q792602257/DataMate
0
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

优化部分问题 (#126)

Browse Source 
* feature: 支持相对路径引用

* feature: 优化本地部署命令

* feature: 优化算子编排展示

* feature: 优化清洗任务失败后重试
This commit is contained in:
hhhhsc701
2025-12-03 16:41:48 +08:00
committed by GitHub
parent 04cff3f7e6
commit c22683d635
5 changed files with 19 additions and 22 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
README-zh.md
View File

@@ -70,15 +70,15 @@ make install-mineru
```
### 部署DeerFlow服务
1. 修改runtime/deer-flow/.env.example，添加SEARCH_API_KEY和EMBEDDING模型配置
2. 修改runtime/deer-flow/.conf.yaml.example，添加基础模型服务配置
3. 执行`make install-deer-flow`
```bash
make install-deer-flow
```
### 本地开发部署
本地代码修改后，请执行以下命令构建镜像并使用本地镜像部署
```bash
make build
make install REGISTRY=""
make install dev=true
```
## 🤝 贡献指南

10
backend/services/data-cleaning-service/src/main/java/com/datamate/cleaning/application/CleaningTaskService.java
View File

@@ -171,10 +171,10 @@ public class CleaningTaskService {
}
public void executeTask(String taskId) {
List<CleaningResultDto> failed = cleaningResultRepo.findByInstanceId(taskId, "FAILED");
Set<String> failedSet = failed.stream().map(CleaningResultDto::getSrcFileId).collect(Collectors.toSet());
List<CleaningResultDto> succeed = cleaningResultRepo.findByInstanceId(taskId, "COMPLETED");
Set<String> succeedSet = succeed.stream().map(CleaningResultDto::getSrcFileId).collect(Collectors.toSet());
CleaningTaskDto task = cleaningTaskRepo.findTaskById(taskId);
scanDataset(taskId, task.getSrcDatasetId(), failedSet);
scanDataset(taskId, task.getSrcDatasetId(), succeedSet);
cleaningResultRepo.deleteByInstanceId(taskId, "FAILED");
taskScheduler.executeTask(taskId);
}
@@ -232,7 +232,7 @@ public class CleaningTaskService {
} while (pageNumber < datasetFiles.getTotalPages());
}
private void scanDataset(String taskId, String srcDatasetId, Set<String> failedFiles) {
private void scanDataset(String taskId, String srcDatasetId, Set<String> succeedFiles) {
int pageNumber = 0;
int pageSize = 500;
PagingQuery pageRequest = new PagingQuery(pageNumber, pageSize);
@@ -243,7 +243,7 @@ public class CleaningTaskService {
break;
}
List<Map<String, Object>> files = datasetFiles.getContent().stream()
.filter(content -> failedFiles.contains(content.getId()))
.filter(content -> !succeedFiles.contains(content.getId()))
.map(content -> Map.of("fileName", (Object) content.getFileName(),
"fileSize", content.getFileSize(),
"filePath", content.getFilePath(),

8
frontend/src/pages/DataCleansing/Create/components/OperatorOrchestration.tsx
View File

@@ -182,14 +182,6 @@ const OperatorFlow: React.FC<OperatorFlowProps> = ({
{operator?.categories?.map((categoryId) => {
return <Tag color="default">{categoryMap[categoryId].name}</Tag>
})}
{/* 参数状态指示 */}
{Object.values(operator.configs).some(
(param: any) =>
(param.type === "input" && !param.value) ||
(param.type === "checkbox" &&
Array.isArray(param.value) &&
param.value.length === 0)
) && <Tag color="red">待配置</Tag>}
{/* 操作按钮 */}
<span
className="cursor-pointer text-red-500"

13
runtime/python-executor/datamate/core/dataset.py
View File

@@ -16,6 +16,8 @@ from datamate.core.base_op import Filter, Mapper, Slicer
from datamate.core.constant import Fields
from datamate.core.base_op import OPERATORS, BaseOp
from core.base_op import Filter as RELATIVE_Filter, Mapper as RELATIVE_Mapper, Slicer as RELATIVE_Slicer
rd.DataContext.get_current().enable_progress_bars = False
@@ -136,7 +138,10 @@ class RayDataset(BasicDataset):
parent_dir = os.path.join(os.path.dirname(os.path.dirname(os.path.abspath(__file__))), "ops")
if parent_dir not in sys.path:
sys.path.insert(0, parent_dir)
registry_content = OPERATORS.modules[op_name]
registry_content = OPERATORS.modules.get(op_name)
if registry_content is None:
from core.base_op import OPERATORS as RELATIVE_OPERATORS
registry_content = RELATIVE_OPERATORS.modules.get(op_name)
if isinstance(registry_content, str):
# registry_content是module的路径
submodule = importlib.import_module(registry_content)
@@ -171,7 +176,7 @@ class RayDataset(BasicDataset):
kwargs.update({"ext_params": {}, "failed_reason": {}, "target_type": None})
try:
if issubclass(operators_cls, Mapper):
if issubclass(operators_cls, (Mapper, RELATIVE_Mapper)):
self.data = self.data.map(operators_cls,
fn_constructor_kwargs=init_kwargs,
fn_kwargs=kwargs,
@@ -179,7 +184,7 @@ class RayDataset(BasicDataset):
num_cpus=0.05,
concurrency=(1, 1 if operators_cls.use_model else int(max_actor_nums)))
elif issubclass(operators_cls, Slicer):
elif issubclass(operators_cls, (Slicer, RELATIVE_Slicer)):
self.data = self.data.flat_map(operators_cls,
fn_constructor_kwargs=init_kwargs,
fn_kwargs=kwargs,
@@ -187,7 +192,7 @@ class RayDataset(BasicDataset):
num_cpus=0.05,
concurrency=(1, int(max_actor_nums)))
elif issubclass(operators_cls, Filter):
elif issubclass(operators_cls, (Filter, RELATIVE_Filter)):
self.data = self.data.filter(operators_cls,
fn_constructor_kwargs=init_kwargs,
fn_kwargs=kwargs,

2
scripts/images/deer-flow-backend/Dockerfile
View File

@@ -26,4 +26,4 @@ RUN --mount=type=cache,target=/root/.cache/uv \
EXPOSE 8000
# Run the application.
CMD ["uv", "run", "python", "server.py", "--host", "0.0.0.0", "--port", "8000"]
CMD ["uv", "run", "--no-sync", "python", "server.py", "--host", "0.0.0.0", "--port", "8000"]