atomgit 1.0.7__tar.gz → 1.0.9__tar.gz

{atomgit-1.0.7 → atomgit-1.0.9}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: atomgit
-Version: 1.0.7
+Version: 1.0.9
 Summary: AtomGit模型文件上传下载CLI工具
 Home-page: https://atomgit.com/gitcode-ai/atomgit_cli
 Author: AtomGit CLI Team

{atomgit-1.0.7 → atomgit-1.0.9}/__init__.py

@@ -8,7 +8,7 @@ AtomGit CLI - 基于Transformers和Hugging Face Hub的模型文件上传下载
 支持模型和数据集的上传、下载等操作。
 """
 
-__version__ = '1.0.7'
+__version__ = '1.0.9'
 __author__ = 'AtomGit CLI Team'
 __description__ = 'AtomGit模型文件上传下载CLI工具'
 
@@ -31,6 +31,6 @@ except ImportError:
 
 __all__ = [
     'config', 'api', 'cli',
-    'snapshot_download', 'hub_download_url', 'download_file', 
+    'snapshot_download', 'hub_download_url', 'download_file',
     'upload_folder', 'create_repository'
-] 
+]

{atomgit-1.0.7 → atomgit-1.0.9}/api.py

@@ -16,6 +16,10 @@ os.environ["HF_HOME"] = cache_dir
 
 
 from huggingface_hub import hf_hub_download, upload_folder, create_repo, snapshot_download, constants as hf_constants
+try:
+    from .atomgit_hub import snapshot_download as atomgit_snapshot_download
+except ImportError:
+    from atomgit_hub import snapshot_download as atomgit_snapshot_download
 try:
     from .rate_limiter import install_hf_rate_limiter
 except ImportError:
@@ -223,7 +227,7 @@ class HuggingFaceAPI:
             # 首先尝试不使用token下载（适用于公开仓库）
             credentials = config.get_credentials()
             try:
-                snapshot_download(
+                atomgit_snapshot_download(
                     repo_id=normalized_repo_id,
                     local_dir=str(local_path),
                     force_download=force_download,  # 根据用户选择决定是否强制下载

{atomgit-1.0.7 → atomgit-1.0.9}/atomgit.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: atomgit
-Version: 1.0.7
+Version: 1.0.9
 Summary: AtomGit模型文件上传下载CLI工具
 Home-page: https://atomgit.com/gitcode-ai/atomgit_cli
 Author: AtomGit CLI Team

{atomgit-1.0.7 → atomgit-1.0.9}/atomgit.egg-info/SOURCES.txt

@@ -9,7 +9,6 @@ config.py
 rate_limiter.py
 requirements.txt
 setup.py
-test.py
 utils.py
 ./__init__.py
 ./__main__.py
@@ -18,7 +17,6 @@ utils.py
 ./cli.py
 ./config.py
 ./rate_limiter.py
-./test.py
 ./utils.py
 atomgit.egg-info/PKG-INFO
 atomgit.egg-info/SOURCES.txt

{atomgit-1.0.7 → atomgit-1.0.9}/atomgit_hub.py

@@ -22,7 +22,9 @@ os.makedirs(cache_dir, exist_ok=True)
 os.environ["HF_HOME"] = cache_dir
 
 from huggingface_hub import snapshot_download as hf_snapshot_download
-from huggingface_hub import hf_hub_download, upload_folder as hf_upload_folder, create_repo
+from huggingface_hub import hf_hub_download, upload_file as hf_upload_file, upload_folder as hf_upload_folder, create_repo
+from huggingface_hub import HfApi
+from concurrent.futures import ThreadPoolExecutor
 try:
     from .rate_limiter import install_hf_rate_limiter
 except ImportError:
@@ -86,6 +88,54 @@ def _get_token() -> Optional[str]:
     return None
 
