intellif-aihub 0.1.14__tar.gz → 0.1.15__tar.gz

{intellif_aihub-0.1.14/src/intellif_aihub.egg-info → intellif_aihub-0.1.15}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: intellif-aihub
-Version: 0.1.14
+Version: 0.1.15
 Summary: Intellif AI-hub SDK.
 Author-email: Platform Team <aihub@example.com>
 License-Expression: Apache-2.0

{intellif_aihub-0.1.14 → intellif_aihub-0.1.15}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name            = "intellif-aihub"
-version         = "0.1.14"
+version = "0.1.15"
 description     = "Intellif AI-hub SDK."
 readme          = {file = "README.md", content-type = "text/markdown"}
 requires-python = ">=3.9"

intellif_aihub-0.1.15/src/aihub/__init__.py

@@ -0,0 +1 @@
+__version__ = "0.1.15"

{intellif_aihub-0.1.14 → intellif_aihub-0.1.15}/src/aihub/models/artifact.py

@@ -14,6 +14,7 @@ from pydantic import BaseModel, Field
 
 class ArtifactType(str, Enum):
     """制品类型枚举："dataset"-数据集类型；"model"-模型类型；"metrics"-指标类型；"log"-日志类型；"checkpoint"-检查点类型；"image"-图像类型；"prediction"-预测结果类型；"other"-其他类型"""
+
     dataset = "dataset"  # 数据集类型
     model = "model"  # 模型类型
     metrics = "metrics"  # 指标类型
@@ -26,24 +27,29 @@ class ArtifactType(str, Enum):
 
 class CreateArtifactsReq(BaseModel):
     """创建制品请求"""
+
     entity_id: str = Field(alias="entity_id", description="实体ID，通常是运行ID，用于关联制品与特定运行")
-    entity_type: ArtifactType = Field(default=ArtifactType.other, alias="entity_type",
-                                      description="制品类型，指定制品的类型，默认为other")
+    entity_type: ArtifactType = Field(
+        default=ArtifactType.other, alias="entity_type", description="制品类型，指定制品的类型，默认为other"
+    )
     src_path: str = Field(alias="src_path", description="源路径，制品在系统中的路径标识")
-    is_dir: bool = Field(default=False, alias="is_dir",
-                         description="是否为目录，True表示制品是一个目录，False表示是单个文件")
+    is_dir: bool = Field(
+        default=False, alias="is_dir", description="是否为目录，True表示制品是一个目录，False表示是单个文件"
+    )
 
     model_config = {"use_enum_values": True}
 
 
 class CreateArtifactsResponseData(BaseModel):
     """创建制品响应数据"""
+
     id: int = Field(description="制品ID")
     s3_path: str = Field(alias="s3_path", description="S3存储路径")
 
 
 class CreateArtifactsResponseModel(BaseModel):
     """创建制品响应模型"""
+
     code: int = Field(description="响应码，0表示成功")
     msg: str = Field(default="", description="响应消息")
     data: Optional[CreateArtifactsResponseData] = Field(default=None, description="响应数据")
@@ -51,6 +57,7 @@ class CreateArtifactsResponseModel(BaseModel):
 
 class CreateEvalReq(BaseModel):
     """创建评估请求"""
+
     dataset_id: int = Field(alias="dataset_id", description="数据集ID")
     dataset_version_id: int = Field(alias="dataset_version_id", description="数据集版本ID")
     prediction_artifact_path: str = Field(alias="prediction_artifact_path", description="预测结果制品路径")
@@ -62,6 +69,7 @@ class CreateEvalReq(BaseModel):
 
 class ArtifactResp(BaseModel):
     """制品响应模型，表示一个制品的详细信息"""
+
     id: int = Field(description="制品ID")
     entity_type: str = Field(alias="entity_type", description="实体类型")
     entity_id: str = Field(alias="entity_id", description="实体ID")
@@ -72,6 +80,7 @@ class ArtifactResp(BaseModel):
 
 class ArtifactRespData(BaseModel):
     """制品分页数据"""
+
     total: int = Field(description="总记录数")
     page_size: int = Field(alias="page_size", description="每页大小")
     page_num: int = Field(alias="page_num", description="页码")
@@ -80,6 +89,7 @@ class ArtifactRespData(BaseModel):
 
