feat: enhance backend deployment, frontend file selection and synthesis task management (#129)

* feat: Implement data synthesis task management with database models and API endpoints

* feat: Update Python version requirements and refine dependency constraints in configuration

* fix: Correctly extract file values from selectedFilesMap in AddDataDialog

* feat: Refactor synthesis task routes and enhance file task management in the API

* feat: Enhance SynthesisTaskTab with tooltip actions and add chunk data retrieval in API

.github/workflows/docker-image-backend-python.yml

@@ -5,12 +5,14 @@ on:
     branches: [ "main" ]
     paths:
       - 'scripts/images/datamate-python/**'
+      - 'runtime/datamate-python/**'
       - '.github/workflows/docker-image-backend-python.yml'
       - '.github/workflows/docker-images-reusable.yml'
   pull_request:
     branches: [ "main" ]
     paths:
       - 'scripts/images/datamate-python/**'
+      - 'runtime/datamate-python/**'
       - '.github/workflows/docker-image-backend-python.yml'
       - '.github/workflows/docker-images-reusable.yml'
   workflow_dispatch:

deployment/docker/milvus/docker-compose.yml

@@ -10,6 +10,7 @@ services:
     volumes:
       - ${DOCKER_VOLUME_DIRECTORY:-.}/volumes/etcd:/etcd
     command: etcd -advertise-client-urls=http://etcd:2379 -listen-client-urls http://0.0.0.0:2379 --data-dir /etcd
+    restart: always
     networks:
       - datamate
     healthcheck:

frontend/src/components/business/DatasetFileTransfer.tsx

@@ -1,4 +1,4 @@
-import React, { useEffect } from "react";
+import React, { useCallback, useEffect } from "react";
 import { Button, Input, Table } from "antd";
 import { RightOutlined } from "@ant-design/icons";
 import { mapDataset } from "@/pages/DataManagement/dataset.const";
@@ -19,6 +19,7 @@ interface DatasetFileTransferProps
   open: boolean;
   selectedFilesMap: { [key: string]: DatasetFile };
   onSelectedFilesChange: (filesMap: { [key: string]: DatasetFile }) => void;
+  onDatasetSelect?: (dataset: Dataset | null) => void;
 }
 
 const fileCols = [
@@ -48,6 +49,7 @@ const DatasetFileTransfer: React.FC<DatasetFileTransferProps> = ({
   open,
   selectedFilesMap,
   onSelectedFilesChange,
+  onDatasetSelect,
   ...props
 }) => {
   const [datasets, setDatasets] = React.useState<Dataset[]>([]);
@@ -96,7 +98,7 @@ const DatasetFileTransfer: React.FC<DatasetFileTransferProps> = ({
     300
   );
 
-  const fetchFiles = async () => {
+  const fetchFiles = useCallback(async () => {
     if (!selectedDataset) return;
     const { data } = await queryDatasetFilesUsingGet(selectedDataset.id, {
       page: filesPagination.current - 1,
@@ -104,23 +106,25 @@ const DatasetFileTransfer: React.FC<DatasetFileTransferProps> = ({
       keyword: filesSearch,
     });
     setFiles(
-      data.content.map((item) => ({
+      (data.content || []).map((item: DatasetFile) => ({
         ...item,
         key: item.id,
         datasetName: selectedDataset.name,
-      })) || []
+      }))
     );
     setFilesPagination((prev) => ({
       ...prev,
       total: data.totalElements,
     }));
-  };
+  }, [filesPagination.current, filesPagination.pageSize, filesSearch, selectedDataset]);
 
   useEffect(() => {
-    if (selectedDataset) {
+    fetchFiles().catch(() => {});
-      fetchFiles();
+  }, [fetchFiles]);
-    }
+
-  }, [selectedDataset]);
+  useEffect(() => {
+    onDatasetSelect?.(selectedDataset);
+  }, [selectedDataset, onDatasetSelect]);
 
   const toggleSelectFile = (record: DatasetFile) => {
     if (!selectedFilesMap[record.id]) {
@@ -147,8 +151,9 @@ const DatasetFileTransfer: React.FC<DatasetFileTransferProps> = ({
       setShowFiles(false);
       setSelectedDataset(null);
       setDatasetSelections([]);
+      onDatasetSelect?.(null);
     }
-  }, [open]);
+  }, [open, onDatasetSelect]);
 
