intellif-aihub 0.1.19__tar.gz → 0.1.21__tar.gz

{intellif_aihub-0.1.19 → intellif_aihub-0.1.21}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: intellif-aihub
-Version: 0.1.19
+Version: 0.1.21
 Summary: Intellif AI-hub SDK.
 Author-email: Platform Team <aihub@example.com>
 License-Expression: Apache-2.0

{intellif_aihub-0.1.19 → intellif_aihub-0.1.21}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name            = "intellif-aihub"
-version = "0.1.19"
+version = "0.1.21"
 description     = "Intellif AI-hub SDK."
 readme          = {file = "README.md", content-type = "text/markdown"}
 requires-python = ">=3.9"

intellif_aihub-0.1.21/src/aihub/__init__.py

@@ -0,0 +1 @@
+__version__ = "0.1.21"

intellif_aihub-0.1.21/src/aihub/models/eval.py

@@ -0,0 +1,81 @@
+# !/usr/bin/env python
+# -*-coding:utf-8 -*-
+from typing import Dict, List, Optional
+
+from pydantic import BaseModel, Field
+
+
+class BaseEvalReq(BaseModel):
+    """评测任务基础请求模型"""
+    run_id: str = Field(description="运行ID")
+    type: str = Field(description="评测类型，支持 'llm' 和 'cv'")
+    prediction_artifact_path: str = Field(description="推理产物的路径")
+    user_id: int = Field(0, description="用户ID，默认0")
+
+
+class CreateLLMEvalReq(BaseEvalReq):
+    """创建LLM类型评测任务请求"""
+    type: str = Field(default="llm", description="评测类型，固定为 'llm'")
+    dataset_id: int = Field(description="数据集ID")
+    dataset_version_id: int = Field(description="数据集版本ID")
+    evaled_artifact_path: str = Field(description="评测结果产物的路径")
+    report: Dict = Field(description="评测报告")
+
+
+class CreateCVEvalReq(BaseEvalReq):
+    """创建CV类型评测任务请求"""
+    type: str = Field(default="cv", description="评测类型，固定为 'cv'")
+    metrics_artifact_path: str = Field(description="指标产物的路径")
+    ground_truth_artifact_path: str = Field(description="真实标签产物的路径")
+
+
+class EvalRun(BaseModel):
+    """评测任务的运行实体"""
+    id: int = Field(description="评测的运行ID")
+    name: str = Field(description="评测名称")
+    description: str = Field(description="评测描述")
+    user_id: int = Field(description="用户ID")
+    model_id: int = Field(description="模型ID")
+    model_name: str = Field(description="模型名称")
+    dataset_id: int = Field(description="数据集ID")
+    dataset_version_id: int = Field(description="数据集版本ID")
+    dataset_name: str = Field(description="数据集名称")
+    status: str = Field(description="状态")
+    prediction_artifact_path: str = Field(description="推理产物路径")
+    evaled_artifact_path: str = Field(description="评测结果产物路径")
+    run_id: str = Field(description="运行ID")
+    dataset_summary: Dict = Field(default_factory=dict, description="数据集摘要")
+    metrics_summary: Dict = Field(default_factory=dict, description="指标摘要")
+    viz_summary: Dict = Field(default_factory=dict, description="可视化摘要")
+    eval_config: Optional[Dict] = Field(default=None, description="评测配置")
+    created_at: int = Field(description="创建时间")
+    updated_at: int = Field(description="更新时间")
+
+
+class CreateEvalResp(BaseModel):
+    """创建评测任务的返回结果"""
+    eval_run: EvalRun = Field(alias="eval_run", description="评测运行信息")
+
+
+class ListEvalReq(BaseModel):
+    """列出评测任务请求"""
+    page_size: int = Field(20, description="页面大小")
+    page_num: int = Field(1, description="页码")
+    status: Optional[str] = Field(None, description="状态过滤")
+    name: Optional[str] = Field(None, description="名称过滤")
+    model_id: Optional[int] = Field(None, description="模型ID过滤")
+    dataset_id: Optional[int] = Field(None, description="数据集ID过滤")
+    dataset_version_id: Optional[int] = Field(None, description="数据集版本ID过滤")
+    run_id: Optional[str] = Field(None, description="运行ID过滤")
+    user_id: Optional[int] = Field(None, description="用户ID过滤")
+    model_ids: Optional[str] = Field(None, description="模型ID列表过滤")
+    dataset_ids: Optional[str] = Field(None, description="数据集ID列表过滤")
+    dataset_version_ids: Optional[str] = Field(None, description="数据集版本ID列表过滤")
+
+
+class ListEvalResp(BaseModel):
+    """列出评测任务响应"""
+    total: int = Field(description="总数")
+    page_size: int = Field(description="页面大小")
+    page_num: int = Field(description="页码")
+    data: List[EvalRun] = Field(description="评测运行列表")

