fiuai-s3 0.3.0__py3-none-any.whl → 0.3.2__py3-none-any.whl

fiuai_s3/__init__.py

@@ -6,6 +6,7 @@
 # Copyright (c) 2025 FiuAI
 
 from .object_storage import ObjectStorage, ObjectStorageFactory, StorageConfig
+from .type import DocFileObject, DocSourceFrom, DocFileType
 
-__version__ = "0.1.0"
-__all__ = ["ObjectStorage", "ObjectStorageFactory", "StorageConfig"] 
+__version__ = "0.3.2"
+__all__ = ["ObjectStorage", "ObjectStorageFactory", "StorageConfig", "DocFileObject", "DocSourceFrom", "DocFileType"] 

fiuai_s3/minio/minio_storage.py

@@ -12,7 +12,7 @@ from minio import Minio
 from minio.error import S3Error
 from minio.commonconfig import Tags
 from ..object_storage import ObjectStorage, StorageConfig
-from ..type import DocFileObject, DocFileType       
+from ..type import DocFileObject, DocFileType, DocSourceFrom
 
 logger = logging.getLogger(__name__)
 
@@ -163,14 +163,30 @@ class MinioStorage(ObjectStorage):
             raise ValueError("auth_tenant_id、auth_company_id、doc_id 不能为空")
         return f"{tenant_id}/{company_id}/{docid}/{filename}"
 
-    def upload_doc_file(self, filename: str, data: bytes, tags: Optional[Dict[str, str]] = None, auth_tenant_id: Optional[str] = None, auth_company_id: Optional[str] = None, doc_id: Optional[str] = None) -> bool:
+    def upload_doc_file(self, 
+                        filename: str, 
+                        data: bytes, 
+                        doc_source_from: Optional[DocSourceFrom], 
+                        doc_file_type: Optional[DocFileType], 
+                        tags: Optional[Dict[str, str]] = None,
+                        auth_tenant_id: Optional[str] = None, 
+                        auth_company_id: Optional[str] = None, 
+                        doc_id: Optional[str] = None) -> bool:
+        """
+        上传单据文件，自动拼接存储路径并打tag
+        """
         try:
+            
             object_key = self._build_doc_path(filename, auth_tenant_id, auth_company_id, doc_id)
-            print(f"tags is {tags}")
             # 将tags转换为Tags对象
             tags_obj = Tags()
             for k, v in tags.items():
                 tags_obj[k] = v
+
+            if doc_source_from:
+                tags_obj["doc_source_from"] = doc_source_from.value
+            if doc_file_type:
+                tags_obj["doc_file_type"] = doc_file_type.value
             
             self.client.put_object(
                 bucket_name=self.bucket_name,

fiuai_s3/object_storage.py

@@ -10,7 +10,7 @@ from typing import Optional, List, Dict, Any
 from pydantic import BaseModel, Field, ConfigDict
 import logging
 from uuid import uuid4
-from .minio.type import DocFileObject
+from .type import DocFileObject, DocSourceFrom, DocFileType
 
 logger = logging.getLogger(__name__)
 
@@ -97,7 +97,13 @@ class ObjectStorage(ABC):
         pass
 
     @abstractmethod
-    def upload_doc_file(self, filename: str, data: bytes, tags: Optional[Dict[str, Any]] = None, auth_tenant_id: Optional[str] = None, auth_company_id: Optional[str] = None, doc_id: Optional[str] = None) -> bool:
+    def upload_doc_file(self, 
+                        filename: str, 
+                        data: bytes, 
+                        doc_source_from: Optional[DocSourceFrom], 
+                        doc_file_type: Optional[DocFileType], 
+                        tags: Optional[Dict[str, str]] = None,
+                        auth_tenant_id: Optional[str] = None, auth_company_id: Optional[str] = None, doc_id: Optional[str] = None) -> bool:
         """
         上传单据文件，自动拼接存储路径并打tag
         Args:

fiuai_s3/type.py

@@ -11,7 +11,7 @@ from typing import Optional, Dict
 
 class DocFileType(Enum):
     """
-    文档文件类型
+    文档文件类型, 避免某些场景不规范的文件名
     """
     PDF = "pdf"
     OFD = "ofd"
@@ -22,6 +22,18 @@ class DocFileType(Enum):
     IMAGE = "image"
     AUDIO = "audio"
     ARCHIVE = "archive"
+    OTHER = "other"
+
+class DocSourceFrom(Enum):
+    """
+    文档来源, 避免某些场景同类文件混淆,比如有2个xml文件,一个用户上传,一个ai生成
+    """
+    USER = "user"
+    AI = "ai"
+    INTEGRATION = "integration"
+    ETAX = "etax"
+    BANK = "bank"
+    OTHER = "other"
 
 class DocFileObject(BaseModel):
     """

