huggingface-hub 0.32.1__py3-none-any.whl → 0.32.2__py3-none-any.whl

huggingface_hub/__init__.py

@@ -46,7 +46,7 @@ import sys
 from typing import TYPE_CHECKING
 
 
-__version__ = "0.32.1"
+__version__ = "0.32.2"
 
 # Alphabetical order of definitions is ensured in tests
 # WARNING: any comment added in this dictionary definition will be lost when

huggingface_hub/_snapshot_download.py

@@ -1,6 +1,6 @@
 import os
 from pathlib import Path
-from typing import Dict, List, Literal, Optional, Union
+from typing import Dict, Iterable, List, Literal, Optional, Union
 
 import requests
 from tqdm.auto import tqdm as base_tqdm
@@ -15,13 +15,15 @@ from .errors import (
     RevisionNotFoundError,
 )
 from .file_download import REGEX_COMMIT_HASH, hf_hub_download, repo_folder_name
-from .hf_api import DatasetInfo, HfApi, ModelInfo, SpaceInfo
+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
 
 
 logger = logging.get_logger(__name__)
 
+VERY_LARGE_REPO_THRESHOLD = 50000  # After this limit, we don't consider `repo_info.siblings` to be reliable enough
+
 
 @validate_hf_hub_args
 def snapshot_download(
@@ -145,20 +147,22 @@ def snapshot_download(
 
     storage_folder = os.path.join(cache_dir, repo_folder_name(repo_id=repo_id, repo_type=repo_type))
 
+    api = HfApi(
+        library_name=library_name,
+        library_version=library_version,
+        user_agent=user_agent,
+        endpoint=endpoint,
+        headers=headers,
+        token=token,
+    )
+
     repo_info: Union[ModelInfo, DatasetInfo, SpaceInfo, None] = None
     api_call_error: Optional[Exception] = None
     if not local_files_only:
         # try/except logic to handle different errors => taken from `hf_hub_download`
         try:
             # if we have internet connection we want to list files to download
-            api = HfApi(
-                library_name=library_name,
-                library_version=library_version,
-                user_agent=user_agent,
-                endpoint=endpoint,
-                headers=headers,
-            )
-            repo_info = api.repo_info(repo_id=repo_id, repo_type=repo_type, revision=revision, token=token)
+            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
             raise
@@ -251,13 +255,31 @@ def snapshot_download(
     # => let's download the files!
     assert repo_info.sha is not None, "Repo info returned from server must have a revision sha."
     assert repo_info.siblings is not None, "Repo info returned from server must have a siblings list."
-    filtered_repo_files = list(
-        filter_repo_objects(
-            items=[f.rfilename for f in repo_info.siblings],
-            allow_patterns=allow_patterns,
-            ignore_patterns=ignore_patterns,
+
+    # Corner case: on very large repos, the siblings list in `repo_info` might not contain all files.
+    # In that case, we need to use the `list_repo_tree` method to prevent caching issues.
+    repo_files: Iterable[str] = [f.rfilename for f in repo_info.siblings]
+    has_many_files = len(repo_info.siblings) > VERY_LARGE_REPO_THRESHOLD
+    if has_many_files:
+        logger.info("The repo has more than 50,000 files. Using `list_repo_tree` to ensure all files are listed.")
+        repo_files = (
+            f.rfilename
+            for f in api.list_repo_tree(repo_id=repo_id, recursive=True, revision=revision, repo_type=repo_type)
+            if isinstance(f, RepoFile)
         )
+
+    filtered_repo_files: Iterable[str] = filter_repo_objects(
+        items=repo_files,
+        allow_patterns=allow_patterns,
+        ignore_patterns=ignore_patterns,
     )
+
+    if not has_many_files:
+        filtered_repo_files = list(filtered_repo_files)
+        tqdm_desc = f"Fetching {len(filtered_repo_files)} files"
+    else:
+        tqdm_desc = "Fetching ... files"
+
     commit_hash = repo_info.sha
     snapshot_folder = os.path.join(storage_folder, "snapshots", commit_hash)
     # if passed revision is not identical to commit_hash
@@ -305,7 +327,7 @@ def snapshot_download(
         thread_map(
             _inner_hf_hub_download,
             filtered_repo_files,
-            desc=f"Fetching {len(filtered_repo_files)} files",
+            desc=tqdm_desc,
             max_workers=max_workers,
             # User can use its own tqdm class or the default one from `huggingface_hub.utils`
             tqdm_class=tqdm_class or hf_tqdm,