intellif_aihub-0.1.21/src/aihub/services/eval.py

@@ -0,0 +1,172 @@
+# !/usr/bin/env python
+# -*- coding:utf-8 -*-
+"""评测平台服务模块
+
+本模块围绕 **“模型评测（Run → Report）”** 提供能力：
+
+- **创建评测任务 / 评测报告**
+- **获取评测任务列表**
+"""
+
+import httpx
+
+from ..exceptions import APIError
+from ..models.common import APIWrapper
+from ..models.eval import CreateLLMEvalReq, CreateCVEvalReq, CreateEvalResp, ListEvalReq, ListEvalResp
+
+_BASE = "/eval-platform/api/v1"
+
+
+class EvalService:
+    """评测服务"""
+
+    def __init__(self, http: httpx.Client):
+        self._http = http
+        self._eval = _Eval(http)
+
+    def create(
+        self,
+        dataset_version_name: str,
+        prediction_artifact_path: str,
+        evaled_artifact_path: str,
+        report_json: dict,
+        run_id,
+        user_id: int = 0,
+    ) -> int:
+        """创建评测报告
+
+        Args:
+            run_id (str): RUN ID
+            report_json (dict): 报告内容
+            evaled_artifact_path:   评测结果制品路径
+            prediction_artifact_path: 推理结果制品路径
+            dataset_version_name (str): 数据集名称
+            user_id (int, optional): 用户ID，默认为0
+
+        Returns:
+            id (int): 评测报告id
+
+        """
+        from .dataset_management import DatasetManagementService
+
+        dataset_service = DatasetManagementService(self._http)
+        dataset_version = dataset_service.get_dataset_version_by_name(
+            dataset_version_name
+        )
+        payload = CreateLLMEvalReq(
+            dataset_id=dataset_version.dataset_id,
+            dataset_version_id=dataset_version.id,
+            evaled_artifact_path=evaled_artifact_path,
+            prediction_artifact_path=prediction_artifact_path,
+            report=report_json,
+            run_id=run_id,
+            user_id=user_id,
+        )
+
+        return self._eval.create(payload)
+
+    def create_cv_run(
+        self,
+        run_id: str,
+        prediction_artifact_path: str,
+        metrics_artifact_path: str,
+        ground_truth_artifact_path: str,
+        user_id: int = 0,
+    ) -> int:
+        """创建 CV 类型评测运行
+        
+        Args:
+            run_id (str): 运行ID
+            prediction_artifact_path (str): 推理产物的路径
+            metrics_artifact_path (str): 指标产物的路径
+            ground_truth_artifact_path (str): 真实标签产物的路径
+            user_id (int, optional): 用户ID，默认为0
+            
+        Returns:
+            id (int): 评测运行id
+        """
+        payload = CreateCVEvalReq(
+            run_id=run_id,
+            prediction_artifact_path=prediction_artifact_path,
+            metrics_artifact_path=metrics_artifact_path,
+            ground_truth_artifact_path=ground_truth_artifact_path,
+            user_id=user_id,
+        )
+        
+        return self._eval.create(payload)
+
+    def list(
+        self,
+        page_size: int = 20,
+        page_num: int = 1,
+        status: str = None,
+        name: str = None,
+        model_id: int = None,
+        dataset_id: int = None,
+        dataset_version_id: int = None,
+        run_id: str = None,
+        user_id: int = None,
+        model_ids: str = None,
+        dataset_ids: str = None,
+        dataset_version_ids: str = None,
+    ) -> ListEvalResp:
+        """列出评测结果
+
+        Args:
+            page_size (int): 页面大小，默认为20
+            page_num (int): 页码，默认为1
+            status (str, optional): 状态过滤
+            name (str, optional): 名称过滤
+            model_id (int, optional): 模型ID过滤
+            dataset_id (int, optional): 数据集ID过滤
+            dataset_version_id (int, optional): 数据集版本ID过滤
+            run_id (str, optional): 运行ID过滤
+            user_id (int, optional): 用户ID过滤
+            model_ids (str, optional): 模型ID列表过滤（逗号分隔）
+            dataset_ids (str, optional): 数据集ID列表过滤（逗号分隔）
+            dataset_version_ids (str, optional): 数据集版本ID列表过滤（逗号分隔）
+
+        Returns:
+            ListEvalResp: 评测结果列表响应
+        """
+        payload = ListEvalReq(
+            page_size=page_size,
+            page_num=page_num,
+            status=status,
+            name=name,
+            model_id=model_id,
+            dataset_id=dataset_id,
+            dataset_version_id=dataset_version_id,
+            run_id=run_id,
+            user_id=user_id,
+            model_ids=model_ids,
+            dataset_ids=dataset_ids,
+            dataset_version_ids=dataset_version_ids,
+        )
+
+        return self._eval.list(payload)
+
+
+class _Eval:
+    def __init__(self, http: httpx.Client):
+        self._http = http
+
+    def create(self, payload) -> int:
+        resp = self._http.post(f"{_BASE}/run/", json=payload.model_dump())
+        wrapper = APIWrapper[CreateEvalResp].model_validate(resp.json())
+        if wrapper.code != 0:
+            raise APIError(f"backend code {wrapper.code}: {wrapper.msg}")
+        return wrapper.data.eval_run.id
+
+    def list(self, payload: ListEvalReq) -> ListEvalResp:
+        # Build query parameters, excluding None values
+        params = {}
+        for field, value in payload.model_dump().items():
+            if value is not None:
+                params[field] = value
+
+        resp = self._http.get(f"{_BASE}/run/", params=params)
+        wrapper = APIWrapper[ListEvalResp].model_validate(resp.json())
+        if wrapper.code != 0:
+            raise APIError(f"backend code {wrapper.code}: {wrapper.msg}")
+        return wrapper.data

