fix push_to_ms (#1901)

(cherry picked from commit eee2323)

swift/trainers/push_to_ms.py

@@ -1,6 +1,5 @@
 import os
 import tempfile
-from concurrent.futures import Future
 from functools import partial
 from pathlib import Path
 from typing import List, Optional, Union
@@ -13,75 +12,75 @@
 logger = logging.get_logger(__name__)
 
 
-class PushToMsHubMixin:
+def create_repo(repo_id: str, *, token: Union[str, bool, None] = None, private: bool = False, **kwargs) -> RepoUrl:
+    from modelscope.hub.repository import Repository
+    hub_model_id = PushToMsHubMixin.create_ms_repo(repo_id, token, private)
+    PushToMsHubMixin.ms_token = token
+    with tempfile.TemporaryDirectory() as temp_cache_dir:
+        repo = Repository(temp_cache_dir, hub_model_id)
+        PushToMsHubMixin.add_patterns_to_gitattributes(repo, ['*.safetensors', '*.bin', '*.pt'])
+        # Add 'runs/' to .gitignore, ignore tensorboard files
+        PushToMsHubMixin.add_patterns_to_gitignore(repo, ['runs/', 'images/'])
+        PushToMsHubMixin.add_patterns_to_file(
+            repo,
+            'configuration.json', ['{"framework": "pytorch", "task": "text-generation", "allow_remote": true}'],
+            ignore_push_error=True)
+        # Add '*.sagemaker' to .gitignore if using SageMaker
+        if os.environ.get('SM_TRAINING_ENV'):
+            PushToMsHubMixin.add_patterns_to_gitignore(repo, ['*.sagemaker-uploading', '*.sagemaker-uploaded'],
+                                                       'Add `*.sagemaker` patterns to .gitignore')
+    return RepoUrl(url=hub_model_id, )
+
+
+@future_compatible
+def upload_folder(
+    self,
+    *,
+    repo_id: str,
+    folder_path: Union[str, Path],
+    path_in_repo: Optional[str] = None,
+    commit_message: Optional[str] = None,
+    commit_description: Optional[str] = None,
+    token: Union[str, bool, None] = None,
+    revision: Optional[str] = 'master',
+    ignore_patterns: Optional[Union[List[str], str]] = None,
+    run_as_future: bool = False,
+    **kwargs,
+):
+    from modelscope import push_to_hub
+    commit_message = commit_message or 'Upload folder using api'
+    if commit_description:
+        commit_message = commit_message + '\n' + commit_description
+    if not os.path.exists(os.path.join(folder_path, 'configuration.json')):
+        with open(os.path.join(folder_path, 'configuration.json'), 'w') as f:
+            f.write('{"framework": "pytorch", "task": "text-generation", "allow_remote": true}')
+    if ignore_patterns:
+        ignore_patterns = [p for p in ignore_patterns if p != '_*']
+    if path_in_repo:
+        # We don't support part submit for now
+        path_in_repo = os.path.basename(folder_path)
+        folder_path = os.path.dirname(folder_path)
+        ignore_patterns = []
+    push_to_hub(
+        repo_id,
+        folder_path,
+        token or PushToMsHubMixin.ms_token,
+        commit_message=commit_message,
+        ignore_file_pattern=ignore_patterns,
+        revision=revision,
+        tag=path_in_repo)
+    return CommitInfo(
+        commit_url=f'https://www.modelscope.cn/models/{repo_id}/files',
+        commit_message=commit_message,
+        commit_description=commit_description,
+        oid=None,
+    )
 
-    _use_hf_hub = strtobool(os.environ.get('USE_HF', 'False'))
-    _token = None
 
-    @staticmethod
-    def create_repo(repo_id: str, *, token: Union[str, bool, None] = None, private: bool = False, **kwargs) -> RepoUrl:
-        from modelscope.hub.repository import Repository
-        hub_model_id = PushToMsHubMixin._create_ms_repo(repo_id, token, private)
-        PushToMsHubMixin._token = token
-        with tempfile.TemporaryDirectory() as temp_cache_dir:
-            repo = Repository(temp_cache_dir, hub_model_id)
-            PushToMsHubMixin._add_patterns_to_gitattributes(repo, ['*.safetensors', '*.bin', '*.pt'])
-            # Add 'runs/' to .gitignore, ignore tensorboard files
-            PushToMsHubMixin._add_patterns_to_gitignore(repo, ['runs/', 'images/'])
-            PushToMsHubMixin._add_patterns_to_file(
-                repo,
-                'configuration.json', ['{"framework": "pytorch", "task": "text-generation", "allow_remote": true}'],
-                ignore_push_error=True)
-            # Add '*.sagemaker' to .gitignore if using SageMaker
-            if os.environ.get('SM_TRAINING_ENV'):
-                PushToMsHubMixin._add_patterns_to_gitignore(repo, ['*.sagemaker-uploading', '*.sagemaker-uploaded'],
-                                                            'Add `*.sagemaker` patterns to .gitignore')
-        return RepoUrl(url=hub_model_id, )
+class PushToMsHubMixin:
 