huggingface_hub/inference/_mcp/agent.py

@@ -20,7 +20,7 @@ class Agent(MCPClient):
     </Tip>
 
     Args:
-        model (`str`):
+        model (`str`, *optional*):
             The model to run inference with. Can be a model id hosted on the Hugging Face Hub, e.g. `meta-llama/Meta-Llama-3-8B-Instruct`
             or a URL to a deployed Inference Endpoint or other local or remote endpoint.
         servers (`Iterable[Dict]`):
@@ -28,6 +28,8 @@ class Agent(MCPClient):
         provider (`str`, *optional*):
             Name of the provider to use for inference. Defaults to "auto" i.e. the first of the providers available for the model, sorted by the user's order in https://hf.co/settings/inference-providers.
             If model is a URL or `base_url` is passed, then `provider` is not used.
+        base_url (`str`, *optional*):
+            The base URL to run inference. Defaults to None.
         api_key (`str`, *optional*):
             Token to use for authentication. Will default to the locally Hugging Face saved token if not provided. You can also use your own provider API key to interact directly with the provider's service.
         prompt (`str`, *optional*):
@@ -37,13 +39,14 @@ class Agent(MCPClient):
     def __init__(
         self,
         *,
-        model: str,
+        model: Optional[str] = None,
         servers: Iterable[Dict],
         provider: Optional[PROVIDER_OR_POLICY_T] = None,
+        base_url: Optional[str] = None,
         api_key: Optional[str] = None,
         prompt: Optional[str] = None,
     ):
-        super().__init__(model=model, provider=provider, api_key=api_key)
+        super().__init__(model=model, provider=provider, base_url=base_url, api_key=api_key)
         self._servers_cfg = list(servers)
         self.messages: List[Union[Dict, ChatCompletionInputMessage]] = [
             {"role": "system", "content": prompt or DEFAULT_SYSTEM_PROMPT}

huggingface_hub/inference/_mcp/cli.py

@@ -1,6 +1,7 @@
 import asyncio
 import os
 import signal
+import traceback
 from functools import partial
 from typing import Any, Dict, List, Optional
 
@@ -71,8 +72,9 @@ async def run_agent(
             # Windows (or any loop that doesn't support it) : fall back to sync
             signal.signal(signal.SIGINT, lambda *_: _sigint_handler())
         async with Agent(
-            provider=config["provider"],
-            model=config["model"],
+            provider=config.get("provider"),
+            model=config.get("model"),
+            base_url=config.get("endpointUrl"),
             servers=servers,
             prompt=prompt,
         ) as agent:
@@ -123,9 +125,15 @@ async def run_agent(
                     print()
 
                 except Exception as e:
-                    print(f"\n[bold red]Error during agent run: {e}[/bold red]", flush=True)
+                    tb_str = traceback.format_exc()
+                    print(f"\n[bold red]Error during agent run: {e}\n{tb_str}[/bold red]", flush=True)
                     first_sigint = True  # Allow graceful interrupt for the next command
 
+    except Exception as e:
+        tb_str = traceback.format_exc()
+        print(f"\n[bold red]An unexpected error occurred: {e}\n{tb_str}[/bold red]", flush=True)
+        raise e
+
     finally:
         if sigint_registered_in_loop:
             try:

huggingface_hub/inference/_mcp/mcp_client.py

@@ -69,6 +69,8 @@ class MCPClient:
         provider (`str`, *optional*):
             Name of the provider to use for inference. Defaults to "auto" i.e. the first of the providers available for the model, sorted by the user's order in https://hf.co/settings/inference-providers.
             If model is a URL or `base_url` is passed, then `provider` is not used.
+        base_url (`str`, *optional*):
+            The base URL to run inference. Defaults to None.
         api_key (`str`, `optional`):
             Token to use for authentication. Will default to the locally Hugging Face saved token if not provided. You can also use your own provider API key to interact directly with the provider's service.
     """
@@ -76,17 +78,25 @@ class MCPClient:
     def __init__(
         self,
         *,
-        model: str,
+        model: Optional[str] = None,
         provider: Optional[PROVIDER_OR_POLICY_T] = None,
+        base_url: Optional[str] = None,
         api_key: Optional[str] = None,
     ):
         # Initialize MCP sessions as a dictionary of ClientSession objects
         self.sessions: Dict[ToolName, "ClientSession"] = {}
         self.exit_stack = AsyncExitStack()
         self.available_tools: List[ChatCompletionInputTool] = []
-
-        # Initialize the AsyncInferenceClient
-        self.client = AsyncInferenceClient(model=model, provider=provider, api_key=api_key)
+        # To be able to send the model in the payload if `base_url` is provided
+        if model is None and base_url is None:
+            raise ValueError("At least one of `model` or `base_url` should be set in `MCPClient`.")
+        self.payload_model = model
+        self.client = AsyncInferenceClient(
+            model=None if base_url is not None else model,
+            provider=provider,
+            api_key=api_key,
+            base_url=base_url,
+        )
 
     async def __aenter__(self):
         """Enter the context manager"""
@@ -244,6 +254,7 @@ class MCPClient:
 