{intellif_aihub-0.1.19 → intellif_aihub-0.1.21}/src/intellif_aihub.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: intellif-aihub
-Version: 0.1.19
+Version: 0.1.21
 Summary: Intellif AI-hub SDK.
 Author-email: Platform Team <aihub@example.com>
 License-Expression: Apache-2.0

{intellif_aihub-0.1.19 → intellif_aihub-0.1.21}/src/intellif_aihub.egg-info/SOURCES.txt

@@ -49,6 +49,7 @@ tests/test_data_warehouse.py
 tests/test_dataset_management.py
 tests/test_di.py
 tests/test_document_center.py
+tests/test_eval.py
 tests/test_labelfree.py
 tests/test_model_center.py
 tests/test_model_training_platform.py

intellif_aihub-0.1.21/tests/test_eval.py

@@ -0,0 +1,520 @@
+# !/usr/bin/env python
+# -*-coding:utf-8 -*-
+import sys
+import os
+import unittest
+import uuid
+from unittest.mock import Mock, patch
+
+import httpx
+
+from aihub.services.eval import EvalService
+from aihub.models.eval import ListEvalResp, EvalRun, CreateEvalResp
+from aihub.models.common import APIWrapper
+
+BASE_URL = "http://192.168.13.160:30052"
+TOKEN = "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJleHAiOjE3NTI1NDQwNDksImlhdCI6MTc1MVkzOTI0OSwidWlkIjoyfQ.MfB_7LK5oR3RAhga3jtgcvJqYESeUPLbz8Bc_y3fouc"
+
+
+class TestEvalService(unittest.TestCase):
+
+    def setUp(self):
+        self.http_client = Mock(spec=httpx.Client)
+        self.eval_service = EvalService(self.http_client)
+
+    def test_list_eval_runs_default(self):
+        mock_eval_run = {
+            "id": 1,
+            "name": "test_eval",
+            "description": "Test evaluation",
+            "user_id": 1,
+            "model_id": 1,
+            "model_name": "test_model",
+            "dataset_id": 1,
+            "dataset_version_id": 1,
+            "dataset_name": "test_dataset",
+            "status": "completed",
+            "prediction_artifact_path": "/path/to/prediction",
+            "evaled_artifact_path": "/path/to/eval",
+            "run_id": "test_run_123",
+            "dataset_summary": {},
+            "metrics_summary": {"accuracy": 0.95},
+            "viz_summary": {},
+            "eval_config": {"metric": "accuracy"},
+            "created_at": 1640995200,
+            "updated_at": 1640995200
+        }
+
+        mock_response = {
+            "code": 0,
+            "msg": None,
+            "data": {
+                "total": 1,
+                "page_size": 20,
+                "page_num": 1,
+                "data": [mock_eval_run]
+            }
+        }
+
+        mock_resp = Mock()
+        mock_resp.json.return_value = mock_response
+        self.http_client.get.return_value = mock_resp
+
+        result = self.eval_service.list()
+
+        self.assertIsInstance(result, ListEvalResp)
+        self.assertEqual(result.total, 1)
+        self.assertEqual(result.page_size, 20)
+        self.assertEqual(result.page_num, 1)
+        self.assertEqual(len(result.data), 1)
+        self.assertEqual(result.data[0].id, 1)
+        self.assertEqual(result.data[0].name, "test_eval")
+
+        self.http_client.get.assert_called_once_with(
+            "/eval-platform/api/v1/run/",
+            params={"page_size": 20, "page_num": 1}
+        )
+
+    def test_list_eval_runs_with_filters(self):
+        mock_response = {
+            "code": 0,
+            "msg": None,
+            "data": {
+                "total": 0,
+                "page_size": 10,
+                "page_num": 1,
+                "data": []
+            }
+        }
+
+        mock_resp = Mock()
+        mock_resp.json.return_value = mock_response
+        self.http_client.get.return_value = mock_resp
+
+        # 带过滤参数
+        result = self.eval_service.list(
+            page_size=10,
+            page_num=1,
+            status="completed",
+            name="test",
+            model_id=1,