 
+def _snapshot_download_from_tree(
+    repo_id: str,
+    local_dir: Union[str, Path],
+    revision: Optional[str],
+    token: Optional[Union[str, bool]],
+    force_download: bool,
+    allow_patterns: Optional[Union[List[str], str]],
+    ignore_patterns: Optional[Union[List[str], str]],
+    max_workers: int,
+) -> str:
+    """Download through list_repo_tree when repo_info.siblings is unavailable."""
+    from huggingface_hub.utils import filter_repo_objects
+    from huggingface_hub.hf_api import RepoFile
+
+    api = HfApi()
+    files = [
+        item.path
+        for item in api.list_repo_tree(
+            repo_id=repo_id, recursive=True, revision=revision, token=token
+        )
+        if isinstance(item, RepoFile)
+    ]
+    files = list(filter_repo_objects(
+        items=files,
+        allow_patterns=allow_patterns,
+        ignore_patterns=ignore_patterns,
+    ))
+    if not files:
+        raise FileNotFoundError(f"仓库没有可下载的文件: {repo_id}")
+
+    destination = Path(local_dir)
+    destination.mkdir(parents=True, exist_ok=True)
+
+    def download_one(filename: str) -> None:
+        hf_hub_download(
+            repo_id=repo_id,
+            filename=filename,
+            revision=revision,
+            token=token,
+            local_dir=str(destination),
+            force_download=force_download,
+        )
+
+    with ThreadPoolExecutor(max_workers=max(1, max_workers)) as executor:
+        list(executor.map(download_one, files))
+    return str(destination.resolve())
+
+
 def snapshot_download(
     repo_id: str,
     revision: Optional[str] = None,
@@ -181,6 +231,24 @@ def snapshot_download(
         return result
     except Exception as e:
         error_msg = str(e)
+        if local_dir is not None and (
+            "min() iterable argument is empty" in error_msg
+            or "min() arg is an empty sequence" in error_msg
+            or "404 Not Found" in error_msg
+        ):
+            try:
+                return _snapshot_download_from_tree(
+                    repo_id=normalized_repo_id,
+                    local_dir=local_dir,
+                    revision=revision,
+                    token=token,
+                    force_download=force_download,
+                    allow_patterns=allow_patterns,
+                    ignore_patterns=ignore_patterns,
+                    max_workers=max_workers,
+                )
+            except Exception as fallback_error:
+                error_msg = str(fallback_error)
         if "401" in error_msg or "403" in error_msg:
             raise Exception(f"认证失败：{error_msg}。请检查token是否正确，或使用 'atomgit login' 重新登录。")
         elif "404" in error_msg:
@@ -291,18 +359,21 @@ def upload_folder(
     upload_timeout: float = 300.0,
 ) -> str:
     """
-    上传文件夹到AtomGit Hub
-    
+    上传文件或文件夹到AtomGit Hub
+
+    自动判断传入路径是文件还是文件夹，两者都支持。
+
     参数:
-        folder_path (str 或 Path): 本地文件夹路径
+        folder_path (str 或 Path): 本地文件或文件夹路径
         repo_id (str): 仓库ID
         token (str, 可选): 认证token
         repo_type (str, 可选): 仓库类型
         revision (str, 可选): 分支名
         commit_message (str, 可选): 提交消息
         commit_description (str, 可选): 提交描述
-        path_in_repo (str, 可选): 在仓库中的路径，默认为根目录
-        ignore_patterns (List[str], 可选): 要忽略的文件模式
+        path_in_repo (str, 可选): 在仓库中的路径，默认为根目录。
+                                  若传入的是文件，文件名会自动拼接到该路径末尾。
+        ignore_patterns (List[str], 可选): 要忽略的文件模式（仅文件夹上传生效）
         upload_timeout (float, 可选): 上传超时时间（秒），默认60秒（1分钟）。
                                       对于大文件，服务器处理响应可能需要较长时间。
     
