError`]
@@ -163,14 +238,10 @@ def snapshot_download(
         try:
             # if we have internet connection we want to list files to download
             repo_info = api.repo_info(repo_id=repo_id, repo_type=repo_type, revision=revision)
-        except (requests.exceptions.SSLError, requests.exceptions.ProxyError):
-            # Actually raise for those subclasses of ConnectionError
+        except httpx.ProxyError:
+            # Actually raise on proxy error
             raise
-        except (
-            requests.exceptions.ConnectionError,
-            requests.exceptions.Timeout,
-            OfflineModeIsEnabled,
-        ) as error:
+        except (httpx.ConnectError, httpx.TimeoutException, OfflineModeIsEnabled) as error:
             # Internet connection is down
             # => will try to use local files only
             api_call_error = error
@@ -178,7 +249,7 @@ def snapshot_download(
         except RevisionNotFoundError:
             # The repo was found but the revision doesn't exist on the Hub (never existed or got deleted)
             raise
-        except requests.HTTPError as error:
+        except HfHubHTTPError as error:
             # Multiple reasons for an http error:
             # - Repository is private and invalid/missing token sent
             # - Repository is gated and invalid/missing token sent
@@ -198,6 +269,11 @@ def snapshot_download(
     #    - f the specified revision is a branch or tag, look inside "refs".
     # => if local_dir is not None, we will return the path to the local folder if it exists.
     if repo_info is None:
+        if dry_run:
+            raise DryRunError(
+                "Dry run cannot be performed as the repository cannot be accessed. Please check your internet connection or authentication token."
+            ) from api_call_error
+
         # Try to get which commit hash corresponds to the specified revision
         commit_hash = None
         if REGEX_COMMIT_HASH.match(revision):
@@ -284,6 +360,8 @@ def snapshot_download(
         tqdm_desc = f"Fetching {len(filtered_repo_files)} files"
     else:
         tqdm_desc = "Fetching ... files"
+    if dry_run:
+        tqdm_desc = "[dry-run] " + tqdm_desc
 
     commit_hash = repo_info.sha
     snapshot_folder = os.path.join(storage_folder, "snapshots", commit_hash)
@@ -299,33 +377,36 @@ def snapshot_download(
         except OSError as e:
             logger.warning(f"Ignored error while writing commit hash to {ref_path}: {e}.")
 
+    results: List[Union[str, DryRunFileInfo]] = []
+
     # we pass the commit_hash to hf_hub_download
     # so no network call happens if we already
     # have the file locally.
-    def _inner_hf_hub_download(repo_file: str):
-        return hf_hub_download(
-            repo_id,
-            filename=repo_file,
-            repo_type=repo_type,
-            revision=commit_hash,
-            endpoint=endpoint,
-            cache_dir=cache_dir,
-            local_dir=local_dir,
-            local_dir_use_symlinks=local_dir_use_symlinks,
-            library_name=library_name,
-            library_version=library_version,
-            user_agent=user_agent,
-            proxies=proxies,
-            etag_timeout=etag_timeout,
-            resume_download=resume_download,
-            force_download=force_download,
-            token=token,
-            headers=headers,
+    def _inner_hf_hub_download(repo_file: str) -> None:
+        results.append(
+            hf_hub_download(  # type: ignore[no-matching-overload] # ty not happy, don't know why :/
+                repo_id,
+                filename=repo_file,
+                repo_type=repo_type,
+                revision=commit_hash,
+                endpoint=endpoint,
+                cache_dir=cache_dir,
+                local_dir=local_dir,
+                library_name=library_name,
+                library_version=library_version,
+                user_agent=user_agent,
+                etag_timeout=etag_timeout,
+                force_download=force_download,
+                token=token,
+                headers=headers,
+                dry_run=dry_run,
+            )
         )
 
-    if constants.HF_HUB_ENABLE_HF_TRANSFER:
-        # when using hf_transfer we don't want extra parallelism
-        # from the one hf_transfer provides
+    if constants.HF_XET_HIGH_PERFORMANCE and not dry_run:
+        # when using hf_xet high performance we don't want extra parallelism
+        # from the one hf_xet provides
+        # TODO: revisit this when xet_session is implemented
         for file in filtered_repo_files:
             _inner_hf_hub_download(file)
     else:
@@ -338,6 +419,10 @@ def snapshot_download(
             tqdm_class=tqdm_class or hf_tqdm,
         )
 
+    if dry_run:
+        assert all(isinstance(r, DryRunFileInfo) for r in results)
+        return results  # type: ignore
+
     if local_dir is not None:
         return str(os.path.realpath(local_dir))
     return snapshot_folder