+            dataset_id=2,
+            dataset_version_id=3,
+            run_id="test_run",
+            user_id=1,
+            model_ids="1,2,3",
+            dataset_ids="2,3,4",
+            dataset_version_ids="3,4,5"
+        )
+
+        self.assertIsInstance(result, ListEvalResp)
+        self.assertEqual(result.total, 0)
+        self.assertEqual(len(result.data), 0)
+
+        expected_params = {
+            "page_size": 10,
+            "page_num": 1,
+            "status": "completed",
+            "name": "test",
+            "model_id": 1,
+            "dataset_id": 2,
+            "dataset_version_id": 3,
+            "run_id": "test_run",
+            "user_id": 1,
+            "model_ids": "1,2,3",
+            "dataset_ids": "2,3,4",
+            "dataset_version_ids": "3,4,5"
+        }
+        self.http_client.get.assert_called_once_with(
+            "/eval-platform/api/v1/run/",
+            params=expected_params
+        )
+
+    def test_list_eval_runs_api_error(self):
+        """测试列出评测运行 - API错误"""
+        # 模拟 API 错误
+        mock_response = {
+            "code": 1001,
+            "msg": "Database connection failed",
+            "data": None
+        }
+
+        mock_resp = Mock()
+        mock_resp.json.return_value = mock_response
+        self.http_client.get.return_value = mock_resp
+
+        with self.assertRaises(Exception) as context:
+            self.eval_service.list()
+
+        self.assertIn("backend code 1001", str(context.exception))
+        self.assertIn("Database connection failed", str(context.exception))
+
+    def test_list_eval_runs_only_specified_filters(self):
+        mock_response = {
+            "code": 0,
+            "msg": None,
+            "data": {
+                "total": 0,
+                "page_size": 20,
+                "page_num": 1,
+                "data": []
+            }
+        }
+
+        mock_resp = Mock()
+        mock_resp.json.return_value = mock_response
+        self.http_client.get.return_value = mock_resp
+
+        result = self.eval_service.list(
+            status="completed",
+            model_id=1
+        )
+
+        expected_params = {
+            "page_size": 20,
+            "page_num": 1,
+            "status": "completed",
+            "model_id": 1
+        }
+        self.http_client.get.assert_called_once_with(
+            "/eval-platform/api/v1/run/",
+            params=expected_params
+        )
+
+    @patch('aihub.services.dataset_management.DatasetManagementService')
+    def test_create_eval_llm_default_user_id(self, mock_dataset_service_class):
+        """测试创建 LLM 类型评测 - 使用默认 user_id"""
+        # 模拟数据集版本
+        mock_dataset_version = Mock()
+        mock_dataset_version.dataset_id = 1
+        mock_dataset_version.id = 1
+        
+        mock_dataset_service = Mock()
+        mock_dataset_service.get_dataset_version_by_name.return_value = mock_dataset_version
+        mock_dataset_service_class.return_value = mock_dataset_service
+
+        # 模拟成功的创建响应
+        mock_eval_run = {
+            "id": 123,
+            "name": "test_eval",
+            "description": "Test evaluation",
+            "user_id": 0,
+            "model_id": 1,
+            "model_name": "test_model",
+            "dataset_id": 1,
+            "dataset_version_id": 1,
+            "dataset_name": "test_dataset",
+            "status": "created",
+            "prediction_artifact_path": "/path/to/prediction.json",
+            "evaled_artifact_path": "/path/to/evaled.json",
+            "run_id": "test_run_123",
+            "dataset_summary": {},
+            "metrics_summary": {},
+            "viz_summary": {},
+            "eval_config": None,
+            "created_at": 1640995200,
+            "updated_at": 1640995200
+        }