-    @staticmethod
-    @future_compatible
-    def upload_folder(
-        self,
-        *,
-        repo_id: str,
-        folder_path: Union[str, Path],
-        path_in_repo: Optional[str] = None,
-        commit_message: Optional[str] = None,
-        commit_description: Optional[str] = None,
-        token: Union[str, bool, None] = None,
-        revision: Optional[str] = 'master',
-        ignore_patterns: Optional[Union[List[str], str]] = None,
-        run_as_future: bool = False,
-        **kwargs,
-    ) -> Union[CommitInfo, str, Future[CommitInfo], Future[str]]:
-        from modelscope import push_to_hub
-        commit_message = commit_message or 'Upload folder using api'
-        if commit_description:
-            commit_message = commit_message + '\n' + commit_description
-        if not os.path.exists(os.path.join(folder_path, 'configuration.json')):
-            with open(os.path.join(folder_path, 'configuration.json'), 'w') as f:
-                f.write('{"framework": "pytorch", "task": "text-generation", "allow_remote": true}')
-        if ignore_patterns:
-            ignore_patterns = [p for p in ignore_patterns if p != '_*']
-        if path_in_repo:
-            # We don't support part submit for now
-            path_in_repo = os.path.basename(folder_path)
-            folder_path = os.path.dirname(folder_path)
-            ignore_patterns = []
-        push_to_hub(
-            repo_id,
-            folder_path,
-            token or PushToMsHubMixin._token,
-            commit_message=commit_message,
-            ignore_file_pattern=ignore_patterns,
-            revision=revision,
-            tag=path_in_repo)
-        return CommitInfo(
-            commit_url=f'https://www.modelscope.cn/models/{repo_id}/files',
-            commit_message=commit_message,
-            commit_description=commit_description,
-            oid=None,
-        )
+    _use_hf_hub = strtobool(os.environ.get('USE_HF', 'False'))
+    ms_token = None
 
     if not _use_hf_hub:
         import huggingface_hub
@@ -93,7 +92,7 @@ def upload_folder(
         trainer.upload_folder = partial(upload_folder, api)
 
     @staticmethod
-    def _create_ms_repo(hub_model_id: str, hub_token: Optional[str] = None, hub_private_repo: bool = False) -> str:
+    def create_ms_repo(hub_model_id: str, hub_token: Optional[str] = None, hub_private_repo: bool = False) -> str:
         from modelscope import HubApi
         from modelscope.hub.api import ModelScopeConfig
         from modelscope.hub.constants import ModelVisibility
@@ -121,11 +120,11 @@ def _create_ms_repo(hub_model_id: str, hub_token: Optional[str] = None, hub_priv
         return hub_model_id
 
     @staticmethod
-    def _add_patterns_to_file(repo,
-                              file_name: str,
-                              patterns: List[str],
-                              commit_message: Optional[str] = None,
-                              ignore_push_error=False) -> None:
+    def add_patterns_to_file(repo,
+                             file_name: str,
+                             patterns: List[str],
+                             commit_message: Optional[str] = None,
+                             ignore_push_error=False) -> None:
         if isinstance(patterns, str):
             patterns = [patterns]
         if commit_message is None:
@@ -161,11 +160,11 @@ def _add_patterns_to_file(repo,
                 raise e
 
     @staticmethod
-    def _add_patterns_to_gitignore(repo, patterns: List[str], commit_message: Optional[str] = None) -> None:
-        PushToMsHubMixin._add_patterns_to_file(repo, '.gitignore', patterns, commit_message, ignore_push_error=True)
+    def add_patterns_to_gitignore(repo, patterns: List[str], commit_message: Optional[str] = None) -> None:
+        PushToMsHubMixin.add_patterns_to_file(repo, '.gitignore', patterns, commit_message, ignore_push_error=True)
 
     @staticmethod
-    def _add_patterns_to_gitattributes(repo, patterns: List[str], commit_message: Optional[str] = None) -> None:
+    def add_patterns_to_gitattributes(repo, patterns: List[str], commit_message: Optional[str] = None) -> None:
         new_patterns = []
         suffix = 'filter=lfs diff=lfs merge=lfs -text'
         for pattern in patterns:
@@ -175,4 +174,4 @@ def _add_patterns_to_gitattributes(repo, patterns: List[str], commit_message: Op
         file_name = '.gitattributes'
         if commit_message is None:
             commit_message = f'Add `{patterns[0]}` patterns to {file_name}'
-        PushToMsHubMixin._add_patterns_to_file(repo, file_name, new_patterns, commit_message, ignore_push_error=True)
+        PushToMsHubMixin.add_patterns_to_file(repo, file_name, new_patterns, commit_message, ignore_push_error=True)