 class ArtifactRespModel(BaseModel):
     """获取制品响应模型"""
+
     code: int = Field(description="响应码，0表示成功")
     msg: str = Field(default="", description="响应消息")
     data: ArtifactRespData = Field(description="响应数据")
@@ -91,8 +101,10 @@ InfinityPageSize = 10000 * 100
 
 class StsResp(BaseModel):
     """STS 临时凭证"""
+
     access_key_id: Optional[str] = Field(default=None, alias="access_key_id", description="访问密钥ID")
     secret_access_key: Optional[str] = Field(default=None, alias="secret_access_key", description="秘密访问密钥")
     session_token: Optional[str] = Field(default=None, alias="session_token", description="会话令牌")
     expiration: Optional[int] = Field(default=None, alias="expiration", description="过期时间")
     endpoint: Optional[str] = Field(default=None, alias="endpoint", description="端点URL")
+    bucket: Optional[str] = Field(default=None, alias="bucket", description="存储桶名称")

{intellif_aihub-0.1.14 → intellif_aihub-0.1.15}/src/aihub/models/dataset_management.py

@@ -8,6 +8,7 @@ from pydantic import BaseModel, Field
 
 class DatasetVersionStatus(IntEnum):
     """数据集版本状态：1-等待中；2-运行中；3-成功；4-失败；5-加载meta；6-构建index"""
+
     Waiting = 1  # 等待中
     Running = 2  # 运行中
     Success = 3  # 成功
@@ -18,6 +19,7 @@ class DatasetVersionStatus(IntEnum):
 
 class UploadType(IntEnum):
     """上传类型：1-本地上传；3-服务器路径上传；4-Labelfree；5-数据接入"""
+
     LOCAL = 1  # 本地上传
     SERVER_PATH = 3  # 服务器路径上传
     LABELFREE = 4  # Labelfree
@@ -26,6 +28,7 @@ class UploadType(IntEnum):
 
 class CreateDatasetRequest(BaseModel):
     """创建数据集请求"""
+
     name: str = Field(description="数据集名称")
     description: str = Field(description="数据集描述")
     tags: List[int] = Field(description="标签ID列表，通过标签管理系统查询")
@@ -37,11 +40,13 @@ class CreateDatasetRequest(BaseModel):
 
 class CreateDatasetResponse(BaseModel):
     """创建数据集返回"""
+
     id: int = Field(alias="id", description="数据集ID")
 
 
 class DatasetVersionBase(BaseModel):
     """数据集版本概要"""
+
     id: int = Field(description="版本ID")
     version: int = Field(description="版本号")
     status: DatasetVersionStatus = Field(description="版本状态")
@@ -53,6 +58,7 @@ class DatasetVersionBase(BaseModel):
 
 class DatasetDetail(BaseModel):
     """数据集详情"""
+
     id: int = Field(description="数据集 ID")
     name: str = Field(description="名称")
     description: str = Field(description="描述")
@@ -69,6 +75,7 @@ class DatasetDetail(BaseModel):
 
 class ExtInfo(BaseModel):
     """扩展信息"""
+
     rec_file_path: Optional[str] = Field(None, alias="rec_file_path", description="rec文件路径")
     idx_file_path: Optional[str] = Field(None, alias="idx_file_path", description="idx文件路径")
     json_file_path: Optional[str] = Field(None, alias="json_file_path", description="json文件路径")
@@ -77,6 +84,7 @@ class ExtInfo(BaseModel):
 
 class CreateDatasetVersionRequest(BaseModel):
     """创建版本请求"""
+
     upload_path: str = Field(alias="upload_path", description="上传路径")
     description: Optional[str] = Field(None, description="版本描述")
     dataset_id: int = Field(alias="dataset_id", description="数据集ID")
@@ -91,11 +99,13 @@ class CreateDatasetVersionRequest(BaseModel):
 
 class CreateDatasetVersionResponse(BaseModel):
     """创建版本返回"""
+
     id: int = Field(alias="id", description="版本ID")
 
 
 class UploadDatasetVersionRequest(BaseModel):
     """上传数据集版本请求"""
+
     upload_path: str = Field(alias="upload_path", description="上传目录")
     upload_type: UploadType = Field(alias="upload_type", description="上传类型")
     dataset_id: int = Field(alias="dataset_id", description="数据集ID")
@@ -107,11 +117,13 @@ class UploadDatasetVersionRequest(BaseModel):
 
 class UploadDatasetVersionResponse(BaseModel):
     """上传数据集版本返回"""
+
     id: int = Field(alias="id", description="版本ID")
 
 
 class DatasetVersionDetail(BaseModel):
     """数据集版本详情"""
+
     id: int = Field(description="版本ID")
     version: int = Field(description="版本号")
     dataset_id: int = Field(alias="dataset_id", description="数据集ID")
@@ -133,6 +145,7 @@ class DatasetVersionDetail(BaseModel):
 
 class FileUploadData(BaseModel):
     """文件上传数据"""
+
     path: str = Field(description="路径")
     url: str = Field(description="URL")
 