         # Create the streaming request
         response = await self.client.chat.completions.create(
+            model=self.payload_model,
             messages=messages,
             tools=tools,
             tool_choice="auto",

huggingface_hub/serialization/_torch.py

@@ -246,7 +246,7 @@ def save_torch_state_dict(
             shared_tensors_to_discard=shared_tensors_to_discard,
         )
     else:
-        from torch import save as save_file_fn  # type: ignore[assignment]
+        from torch import save as save_file_fn  # type: ignore[assignment, no-redef]
 
         logger.warning(
             "You are using unsafe serialization. Due to security reasons, it is recommended not to load "

{huggingface_hub-0.32.1.dist-info → huggingface_hub-0.32.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: huggingface-hub
-Version: 0.32.1
+Version: 0.32.2
 Summary: Client library to download and publish models, datasets and other repos on the huggingface.co hub
 Home-page: https://github.com/huggingface/huggingface_hub
 Author: Hugging Face, Inc.

{huggingface_hub-0.32.1.dist-info → huggingface_hub-0.32.2.dist-info}/RECORD

@@ -1,11 +1,11 @@
-huggingface_hub/__init__.py,sha256=MraPcpVeTtz7kvp75UQIcplF8NG871xwGxDI4QHg_Mg,50644
+huggingface_hub/__init__.py,sha256=j0CPc82aFQA-FV3ew00Ukp6u4v-Lz3cBrH-K33Z50yE,50644
 huggingface_hub/_commit_api.py,sha256=ZbmuIhFdF8B3F_cvGtxorka7MmIQOk8oBkCtYltnCvI,39456
 huggingface_hub/_commit_scheduler.py,sha256=tfIoO1xWHjTJ6qy6VS6HIoymDycFPg0d6pBSZprrU2U,14679
 huggingface_hub/_inference_endpoints.py,sha256=qXR0utAYRaEWTI8EXzAsDpVDcYpp8bJPEBbcOxRS52E,17413
 huggingface_hub/_local_folder.py,sha256=7Uce_z51D7ZZ58GF7eUOtcq1cCuYQMOEF-W4p85iQTo,16885
 huggingface_hub/_login.py,sha256=ssf4viT5BhHI2ZidnSuAZcrwSxzaLOrf8xgRVKuvu_A,20298
 huggingface_hub/_oauth.py,sha256=YNbSSZCNZLiCqwMoYboSAfI3XjEsbyAADJcwgRAdhBc,18802
-huggingface_hub/_snapshot_download.py,sha256=RqhfsESBHwXAoZxVvw68W7vGhmXSbl7RoEFOPLvw3Ls,15186
+huggingface_hub/_snapshot_download.py,sha256=5BvLNm_1DgdBNXNWOP8omK-9unIyqNSVpDNmxfClxFk,16078
 huggingface_hub/_space_api.py,sha256=jb6rF8qLtjaNU12D-8ygAPM26xDiHCu8CHXHowhGTmg,5470
 huggingface_hub/_tensorboard_logger.py,sha256=ZkYcAUiRC8RGL214QUYtp58O8G5tn-HF6DCWha9imcA,8358
 huggingface_hub/_upload_large_folder.py,sha256=elY5Rv2YVJECVpdZ9PM1zdO8kG-jmi8DifLOa7aC3EU,24178
@@ -81,10 +81,10 @@ huggingface_hub/inference/_generated/types/zero_shot_classification.py,sha256=BA
 huggingface_hub/inference/_generated/types/zero_shot_image_classification.py,sha256=8J9n6VqFARkWvPfAZNWEG70AlrMGldU95EGQQwn06zI,1487
 huggingface_hub/inference/_generated/types/zero_shot_object_detection.py,sha256=GUd81LIV7oEbRWayDlAVgyLmY596r1M3AW0jXDp1yTA,1630
 huggingface_hub/inference/_mcp/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-huggingface_hub/inference/_mcp/agent.py,sha256=RDpqy2rcVKy0euIK43OAAczLlx0UH1QIACHMFjTItq0,4077
-huggingface_hub/inference/_mcp/cli.py,sha256=Pw6nZP57uzDWZGXRVCAqLHsLYO_WiQd4UPcQdXiCa-o,5557
+huggingface_hub/inference/_mcp/agent.py,sha256=azX9_lsFjNlgsEvRYdKgsmOmpNReWIcbuMeIVWc852k,4264