@@ -313,6 +384,7 @@ def upload_folder(
         >>> upload_folder("./my-model/", "username/repo-name")
         >>> upload_folder("./data/", "username/repo", path_in_repo="datasets/")
         >>> upload_folder("./big-files/", "username/repo", upload_timeout=1200.0)  # 20分钟超时
+        >>> upload_folder("./weights.pth", "username/repo", path_in_repo="checkpoints/")  # 上传单个文件
     """
     # 标准化仓库ID
     normalized_repo_id = _normalize_repo_id(repo_id)
@@ -321,10 +393,10 @@ def upload_folder(
     folder_path = Path(folder_path)
     
     if not folder_path.exists():
-        raise FileNotFoundError(f"文件夹不存在: {folder_path}")
-    
-    if not folder_path.is_dir():
-        raise NotADirectoryError(f"路径不是目录: {folder_path}")
+        raise FileNotFoundError(f"路径不存在: {folder_path}")
+
+    if not folder_path.is_file() and not folder_path.is_dir():
+        raise ValueError(f"路径既不是文件也不是目录: {folder_path}")
     
     # 如果没有提供token，尝试使用保存的token
     if token is None:
@@ -332,42 +404,62 @@ def upload_folder(
         if token is None:
             raise Exception("上传需要认证token，请先使用 'atomgit login' 登录，或提供token参数。")
     
-    # 直接使用原始目录，或者创建临时目录来重新组织结构
+    # 使用 Monkey Patch 方式临时修改 huggingface_hub 的默认超时配置
+    from huggingface_hub import constants as hf_constants
     import tempfile
     import shutil
 
-    if path_in_repo == "./" or path_in_repo == "." or path_in_repo == "":
-        # 如果要上传到根目录，直接使用源文件夹
-        upload_path = str(folder_path)
-    else:
-        # 如果要上传到特定路径，需要重新组织目录结构。
-        # 注意：不能提前退出 with TemporaryDirectory()，否则上传时临时目录已被删除。
-        temp_dir = tempfile.mkdtemp(prefix="atomgit_upload_")
-        try:
-            temp_path = Path(temp_dir)
-            
-            # 创建目标路径
-            target_path = temp_path / path_in_repo.strip('./')
-            target_path.parent.mkdir(parents=True, exist_ok=True)
-            
-            # 复制整个目录树
-            shutil.copytree(folder_path, target_path, dirs_exist_ok=True)
-            
-        except Exception:
-            shutil.rmtree(temp_dir, ignore_errors=True)
-            raise
-
-        upload_path = str(temp_path)
-    # 使用 Monkey Patch 方式临时修改 huggingface_hub 的默认超时配置
-    from huggingface_hub import constants as hf_constants
-    
     # 保存原始超时配置
     original_timeout = hf_constants.DEFAULT_REQUEST_TIMEOUT
-    
+
     # 临时修改超时配置
     hf_constants.DEFAULT_REQUEST_TIMEOUT = upload_timeout
-    
+
+    temp_dir = None
     try:
+        # 上传单个文件
+        if folder_path.is_file():
+            # path_in_repo 按目录语义处理（与文件夹分支一致，去掉首尾 ./ 和 /），
+            # 文件名自动拼接到末尾
+            clean_path = path_in_repo.strip().strip("./")
+            if not clean_path:
+                target_path = folder_path.name
+            else:
+                target_path = clean_path + "/" + folder_path.name
+            commit_msg = commit_message or f"Upload file {folder_path.name}"
+            result = hf_upload_file(
+                repo_id=normalized_repo_id,
+                path_or_fileobj=str(folder_path),
+                path_in_repo=target_path,
+                token=token,
+                commit_message=commit_msg
+            )
+            return result
+
+        # 上传文件夹：直接使用原始目录，或者创建临时目录来重新组织结构
+        if path_in_repo == "./" or path_in_repo == "." or path_in_repo == "":
+            # 如果要上传到根目录，直接使用源文件夹
+            upload_path = str(folder_path)
+        else:
+            # 如果要上传到特定路径，需要重新组织目录结构。
+            # 注意：不能提前退出 with TemporaryDirectory()，否则上传时临时目录已被删除。
+            temp_dir = tempfile.mkdtemp(prefix="atomgit_upload_")
+            try:
+                temp_path = Path(temp_dir)
+
+                # 创建目标路径
+                target_path = temp_path / path_in_repo.strip('./')
+                target_path.parent.mkdir(parents=True, exist_ok=True)
+
+                # 复制整个目录树
+                shutil.copytree(folder_path, target_path, dirs_exist_ok=True)
+
+            except Exception:
+                shutil.rmtree(temp_dir, ignore_errors=True)
+                raise
+
+            upload_path = str(temp_path)
+
         # 使用huggingface_hub的upload_folder上传
         commit_msg = commit_message or f"Upload folder {folder_path.name}"
         result = hf_upload_folder(
@@ -376,7 +468,7 @@ def upload_folder(
             token=token,
             commit_message=commit_msg
         )
-        
+
         return result
 