@@ -203,3 +216,13 @@ class ListDatasetVersionResp(BaseModel):
     page_size: int = Field(alias="page_size", description="每页大小")
     page_num: int = Field(alias="page_num", description="当前页码")
     data: List[ListDatasetVersionItem] = Field(description="数据集版本列表")
+
+
+class CreateDatasetVersionByDataIngestReqV2(BaseModel):
+    """通过数据集成创建数据集版本请求"""
+
+    description: Optional[str] = Field(None, description="描述")
+    dataset_id: int = Field(..., description="数据集ID")
+    s3_object_sheet: str = Field(..., description="S3对象表")
+    object_cnt: Optional[int] = Field(None, description="对象数量")
+    data_size: Optional[int] = Field(None, description="数据大小")

{intellif_aihub-0.1.14 → intellif_aihub-0.1.15}/src/aihub/services/artifact.py

@@ -98,9 +98,7 @@ class ArtifactService:
         """
         return self._artifact.get_sts()
 
-    def get_by_run_id(
-            self, run_id: str, artifact_path: Optional[str] = None
-    ) -> List[ArtifactResp]:
+    def get_by_run_id(self, run_id: str, artifact_path: Optional[str] = None) -> List[ArtifactResp]:
         """根据运行ID获取制品列表
 
         Args:
@@ -116,11 +114,11 @@ class ArtifactService:
         return self._artifact.get_by_run_id(run_id, artifact_path)
 
     def create_artifact(
-            self,
-            local_path: str,
-            artifact_path: Optional[str] = None,
-            run_id: Optional[str] = None,
-            artifact_type: ArtifactType = ArtifactType.other,
+        self,
+        local_path: str,
+        artifact_path: Optional[str] = None,
+        run_id: Optional[str] = None,
+        artifact_type: ArtifactType = ArtifactType.other,
     ) -> None:
         """创建单个文件制品并上传
 
@@ -171,11 +169,11 @@ class ArtifactService:
         return
 
     def create_artifacts(
-            self,
-            local_dir: str,
-            artifact_path: Optional[str] = None,
-            run_id: Optional[str] = None,
-            artifact_type: ArtifactType = ArtifactType.other,
+        self,
+        local_dir: str,
+        artifact_path: Optional[str] = None,
+        run_id: Optional[str] = None,
+        artifact_type: ArtifactType = ArtifactType.other,
     ) -> None:
         """创建目录制品并上传
 
@@ -223,9 +221,7 @@ class ArtifactService:
         logger.info(f"log artifact done: {artifact_path}")
         return
 
-    def download_artifacts(
-            self, run_id: str, artifact_path: Optional[str], local_dir: str
-    ) -> None:
+    def download_artifacts(self, run_id: str, artifact_path: Optional[str], local_dir: str) -> None:
         """下载制品
 
         Args:
@@ -252,9 +248,7 @@ class ArtifactService:
             if artifact_item.is_dir:
                 download_dir_from_s3(self.s3_client, bucket, object_name, local_dir)
             else:
-                self.s3_client.fget_object(
-                    bucket, object_name, str(Path(local_dir) / artifact_item.src_path)
-                )
+                self.s3_client.fget_object(bucket, object_name, str(Path(local_dir) / artifact_item.src_path))
 
         logger.info(f"download artifact done: {artifact_path}")
         return
@@ -311,9 +305,7 @@ class _Artifact:
             raise APIError(f"backend code {wrapper.code}: {wrapper.msg}")
         return
 
-    def get_by_run_id(
-            self, run_id: str, artifact_path: Optional[str]
-    ) -> List[ArtifactResp]:
+    def get_by_run_id(self, run_id: str, artifact_path: Optional[str]) -> List[ArtifactResp]:
         """根据运行ID获取制品列表
 