+
+        mock_response = {
+            "code": 0,
+            "msg": None,
+            "data": {"eval_run": mock_eval_run}
+        }
+
+        mock_resp = Mock()
+        mock_resp.json.return_value = mock_response
+        self.http_client.post.return_value = mock_resp
+
+        # 调用方法（不提供 user_id，使用默认值）
+        result = self.eval_service.create(
+            dataset_version_name="test_dataset_v1",
+            prediction_artifact_path="/path/to/prediction.json",
+            evaled_artifact_path="/path/to/evaled.json",
+            report_json={"accuracy": 0.95},
+            run_id="test_run_123"
+        )
+
+        # 验证结果
+        self.assertEqual(result, 123)
+
+        # 验证 HTTP 调用
+        self.http_client.post.assert_called_once()
+        call_args = self.http_client.post.call_args
+        
+        # 验证端点
+        self.assertEqual(call_args[0][0], "/eval-platform/api/v1/run/")
+        
+        # 验证 payload 内容
+        actual_payload = call_args[1]["json"]
+        
+        # 验证 LLM 类型必需字段
+        self.assertEqual(actual_payload["run_id"], "test_run_123")
+        self.assertEqual(actual_payload["type"], "llm")
+        self.assertEqual(actual_payload["prediction_artifact_path"], "/path/to/prediction.json")
+        self.assertEqual(actual_payload["user_id"], 0)
+        self.assertEqual(actual_payload["dataset_id"], 1)
+        self.assertEqual(actual_payload["dataset_version_id"], 1)
+        self.assertEqual(actual_payload["evaled_artifact_path"], "/path/to/evaled.json")
+        self.assertEqual(actual_payload["report"], {"accuracy": 0.95})
+        
+        # CV 类型字段不会出现在 LLM 类型评测的 payload 中
+        self.assertNotIn("metrics_artifact_path", actual_payload)
+        self.assertNotIn("ground_truth_artifact_path", actual_payload)
+
+    @patch('aihub.services.dataset_management.DatasetManagementService')
+    def test_create_eval_llm_custom_user_id(self, mock_dataset_service_class):
+        """测试创建 LLM 类型评测 - 自定义 user_id"""
+        # 模拟数据集版本
+        mock_dataset_version = Mock()
+        mock_dataset_version.dataset_id = 2
+        mock_dataset_version.id = 3
+        
+        mock_dataset_service = Mock()
+        mock_dataset_service.get_dataset_version_by_name.return_value = mock_dataset_version
+        mock_dataset_service_class.return_value = mock_dataset_service
+
+        # 模拟成功的创建响应
+        mock_eval_run = {
+            "id": 456,
+            "name": "test_eval",
+            "description": "Test evaluation",
+            "user_id": 3750,
+            "model_id": 1,
+            "model_name": "test_model",
+            "dataset_id": 2,
+            "dataset_version_id": 3,
+            "dataset_name": "test_dataset",
+            "status": "created",
+            "prediction_artifact_path": "/path/to/prediction.json",
+            "evaled_artifact_path": "/path/to/evaled.json",
+            "run_id": "test_run_456",
+            "dataset_summary": {},
+            "metrics_summary": {},
+            "viz_summary": {},
+            "eval_config": None,
+            "created_at": 1640995200,
+            "updated_at": 1640995200
+        }
+
+        mock_response = {
+            "code": 0,
+            "msg": None,
+            "data": {"eval_run": mock_eval_run}
+        }
+
+        mock_resp = Mock()
+        mock_resp.json.return_value = mock_response
+        self.http_client.post.return_value = mock_resp
+
+        # 调用方法（提供自定义 user_id）
+        result = self.eval_service.create(
+            dataset_version_name="test_dataset_v2",
+            prediction_artifact_path="/path/to/prediction.json",
+            evaled_artifact_path="/path/to/evaled.json",
+            report_json={"f1_score": 0.88},
+            run_id="test_run_456",
+            user_id=3750
+        )
+
+        # 验证结果