{fiuai_s3-0.3.0.dist-info → fiuai_s3-0.3.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: fiuai-s3
-Version: 0.3.0
+Version: 0.3.2
 Summary: 一个支持阿里云OSS和MinIO的对象存储抽象包
 Project-URL: Homepage, https://github.com/fiuai-sz/fiuai-s3
 Project-URL: Repository, https://github.com/fiuai-sz/fiuai-s3.git
@@ -322,6 +322,52 @@ S3_Client.upload_doc_file(
 )
 ```
 
+为了规范文件, 文档类型的文件上传请按规范进行打标  
+```python
+class DocFileType(Enum):
+    """
+    文档文件类型, 避免某些场景不规范的文件名
+    """
+    PDF = "pdf"
+    OFD = "ofd"
+    XML = "xml"
+    DOCTYPE = "doctype"
+    TABLE = "table"
+    TEXT = "text"
+    IMAGE = "image"
+    AUDIO = "audio"
+    ARCHIVE = "archive"
+    OTHER = "other"
+```
+```python
+class DocSourceFrom(Enum):
+    """
+    文档来源, 避免某些场景同类文件混淆,比如有2个xml文件,一个用户上传,一个ai生成
+    """
+    USER = "user"
+    AI = "ai"
+    INTEGRATION = "integration"
+    ETAX = "etax"
+    BANK = "bank"
+    OTHER = "other"
+```
+
+上传时标注  
+```python
+c = ObjectStorageFactory.get_instance()
+    c.upload_doc_file(
+        filename="test.txt",
+        data=b"test",
+        doc_source_from=DocSourceFrom.USER,
+        doc_file_type=DocFileType.XML,
+        tags={"name": "hilton", "age": "10"},
+        auth_tenant_id="t1",
+        auth_company_id="c1",
+        doc_id="id111",
+    )
+
+```
+
 #### 下载单据文件
 ```python
 data = S3_Client.download_doc_file(
@@ -376,4 +422,8 @@ MIT License
 
 ## 贡献
 
-欢迎提交 Issue 和 Pull Request！
+欢迎提交 Issue 和 Pull Request！
+
+
+# 发布
+uv publish --username __token__ --password your-pypi-token-here

fiuai_s3-0.3.2.dist-info/RECORD

@@ -0,0 +1,11 @@
+fiuai_s3/__init__.py,sha256=PhRotxS_rfqOgdDkci63ToVqhtDi4uSC3-3aJTNebWo,421
+fiuai_s3/object_storage.py,sha256=idhKF4BFoP_ZPHY3PzlzHxVecrv1EnnBYyJ3QjGHiDE,8812
+fiuai_s3/type.py,sha256=-UBL0gVyMDC-4WRKNKyAupAE7s04TCOiFbM-YQPxh_Y,945
+fiuai_s3/alicloud/__init__.py,sha256=mmrCrFp5DzRF5fViJDq7_LpsqCViwTPXOPz4qoaSscI,218
+fiuai_s3/alicloud/alicloud_storage.py,sha256=d24AFYDs7RsTsfL9IiVOxYlGyLzsaUa_WUYD5VPLP4s,6035
+fiuai_s3/minio/__init__.py,sha256=hOmpUkTq8TPgYOxfeOMcWq1_YsJ2r8Zf9VTX3w_v1Lk,209
+fiuai_s3/minio/minio_storage.py,sha256=xsEt70WEo5X7RRLjIno4uXsoGbuATvwtv_wnQ2QSFBY,9021
+fiuai_s3-0.3.2.dist-info/METADATA,sha256=ltEia0FzulOArPzwYgy2_d4CfsM0_eCD2hM9W02HiRA,18727
+fiuai_s3-0.3.2.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+fiuai_s3-0.3.2.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+fiuai_s3-0.3.2.dist-info/RECORD,,

fiuai_s3-0.3.0.dist-info/RECORD

@@ -1,11 +0,0 @@
-fiuai_s3/__init__.py,sha256=SZNl4Ohwf19PLyjRjIASpzhg3rvutd1WbLpblz_8G3s,312
-fiuai_s3/object_storage.py,sha256=K_g2i3ohbT3BIJW-ZB5IbwYHtPuQMF4ZTwEoKR3WKr4,8561
-fiuai_s3/type.py,sha256=3LdjA9wO9IKKfKONLuI0SQ9FVSablhDnRphJqpDosRQ,608
-fiuai_s3/alicloud/__init__.py,sha256=mmrCrFp5DzRF5fViJDq7_LpsqCViwTPXOPz4qoaSscI,218
-fiuai_s3/alicloud/alicloud_storage.py,sha256=d24AFYDs7RsTsfL9IiVOxYlGyLzsaUa_WUYD5VPLP4s,6035
-fiuai_s3/minio/__init__.py,sha256=hOmpUkTq8TPgYOxfeOMcWq1_YsJ2r8Zf9VTX3w_v1Lk,209
-fiuai_s3/minio/minio_storage.py,sha256=SU743Nr7w8HMran-xkccEoeecSNKYySVNCGjo462-Lk,8476
-fiuai_s3-0.3.0.dist-info/METADATA,sha256=WEEnAxbTp6ogOVIE8r41u7WwwYcAZ3KcEW0iq4FDUr4,17609
-fiuai_s3-0.3.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-fiuai_s3-0.3.0.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-fiuai_s3-0.3.0.dist-info/RECORD,,