         Args:
@@ -326,18 +318,12 @@ class _Artifact:
         Raises:
             APIError: 当API调用失败时抛出
         """
-        resp = self._http.get(
-            f"{_Base}/artifacts?entity_id={run_id}&page_num=1&page_size={InfinityPageSize}"
-        )
+        resp = self._http.get(f"{_Base}/artifacts?entity_id={run_id}&page_num=1&page_size={InfinityPageSize}")
         wrapper = APIWrapper[ArtifactRespData].model_validate(resp.json())
         if wrapper.code != 0:
             raise APIError(f"backend code {wrapper.code}: {wrapper.msg}")
         if artifact_path:
-            return [
-                artifact
-                for artifact in wrapper.data.data
-                if artifact.src_path == artifact_path
-            ]
+            return [artifact for artifact in wrapper.data.data if artifact.src_path == artifact_path]
         else:
             return wrapper.data.data
 

{intellif_aihub-0.1.14 → intellif_aihub-0.1.15}/src/aihub/services/dataset_management.py

@@ -17,7 +17,6 @@ from __future__ import annotations
 import mimetypes
 import os
 import pathlib
-import tempfile
 import time
 import uuid
 
@@ -25,6 +24,7 @@ import httpx
 from loguru import logger
 
 from ..exceptions import APIError
+from ..models.artifact import StsResp
 from ..models.common import APIWrapper
 from ..models.dataset_management import (
     CreateDatasetRequest,
@@ -40,9 +40,12 @@ from ..models.dataset_management import (
     ListDatasetResp,
     ListDatasetVersionReq,
     ListDatasetVersionResp,
+    CreateDatasetVersionByDataIngestReqV2,
+    UploadType,
 )
 from ..models.dataset_management import DatasetVersionStatus
-from ..utils.download import dataset_download, zip_dir
+from ..utils.di import SimpleS3Client, DataUploader
+from ..utils.download import dataset_download
 
 _BASE = "/dataset-mng/api/v2"
 
@@ -138,20 +141,29 @@ class DatasetManagementService:
     def dataset(self) -> _Dataset:
         return self._dataset
 
+    def _get_sts(self) -> StsResp:
+        return self.dataset_version.get_sts()
+
     @property
     def dataset_version(self) -> _DatasetVersion:
         return self._dataset_version
 
+    def upload_by_data_ingest(
+        self,
+        req: CreateDatasetVersionByDataIngestReqV2,
+    ) -> CreateDatasetVersionResponse:
+        return self.dataset_version.upload_by_data_ingest(req)
+
     def create_dataset_and_version(
-            self,
-            *,
-            dataset_name: str,
-            dataset_description: str = "",
-            is_local_upload: bool,
-            local_file_path: str | None = None,
-            server_file_path: str | None = None,
-            version_description: str = "",
-            timeout: int = 1_800,
+        self,
+        *,
+        dataset_name: str,
+        dataset_description: str = "",
+        is_local_upload: bool,
+        local_file_path: str | None = None,
+        server_file_path: str | None = None,
+        version_description: str = "",
+        timeout: int = 1_800,
     ) -> tuple[int, int, str]:
         """创建数据集及其版本，并等待版本状态变为 *Success*。
 
@@ -169,17 +181,51 @@ class DatasetManagementService:
 
         Returns:
            tuple[int, int, str]: 一个三元组，包含：[数据集 ID,数据集版本 ID, 数据集版本标签（格式为 <dataset_name>/V<version_number>)]
+
+        Raises:
+            ValueError: 当参数不满足要求时
+            APIError: 当后端返回错误时
+            TimeoutError: 当等待超时时
         """
+        # 参数校验
+        self._validate_create_params(is_local_upload, local_file_path, server_file_path)
+
+        # 创建数据集
+        dataset_id = self._create_dataset(dataset_name, dataset_description)
+        logger.info(f"创建数据集成功，名称为 {dataset_name} ,开始准备创建版本、上传数据")
+
+        # 创建数据集版本
+        version_id = self._create_dataset_version(
+            dataset_id=dataset_id,
+            is_local_upload=is_local_upload,
+            local_file_path=local_file_path,
+            server_file_path=server_file_path,
+            version_description=version_description,
+        )
+
+        # 获取版本标签
+        version_tag = self._get_version_tag(dataset_id, version_id)
+        logger.info(f"数据集版本创建成功，名称为 {version_tag}，开始轮询状态…")
+
+        # 轮询等待版本状态变为成功
+        self._wait_for_version_success(version_id, timeout)
+
+        return dataset_id, version_id, version_tag
+
+    def _validate_create_params(
+        self, is_local_upload: bool, local_file_path: str | None, server_file_path: str | None
+    ) -> None:
+        """验证创建数据集和版本所需的参数"""
         if is_local_upload:
             if not local_file_path:
                 raise ValueError("is_local_upload=True 时必须提供 local_file_path")
-            upload_type = 1
         else:
             if not server_file_path:
                 raise ValueError("is_local_upload=False 时必须提供 server_file_path")
-            upload_type = 3
 