     except Exception as e:
@@ -388,8 +480,10 @@ def upload_folder(
         else:
             raise Exception(f"上传失败：{error_msg}")
     finally:
+        # 恢复原始超时配置
+        hf_constants.DEFAULT_REQUEST_TIMEOUT = original_timeout
         # 清理临时目录（仅重新组织目录结构时创建）
-        if path_in_repo not in ("./", ".", ""):
+        if temp_dir is not None:
             shutil.rmtree(temp_dir, ignore_errors=True)
 
 

{atomgit-1.0.7 → atomgit-1.0.9}/cli.py

@@ -39,7 +39,7 @@ except ImportError:
 
 
 @click.group()
-@click.version_option(version='1.0.7')
+@click.version_option(version='1.0.9')
 def cli():
     """AtomGit CLI - 基于Transformers和Hugging Face Hub的AtomGit平台模型文件上传下载工具"""
     pass
@@ -292,4 +292,4 @@ def config_show():
 
 
 if __name__ == '__main__':
-    cli() 
+    cli()

{atomgit-1.0.7 → atomgit-1.0.9}/setup.py

@@ -31,7 +31,7 @@ def read_requirements():
 
 setup(
     name='atomgit',
-    version='1.0.7',
+    version='1.0.9',
     author='AtomGit CLI Team',
     author_email='sa@atomgit.com',
     description='AtomGit模型文件上传下载CLI工具',
@@ -69,4 +69,4 @@ setup(
         'Bug Reports': 'https://atomgit.com/gitcode-ai/atomgit_cli/issues',
         'Source': 'https://atomgit.com/gitcode-ai/atomgit_cli',
     },
-) 
+)