   const datasetCols = [
     {
@@ -206,7 +211,15 @@ const DatasetFileTransfer: React.FC<DatasetFileTransferProps> = ({
             })}
             dataSource={datasets}
             columns={datasetCols}
-            pagination={datasetPagination}
+            pagination={{
+                ...datasetPagination,
+                onChange: (page, pageSize) =>
+                    setDatasetPagination({
+                        current: page,
+                        pageSize: pageSize || datasetPagination.pageSize,
+                        total: datasetPagination.total,
+                    }),
+            }}
           />
         </div>
         <RightOutlined />
@@ -231,21 +244,11 @@ const DatasetFileTransfer: React.FC<DatasetFileTransferProps> = ({
             })}
             rowSelection={{
               type: "checkbox",
-              onSelectAll: (selected, _, changeRows) => {
-                const newSelectedFiles = { ...selectedFilesMap };
-                if (selected) {
-                  changeRows.forEach((row) => {
-                    newSelectedFiles[row.id] = row;
-                  });
-                } else {
-                  changeRows.forEach((row) => {
-                    delete newSelectedFiles[row.id];
-                  });
-                }
-                onSelectedFilesChange(newSelectedFiles);
-              },
               selectedRowKeys: Object.keys(selectedFilesMap),
               onSelect: toggleSelectFile,
+              getCheckboxProps: (record: DatasetFile) => ({
+                name: record.fileName,
+              }),
             }}
           />
         </div>

frontend/src/pages/SynthesisTask/SynthFileTask.tsx

@@ -0,0 +1,186 @@
+import { useEffect, useState } from "react";
+import { useParams, useNavigate } from "react-router";
+import { Table, Badge, Button } from "antd";
+import type { ColumnsType, TablePaginationConfig } from "antd/es/table";
+import { querySynthesisFileTasksUsingGet, querySynthesisTaskByIdUsingGet } from "@/pages/SynthesisTask/synthesis-api";
+import type { BadgeProps } from "antd";
+import { formatDateTime } from "@/utils/unit";
+
+interface SynthesisFileTaskItem {
+  id: string;
+  synthesis_instance_id: string;
+  file_name: string;
+  source_file_id: string;
+  target_file_location: string;
+  status?: string;
+  total_chunks: number;
+  processed_chunks: number;
+  created_at?: string;
+  updated_at?: string;
+}
+
+interface PagedResponse<T> {
+  content: T[];
+  totalElements: number;
+  totalPages: number;
+  page: number;
+  size: number;
+}
+
+interface SynthesisTaskInfo {
+  id: string;
+  name: string;
+  synthesis_type: string;
+  status: string;
+  created_at: string;
+  model_id: string;
+}
+
+export default function SynthFileTask() {
+  const { id: taskId = "" } = useParams();
+  const navigate = useNavigate();
+  const [loading, setLoading] = useState(false);
+  const [data, setData] = useState<SynthesisFileTaskItem[]>([]);
+  const [pagination, setPagination] = useState<TablePaginationConfig>({
+    current: 1,
+    pageSize: 10,
+    total: 0,
+  });
+  const [taskInfo, setTaskInfo] = useState<SynthesisTaskInfo | null>(null);
+
+  // 查询总任务详情
+  useEffect(() => {
+    if (!taskId) return;
+    querySynthesisTaskByIdUsingGet(taskId).then((res) => {
+      setTaskInfo(res?.data?.data || null);
+    });
+  }, [taskId]);
+
+  const fetchData = async (page = 1, pageSize = 10) => {
+    if (!taskId) return;
+    setLoading(true);
+    try {
+      const res = await querySynthesisFileTasksUsingGet(taskId, {
+        page,
+        page_size: pageSize,
+      });
+      const payload: PagedResponse<SynthesisFileTaskItem> =
+        res?.data?.data ?? res?.data ?? {
+          content: [],
+          totalElements: 0,
+          totalPages: 0,
+          page,
+          size: pageSize,
+        };
+      setData(payload.content || []);
+      setPagination({
+        current: payload.page ?? page,
+        pageSize: payload.size ?? pageSize,
+        total: payload.totalElements ?? payload.content?.length ?? 0,
+      });
+    } finally {
+      setLoading(false);
+    }
+  };
+
+  useEffect(() => {
+    fetchData(1, pagination.pageSize || 10);
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [taskId]);
+
+  const handleTableChange = (pag: TablePaginationConfig) => {
+    fetchData(pag.current || 1, pag.pageSize || 10);
+  };
+
+  const columns: ColumnsType<SynthesisFileTaskItem> = [
+    {
+      title: "文件名",
+      dataIndex: "file_name",
+      key: "file_name",
+    },
+    {
+      title: "状态",
+      dataIndex: "status",
+      key: "status",
+      render: (status?: string) => {
+        let badgeStatus: BadgeProps["status"] = "default";
+        let text = status || "未知";
+        if (status === "pending" || status === "processing") {
+          badgeStatus = "processing";
+          text = "处理中";
+        } else if (status === "completed") {
+          badgeStatus = "success";
+          text = "已完成";
+        } else if (status === "failed") {
+          badgeStatus = "error";
+          text = "失败";
+        }
+        return <Badge status={badgeStatus} text={text} />;
+      },
+    },
+    {
+      title: "切片进度",
+      key: "chunks",
+      render: (_text, record) => (
+        <span>
+          {record.processed_chunks}/{record.total_chunks}
+        </span>
+      ),
+    },
+    {
+      title: "目标文件路径",
+      dataIndex: "target_file_location",
+      key: "target_file_location",
+      ellipsis: true,
+    },
+    {
+      title: "创建时间",
+      dataIndex: "created_at",
+      key: "created_at",
+      render: (val?: string) => (val ? formatDateTime(val) : "-"),
+    },
+    {
+      title: "更新时间",
+      dataIndex: "updated_at",
+      key: "updated_at",
+      render: (val?: string) => (val ? formatDateTime(val) : "-"),
+    },
+  ];
+
+  return (
+    <div className="p-4 bg-white rounded-lg h-full flex flex-col">
+      {/* 顶部任务信息和返回按钮 */}
+      <div className="flex items-center justify-between mb-4">
+        <div className="space-y-1">
+          {taskInfo && (
+            <>
+              <div className="text-lg font-medium flex items-center gap-2">
+                <span>{taskInfo.name}</span>
+                <span className="text-xs px-2 py-0.5 rounded bg-blue-50 text-blue-700 border border-blue-200">
+                  {taskInfo.synthesis_type === "QA" ? "问答对生成" : taskInfo.synthesis_type === "COT" ? "链式推理生成" : taskInfo.synthesis_type}
+                </span>
+                <span className="text-xs px-2 py-0.5 rounded bg-gray-50 text-gray-700 border border-gray-200">
+                  状态：{taskInfo.status === "pending" ? "等待中" : taskInfo.status === "completed" ? "已完成" : taskInfo.status === "failed" ? "失败" : taskInfo.status}
+                </span>
+              </div>
+              <div className="text-xs text-gray-500 flex gap-4">
+                <span>创建时间：{formatDateTime(taskInfo.created_at)}</span>
+                <span>模型ID：{taskInfo.model_id}</span>
+              </div>
+            </>
+          )}
+        </div>
+        <Button type="default" onClick={() => navigate("/data/synthesis/task")}>返回任务首页</Button>
+      </div>
+      {/* 文件任务表格 */}
+      <Table<SynthesisFileTaskItem>
+        rowKey="id"
+        loading={loading}
+        dataSource={data}
+        columns={columns}
+        pagination={pagination}
+        onChange={handleTableChange}
+      />
+    </div>
+  );
+}