-        dataset_id = self._dataset.create(
+    def _create_dataset(self, dataset_name: str, dataset_description: str) -> int:
+        """创建数据集"""
+        return self._dataset.create(
             CreateDatasetRequest(
                 name=dataset_name,
                 description=dataset_description,
@@ -190,39 +236,96 @@ class DatasetManagementService:
                 access_user_ids=None,
             )
         )
-        logger.info(
-            f"创建数据集成功，名称为 {dataset_name} ,开始准备创建版本、上传数据"
-        )
 
+    def _create_dataset_version(
+        self,
+        dataset_id: int,
+        is_local_upload: bool,
+        local_file_path: str | None,
+        server_file_path: str | None,
+        version_description: str,
+    ) -> int:
+        """根据上传类型创建数据集版本"""
         if is_local_upload:
-            # 上传文件，检查是文件夹还是zip
-            local_file_path = pathlib.Path(local_file_path)
-            if local_file_path.is_dir():
-                # 把文件夹打包为一个 zip
-                temp_zip_path = (
-                        pathlib.Path(tempfile.mkdtemp()) / f" {uuid.uuid4().hex}.zip"
-                )
-                zip_dir(local_file_path, temp_zip_path)
-                upload_data = self._upload.upload_file(temp_zip_path)
-                os.remove(temp_zip_path)
-            else:
-                upload_data = self._upload.upload_file(local_file_path)
-
-            upload_path = upload_data.path
+            return self._create_local_dataset_version(dataset_id, local_file_path, version_description)
         else:
-            upload_path = server_file_path
-        logger.info(f"文件上传成功：{local_file_path}")
+            return self._create_server_dataset_version(dataset_id, server_file_path, version_description)
+
+    def _create_local_dataset_version(
+        self, dataset_id: int, local_file_path: str | None, version_description: str
+    ) -> int:
+        """创建本地文件数据集版本"""
+        if pathlib.Path(local_file_path).is_dir():
+            return self._create_local_dir_dataset_version(dataset_id, local_file_path)
+        elif pathlib.Path(local_file_path).is_file():
+            return self._create_local_file_dataset_version(dataset_id, local_file_path, version_description)
+        else:
+            raise ValueError(f"本地路径既不是文件也不是目录: {local_file_path}")
+
+    def _create_local_dir_dataset_version(self, dataset_id: int, local_file_path: str) -> int:
+        """处理本地目录上传"""
+        sts = self._get_sts()
+        s3_client = SimpleS3Client(
+            sts.endpoint, sts.access_key_id, sts.secret_access_key, session_token=sts.session_token
+        )
+        uid = uuid.uuid4().hex
+        s3_target = f"s3://{sts.bucket}/dataset_workspace/{dataset_id}/{uid}"
+        s3_csv_path = f"s3://{sts.bucket}/dataset_workspace/{dataset_id}/{uid}.csv"
+        s3_status_path = f"s3://{sts.bucket}/dataset_workspace/{dataset_id}/{uid}.json"
+
+        # 创建上传器并执行
+        uploader = DataUploader(
+            task_id=dataset_id,
+            local_path=str(local_file_path),
+            s3_target=s3_target,
+            csv_path=s3_csv_path,
+            status_path=s3_status_path,
+            num_workers=40,
+        )
 