atomgit-1.0.7/test.py

@@ -1,154 +0,0 @@
-#!/usr/bin/env python3
-# -*- coding: utf-8 -*-
-"""
-AtomGit Hub 功能测试
-"""
-
-from atomgit_hub import (
-    create_repository, 
-    upload_folder, 
-    snapshot_download, 
-    hub_download_url, 
-    download_file, 
-    load_dataset
-)
-from pathlib import Path
-
-
-def test_create_model(repo_id: str, repo_type: str):
-    print("\n=== 测试 create_repository ===")
-    try:
-        result = create_repository(
-            repo_id=repo_id,
-            token=None,  # 使用保存的token
-            private=True,
-            repo_type=repo_type,
-            exist_ok=True,
-            space_sdk=None,
-            space_hardware=None,
-            space_storage=None,
-            space_sleep_time=None,
-            space_secrets=None,
-            space_variables=None
-        )
-        print(f"✅ 创建仓库成功: {result}")
-    except Exception as e:
-        print(f"❌ 创建仓库失败: {e}")
-
-def test_upload_folder(repo_id: str, folder_path: str):
-    print("\n=== 测试 upload_folder ===")
-    
-    try:
-        result = upload_folder(
-            folder_path=folder_path,
-            repo_id=repo_id,
-            token=None,  # 使用保存的token
-            commit_message="Test upload with all parameters",
-            commit_description="Full parameter test for upload_folder function",
-            upload_timeout=100
-        )
-        print(f"✅ 上传文件夹成功: {result}")
-    except Exception as e:
-        print(f"❌ 上传文件夹失败: {e}")
-
-def test_snapshot_download(repo_id: str, local_dir: str):
-    print("\n=== 测试 snapshot_download ===")
-    try:
-        local_path = snapshot_download(
-            repo_id=repo_id,
-            local_dir=local_dir,
-            proxies=None,
-            etag_timeout=10,
-            resume_download=False,
-            force_download=True,
-            token=None,  # 使用保存的token
-            local_files_only=False,
-            max_workers=8,
-            tqdm_class=None
-        )
-        print(f"✅ 下载快照成功: {local_path}")
-    except Exception as e:
-        print(f"❌ 下载快照失败: {e}")
-
-def test_download_file(repo_id: str, filename: str, local_dir: str):
-    print("\n=== 测试 download_file ===")
-    try:
-        file_path = download_file(
-            repo_id=repo_id,
-            filename=filename,
-            local_dir=local_dir,
-            token=None,  # 使用保存的token
-            force_download=True
-        )
-        print(f"✅ 下载文件成功: {file_path}")
-    except Exception as e:
-        print(f"❌ 下载文件失败: {e}")
-
-
-def test_hub_download_url(repo_id: str, filename: str, repo_type: str):
-    print("\n=== 测试 hub_download_url ===")
-    try:
-        url = hub_download_url(
-            repo_id=repo_id,
-            filename=filename,
-            repo_type=repo_type
-        )
-        print(f"✅ 获取URL成功: {url}")
-    except Exception as e:
-        print(f"❌ 获取URL失败: {e}")
-
-def test_load_dataset(repo_id: str, local_dir: str):
-    print("\n=== 测试 load_dataset ===")
-    try:
-        dataset = load_dataset(
-            path=repo_id,
-            cache_dir=local_dir,
-            token=None
-        )
-        print(f"✅ 加载数据集成功: {dataset}")
-    except Exception as e:
-        print(f"❌ 加载数据集失败: {e}")
-
-
-def run_all_tests():
-    """运行所有测试"""
-    print("=" * 60)
-    print("开始运行 AtomGit Hub 完整参数测试")
-    print("=" * 60)
-    model = "ai-test/atomgit-cli-test-model-full"
-    dataset = "ai-test/atomgit-cli-test-dataset-full"
-    
-    model_folder_path="/Users/yanlp/csdn/aipython/gitcode_cli/site/yanlp-model-1"
-    dataset_folder_path="/Users/yanlp/csdn/aipython/gitcode_cli/site/yanlp-model-1"
-
-    test_dataset = "yanlp/glaive_toolcall_zh"
-    # 注意：按照依赖顺序执行测试
-    # 1. 先创建仓库
-    # test_create_model(repo_id=model, repo_type="model")
-    # test_create_model(repo_id=dataset, repo_type="dataset")
-    
-    # 2. 测试upload
-    # test_upload_folder(repo_id=model, folder_path=model_folder_path)
-    # test_upload_folder(repo_id=dataset, folder_path=dataset_folder_path)
-    test_upload_folder(repo_id="yanlp/dataset-t2", folder_path="/Users/yanlp/csdn/IdeaProjects/gitcode/gitcode-hf-registry/dataset-t11")
-
-    # 3. 测试下载
-    # test_snapshot_download(repo_id=model, local_dir="./test_downloads/model/snapshot_full")
-    # test_snapshot_download(repo_id="hf_mirrors/Qwen/Qwen3-0.6B", local_dir="./test_downloads/dataset/snapshot_full2")
-    
-    # 4. 测试下载文件
-    # test_download_file(repo_id=model, filename="README.md", local_dir="./test_downloads/model/single_file")
-
-    # 5. 测试URL获取
-    # test_hub_download_url(repo_id=model, filename="model.safetensors", repo_type="model")
-
-    # 6. 测试数据集加载
-    # test_load_dataset(repo_id=test_dataset, local_dir="./test_downloads/dataset/load_dataset")
-    
-    print("\n" + "=" * 60)
-    print("所有测试完成")
-    print("=" * 60)
-
-
-if __name__ == "__main__":
-    run_all_tests()