+        self.assertEqual(result, 456)
+
+        # 验证 HTTP 调用
+        self.http_client.post.assert_called_once()
+        call_args = self.http_client.post.call_args
+        actual_payload = call_args[1]["json"]
+        
+        # 验证关键字段
+        self.assertEqual(actual_payload["type"], "llm")
+        self.assertEqual(actual_payload["user_id"], 3750)
+        self.assertEqual(actual_payload["dataset_id"], 2)
+        self.assertEqual(actual_payload["dataset_version_id"], 3)
+
+    def test_create_cv_run_default_user_id(self):
+        """测试创建 CV 类型评测运行 - 使用默认 user_id"""
+        # 模拟成功的创建响应
+        mock_eval_run = {
+            "id": 789,
+            "name": "cv_eval",
+            "description": "CV evaluation",
+            "user_id": 0,
+            "model_id": 2,
+            "model_name": "cv_model",
+            "dataset_id": 0,  # 使用默认值而不是 None
+            "dataset_version_id": 0,  # 使用默认值而不是 None
+            "dataset_name": "",  # 使用空字符串而不是 None
+            "status": "created",
+            "prediction_artifact_path": "coco_dt.json",
+            "evaled_artifact_path": "",  # 使用空字符串而不是 None
+            "run_id": "cv_run_789",
+            "dataset_summary": {},
+            "metrics_summary": {},
+            "viz_summary": {},
+            "eval_config": None,
+            "created_at": 1640995200,
+            "updated_at": 1640995200
+        }
+
+        mock_response = {
+            "code": 0,
+            "msg": None,
+            "data": {"eval_run": mock_eval_run}
+        }
+
+        mock_resp = Mock()
+        mock_resp.json.return_value = mock_response
+        self.http_client.post.return_value = mock_resp
+
+        # 调用方法（不提供 user_id，使用默认值）
+        result = self.eval_service.create_cv_run(
+            run_id="cv_run_789",
+            prediction_artifact_path="coco_dt.json",
+            metrics_artifact_path="metrics.json",
+            ground_truth_artifact_path="coco_gt.json"
+        )
+
+        # 验证结果
+        self.assertEqual(result, 789)
+
+        # 验证 HTTP 调用
+        self.http_client.post.assert_called_once()
+        call_args = self.http_client.post.call_args
+        actual_payload = call_args[1]["json"]
+        
+        # 验证 CV 类型必需字段
+        self.assertEqual(actual_payload["type"], "cv")
+        self.assertEqual(actual_payload["user_id"], 0)
+        self.assertEqual(actual_payload["run_id"], "cv_run_789")
+        self.assertEqual(actual_payload["prediction_artifact_path"], "coco_dt.json")
+        self.assertEqual(actual_payload["metrics_artifact_path"], "metrics.json")
+        self.assertEqual(actual_payload["ground_truth_artifact_path"], "coco_gt.json")
+        
+        # LLM 类型字段不会出现在 CV 类型评测的 payload 中（Pydantic v2 自动排除未设置的可选字段）
+        self.assertNotIn("dataset_id", actual_payload)
+        self.assertNotIn("dataset_version_id", actual_payload)
+        self.assertNotIn("evaled_artifact_path", actual_payload)
+        self.assertNotIn("report", actual_payload)
+
+    def test_create_cv_run_custom_user_id(self):
+        """测试创建 CV 类型评测运行 - 自定义 user_id"""
+        # 模拟成功的创建响应
+        mock_eval_run = {
+            "id": 999,
+            "name": "cv_eval_custom",
+            "description": "CV evaluation custom",
+            "user_id": 3750,
+            "model_id": 3,
+            "model_name": "cv_model_v2",
+            "dataset_id": 0,  # 使用默认值而不是 None
+            "dataset_version_id": 0,  # 使用默认值而不是 None
+            "dataset_name": "",  # 使用空字符串而不是 None
+            "status": "created",
+            "prediction_artifact_path": "coco_dt.json",
+            "evaled_artifact_path": "",  # 使用空字符串而不是 None
+            "run_id": "cv_run_999",
+            "dataset_summary": {},
+            "metrics_summary": {},
+            "viz_summary": {},
+            "eval_config": None,