+huggingface_hub/inference/_mcp/cli.py,sha256=q7WUlc7K9cHffKnS8uVydqM7IGxDz2gnwoNKkiRF5GU,5873
 huggingface_hub/inference/_mcp/constants.py,sha256=tE_V6qcvsmvVoJa4eg04jhoTR2Cx1cNHieY2ENrm1_M,2511
-huggingface_hub/inference/_mcp/mcp_client.py,sha256=MwoovV2UTf7F8rxxQucAd4WSnIiJ71wx09QCrXDrP0Q,13477
+huggingface_hub/inference/_mcp/mcp_client.py,sha256=HUDkEURrWeQ9AY7W4_H4jca0BM12rKPbjdP4vNaXiwc,13998
 huggingface_hub/inference/_mcp/utils.py,sha256=K7rr4FxCh9OYWwYNlnvQraNLy9y3z-5yVMBIaoCQMjA,4052
 huggingface_hub/inference/_providers/__init__.py,sha256=IrLTMERrbRuPiVdBQEMK9TMvXrsGId4-u2ucMkG-vTU,7671
 huggingface_hub/inference/_providers/_common.py,sha256=Octgz-PbHw62iW3Oa8rF7rxvBJR0ZmL4ouv3NoX-weE,10131
@@ -106,7 +106,7 @@ huggingface_hub/serialization/__init__.py,sha256=kn-Fa-m4FzMnN8lNsF-SwFcfzug4Cuc
 huggingface_hub/serialization/_base.py,sha256=Df3GwGR9NzeK_SD75prXLucJAzPiNPgHbgXSw-_LTk8,8126
 huggingface_hub/serialization/_dduf.py,sha256=s42239rLiHwaJE36QDEmS5GH7DSmQ__BffiHJO5RjIg,15424
 huggingface_hub/serialization/_tensorflow.py,sha256=zHOvEMg-JHC55Fm4roDT3LUCDO5zB9qtXZffG065RAM,3625
-huggingface_hub/serialization/_torch.py,sha256=hJglq5F56s3k06GfLuYKQV4bSSjXQLuk1CC9l1M3Fmo,45191
+huggingface_hub/serialization/_torch.py,sha256=jpBmuSZJymMpvLcDcMaNxDu_fE5VkY_pAVH8e8stYIo,45201
 huggingface_hub/templates/datasetcard_template.md,sha256=W-EMqR6wndbrnZorkVv56URWPG49l7MATGeI015kTvs,5503
 huggingface_hub/templates/modelcard_template.md,sha256=4AqArS3cqdtbit5Bo-DhjcnDFR-pza5hErLLTPM4Yuc,6870
 huggingface_hub/utils/__init__.py,sha256=ORfVkn5D0wuLIq12jjhTzn5_c4F8fRPxB7TG-iednuQ,3722
@@ -137,9 +137,9 @@ huggingface_hub/utils/insecure_hashlib.py,sha256=iAaepavFZ5Dhfa5n8KozRfQprKmvcjS
 huggingface_hub/utils/logging.py,sha256=0A8fF1yh3L9Ka_bCDX2ml4U5Ht0tY8Dr3JcbRvWFuwo,4909
 huggingface_hub/utils/sha.py,sha256=OFnNGCba0sNcT2gUwaVCJnldxlltrHHe0DS_PCpV3C4,2134
 huggingface_hub/utils/tqdm.py,sha256=xAKcyfnNHsZ7L09WuEM5Ew5-MDhiahLACbbN2zMmcLs,10671
-huggingface_hub-0.32.1.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-huggingface_hub-0.32.1.dist-info/METADATA,sha256=y3e9ulpqrSxlIWNwdsA0aXk-ZK9nVDChQ5fjApF_hNA,14777
-huggingface_hub-0.32.1.dist-info/WHEEL,sha256=tZoeGjtWxWRfdplE7E3d45VPlLNQnvbKiYnx7gwAy8A,92
-huggingface_hub-0.32.1.dist-info/entry_points.txt,sha256=uelw0-fu0kd-CxIuOsR1bsjLIFnAaMQ6AIqluJYDhQw,184
-huggingface_hub-0.32.1.dist-info/top_level.txt,sha256=8KzlQJAY4miUvjAssOAJodqKOw3harNzuiwGQ9qLSSk,16
-huggingface_hub-0.32.1.dist-info/RECORD,,
+huggingface_hub-0.32.2.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+huggingface_hub-0.32.2.dist-info/METADATA,sha256=w2Df9xLc5FhDmwO11yIxMDiHc-i7aqLvXFRzRvzgI_E,14777
+huggingface_hub-0.32.2.dist-info/WHEEL,sha256=tZoeGjtWxWRfdplE7E3d45VPlLNQnvbKiYnx7gwAy8A,92
+huggingface_hub-0.32.2.dist-info/entry_points.txt,sha256=uelw0-fu0kd-CxIuOsR1bsjLIFnAaMQ6AIqluJYDhQw,184
+huggingface_hub-0.32.2.dist-info/top_level.txt,sha256=8KzlQJAY4miUvjAssOAJodqKOw3harNzuiwGQ9qLSSk,16
+huggingface_hub-0.32.2.dist-info/RECORD,,