-        version_id = self._dataset_version.upload(
+        upload_stats = uploader.run(s3_client)
+        req = CreateDatasetVersionByDataIngestReqV2(
+            description=f"sdk 上传",
+            dataset_id=dataset_id,
+            s3_object_sheet=s3_csv_path,
+            object_cnt=upload_stats.uploaded_count,
+            data_size=upload_stats.uploaded_size,
+        )
+        return self.upload_by_data_ingest(req).id
+
+    def _create_local_file_dataset_version(
+        self, dataset_id: int, local_file_path: str, version_description: str
+    ) -> int:
+        """处理本地文件上传"""
+        upload_data = self._upload.upload_file(local_file_path)
+        upload_path = upload_data.path
+        logger.info(f"文件上传成功：{local_file_path}")
+        return self._dataset_version.upload(
             UploadDatasetVersionRequest(
                 upload_path=upload_path,
-                upload_type=upload_type,
+                upload_type=UploadType.LOCAL,  # 本地上传类型
+                dataset_id=dataset_id,
+                description=version_description,
+                parent_version_id=0,
+            )
+        )
+
+    def _create_server_dataset_version(
+        self, dataset_id: int, server_file_path: str | None, version_description: str
+    ) -> int:
+        """创建服务器文件数据集版本"""
+        return self._dataset_version.upload(
+            UploadDatasetVersionRequest(
+                upload_path=server_file_path,
+                upload_type=UploadType.SERVER_PATH,  # 服务器文件上传类型
                 dataset_id=dataset_id,
                 description=version_description,
                 parent_version_id=0,
             )
         )
 
+    def _get_version_tag(self, dataset_id: int, version_id: int) -> str:
+        """获取版本标签"""
         detail = self._dataset.get(dataset_id)
         ver_num = next(
             (v.version for v in detail.versions if v.id == version_id),
@@ -231,9 +334,10 @@ class DatasetManagementService:
         if ver_num is None:
             ver_num = 1
 
-        version_tag = f"{detail.name}/V{ver_num}"
-        logger.info(f"数据集版本创建成功，名称为 {version_tag}，开始轮询状态…")
+        return f"{detail.name}/V{ver_num}"
 
+    def _wait_for_version_success(self, version_id: int, timeout: int) -> None:
+        """轮询等待版本状态变为成功"""
         start_ts = time.time()
         poll_interval = 10
 
@@ -255,8 +359,6 @@ class DatasetManagementService:
             logger.debug(f"已等待 {elapsed:.0f}s，继续轮询…")
             time.sleep(poll_interval)
 
-        return dataset_id, version_id, version_tag
-
     def run_download(self, dataset_version_name: str, local_dir: str, worker: int = 4) -> None:
         """根据数据集版本名称下载对应的数据集文件。
 
@@ -400,9 +502,7 @@ class _DatasetVersion:
         return wrapper.data
 
     def get_by_name(self, version_name: str) -> DatasetVersionDetail:
-        resp = self._http.get(
-            f"{_BASE}/dataset-versions-detail", params={"name": version_name}
-        )
+        resp = self._http.get(f"{_BASE}/dataset-versions-detail", params={"name": version_name})
         wrapper = APIWrapper[DatasetVersionDetail].model_validate(resp.json())
         if wrapper.code != 0:
             raise APIError(f"backend code {wrapper.code}: {wrapper.msg}")
@@ -417,6 +517,40 @@ class _DatasetVersion:
             raise APIError(f"backend code {wrapper.code}: {wrapper.msg}")
         return wrapper.data
 
+    def get_sts(self) -> StsResp:
+        """获取STS临时凭证
+
+        获取用于访问S3存储的临时凭证。
+
+        Returns:
+            StsResp: STS临时凭证信息
+
+        Raises:
+            APIError: 当API调用失败时抛出
+        """
+        resp = self._http.get(f"{_BASE}/dataset-versions/get-sts")
+        logger.info(f"get sts: {resp.text}")
+        wrapper = APIWrapper[StsResp].model_validate(resp.json())
+        if wrapper.code != 0:
+            raise APIError(f"backend code {wrapper.code}: {wrapper.msg}")
+        return wrapper.data
+
+    def upload_by_data_ingest(self, req: CreateDatasetVersionByDataIngestReqV2) -> CreateDatasetVersionResponse:
+        """上传数据集版本（数据集导入）
+        Args:
+            req
+
+        """
+        resp = self._http.post(
+            f"{_BASE}/dataset-versions/data-ingest",
+            json=req.model_dump(),
+        )
+        logger.debug(f"upload_by_data_ingest: {resp.text}")
+        wrapper = APIWrapper[CreateDatasetVersionResponse].model_validate(resp.json())
+        if wrapper.code != 0:
+            raise APIError(f"backend code {wrapper.code}: {wrapper.msg}")
+        return wrapper.data
+
 
 class _Upload:
     def __init__(self, http: httpx.Client):