+            "created_at": 1640995200,
+            "updated_at": 1640995200
+        }
+
+        mock_response = {
+            "code": 0,
+            "msg": None,
+            "data": {"eval_run": mock_eval_run}
+        }
+
+        mock_resp = Mock()
+        mock_resp.json.return_value = mock_response
+        self.http_client.post.return_value = mock_resp
+
+        # 调用方法（提供自定义 user_id）
+        result = self.eval_service.create_cv_run(
+            run_id="cv_run_999",
+            prediction_artifact_path="coco_dt.json",
+            metrics_artifact_path="metrics.json",
+            ground_truth_artifact_path="coco_gt.json",
+            user_id=3750
+        )
+
+        # 验证结果
+        self.assertEqual(result, 999)
+
+        # 验证 HTTP 调用
+        self.http_client.post.assert_called_once()
+        call_args = self.http_client.post.call_args
+        actual_payload = call_args[1]["json"]
+        
+        # 验证 CV 类型关键字段
+        self.assertEqual(actual_payload["type"], "cv")
+        self.assertEqual(actual_payload["user_id"], 3750)
+        self.assertEqual(actual_payload["run_id"], "cv_run_999")
+        self.assertEqual(actual_payload["prediction_artifact_path"], "coco_dt.json")
+        self.assertEqual(actual_payload["metrics_artifact_path"], "metrics.json")
+        self.assertEqual(actual_payload["ground_truth_artifact_path"], "coco_gt.json")
+
+    def test_create_eval_api_error(self):
+        """测试创建评测 - API错误"""
+        with patch('aihub.services.dataset_management.DatasetManagementService') as mock_dataset_service_class:
+            # 模拟数据集版本
+            mock_dataset_version = Mock()
+            mock_dataset_version.dataset_id = 1
+            mock_dataset_version.id = 1
+            
+            mock_dataset_service = Mock()
+            mock_dataset_service.get_dataset_version_by_name.return_value = mock_dataset_version
+            mock_dataset_service_class.return_value = mock_dataset_service
+
+            # 模拟 API 错误
+            mock_response = {
+                "code": 2001,
+                "msg": "Invalid dataset version",
+                "data": None
+            }
+
+            mock_resp = Mock()
+            mock_resp.json.return_value = mock_response
+            self.http_client.post.return_value = mock_resp
+
+            with self.assertRaises(Exception) as context:
+                self.eval_service.create(
+                    dataset_version_name="invalid_dataset",
+                    prediction_artifact_path="/path/to/prediction.json",
+                    evaled_artifact_path="/path/to/evaled.json",
+                    report_json={"accuracy": 0.95},
+                    run_id="test_run_error"
+                )
+
+            self.assertIn("backend code 2001", str(context.exception))
+            self.assertIn("Invalid dataset version", str(context.exception))
+
+    def test_create_cv_run_api_error(self):
+        """测试创建 CV 评测运行 - API错误"""
+        # 模拟 API 错误
+        mock_response = {
+            "code": 3001,
+            "msg": "Invalid CV run parameters",
+            "data": None
+        }
+
+        mock_resp = Mock()
+        mock_resp.json.return_value = mock_response
+        self.http_client.post.return_value = mock_resp
+
+        with self.assertRaises(Exception) as context:
+            self.eval_service.create_cv_run(
+                run_id="invalid_cv_run",
+                prediction_artifact_path="invalid.json",
+                metrics_artifact_path="invalid_metrics.json",
+                ground_truth_artifact_path="invalid_gt.json"
+            )
+
+        self.assertIn("backend code 3001", str(context.exception))
+        self.assertIn("Invalid CV run parameters", str(context.exception))
+
+
+if __name__ == "__main__":
+    unittest.main()

intellif_aihub-0.1.19/src/aihub/__init__.py

@@ -1 +0,0 @@
-__version__ = "0.1.19"

intellif_aihub-0.1.19/src/aihub/models/eval.py

@@ -1,26 +0,0 @@
-# !/usr/bin/env python
-# -*-coding:utf-8 -*-
-from typing import Dict
-
-from pydantic import BaseModel, Field
-
-
-class CreateEvalReq(BaseModel):
-    """创建评测任务"""
-    dataset_id: int = Field(description="数据集ID")
-    dataset_version_id: int = Field(description="数据集版本ID")
-    prediction_artifact_path: str = Field(description="推理产物的路径")
-    evaled_artifact_path: str = Field(description="评测结果产物的路径")
-    run_id: str = Field(description="运行ID")
-    user_id: int = Field(0, description="用户ID")
-    report: Dict = Field(default_factory=dict, description="评测报告")
-
-
-class EvalRun(BaseModel):
-    """评测任务的运行实体"""
-    id: int = Field(description="评测的运行ID")
-
-
-class CreateEvalResp(BaseModel):
-    """创建评测任务的返回结果"""
-    eval_run: EvalRun = Field(alias="eval_run", description="评测运行信息")

intellif_aihub-0.1.19/src/aihub/services/eval.py

@@ -1,75 +0,0 @@
-# !/usr/bin/env python
-# -*- coding:utf-8 -*-
-"""评测平台服务模块
-
-本模块围绕 **“模型评测（Run → Report）”** 提供能力：
-
-- **创建评测任务 / 评测报告**
-"""
-
-import httpx
-
-from ..exceptions import APIError
-from ..models.common import APIWrapper
-from ..models.eval import CreateEvalReq, CreateEvalResp
-
-_BASE = "/eval-platform/api/v1"
-
-
-class EvalService:
-    """评测服务"""
-
-    def __init__(self, http: httpx.Client):
-        self._http = http
-        self._eval = _Eval(http)
-
-    def create(
-        self,
-        dataset_version_name: str,
-        prediction_artifact_path: str,
-        evaled_artifact_path: str,
-        report_json: dict,
-        run_id,
-    ) -> int:
-        """创建评测报告
-
-        Args:
-            run_id (str): RUN ID
-            report_json (dict): 报告内容
-            evaled_artifact_path:   评测结果制品路径
-            prediction_artifact_path: 推理结果制品路径
-            dataset_version_name (str): 数据集名称
-
-
-        Returns:
-            id (int): 评测报告id
-
-        """
-        from .dataset_management import DatasetManagementService
-
-        dataset_service = DatasetManagementService(self._http)
-        dataset_version = dataset_service.get_dataset_version_by_name(
-            dataset_version_name
-        )
-        payload = CreateEvalReq(
-            dataset_id=dataset_version.dataset_id,
-            dataset_version_id=dataset_version.id,
-            evaled_artifact_path=evaled_artifact_path,
-            prediction_artifact_path=prediction_artifact_path,
-            report=report_json,
-            run_id=run_id,
-        )
-
-        return self._eval.create(payload)
-
-
-class _Eval:
-    def __init__(self, http: httpx.Client):
-        self._http = http
-
-    def create(self, payload: CreateEvalReq) -> int:
-        resp = self._http.post(f"{_BASE}/run/", json=payload.model_dump())
-        wrapper = APIWrapper[CreateEvalResp].model_validate(resp.json())
-        if wrapper.code != 0:
-            raise APIError(f"backend code {wrapper.code}: {wrapper.msg}")
-        return wrapper.data.eval_run.id