frontend/src/pages/SynthesisTask/components/SynthesisTaskTab.tsx

@@ -1,112 +1,116 @@
-import { useState } from "react";
+import { useState, useEffect, ElementType } from "react";
-import { Card, Button, Badge, Table, Progress } from "antd";
+import { Card, Button, Badge, Table, Modal, message, Tooltip } from "antd";
 import {
   Plus,
-  Sparkles,
   ArrowUp,
   ArrowDown,
   Pause,
   Play,
-  DownloadIcon,
   CheckCircle,
-  Check,
+  Sparkles,
-  StopCircle,
 } from "lucide-react";
-import type { SynthesisTask } from "@/pages/SynthesisTask/synthesis";
+import { DeleteOutlined, EyeOutlined } from "@ant-design/icons";
-import { mockSynthesisTasks } from "@/mock/synthesis";
 import { Link, useNavigate } from "react-router";
 import { SearchControls } from "@/components/SearchControls";
 import { formatDateTime } from "@/utils/unit";
+import {
+  querySynthesisTasksUsingGet,
+  deleteSynthesisTaskByIdUsingDelete,
+} from "@/pages/SynthesisTask/synthesis-api";
+
+interface SynthesisTask {
+  id: string;
+  name: string;
+  description?: string;
+  status: string;
+  synthesis_type: string;
+  model_id: string;
+  progress?: number;
+  result_data_location?: string;
+  text_split_config?: {
+    chunk_size: number;
+    chunk_overlap: number;
+  };
+  synthesis_config?: {
+    temperature?: number | null;
+    prompt_template?: string;
+    synthesis_count?: number | null;
+  };
+  source_file_id?: string[];
+  total_files?: number;
+  processed_files?: number;
+  total_chunks?: number;
+  processed_chunks?: number;
+  total_synthesis_data?: number;
+  created_at: string;
+  updated_at?: string;
+  created_by?: string;
+  updated_by?: string;
+}
 
 export default function SynthesisTaskTab() {
   const navigate = useNavigate();
   const [searchQuery, setSearchQuery] = useState("");
-  const [tasks, setTasks] = useState<SynthesisTask[]>(mockSynthesisTasks);
+  const [tasks, setTasks] = useState<SynthesisTask[]>([]);
   const [filterStatus, setFilterStatus] = useState("all");
   const [sortBy, setSortBy] = useState<"createdAt" | "name">("createdAt");
   const [sortOrder, setSortOrder] = useState<"asc" | "desc">("desc");
+  const [page, setPage] = useState(1);
+  const [pageSize, setPageSize] = useState(10);
+  const [total, setTotal] = useState(0);
+  const [loading, setLoading] = useState(false);
 
-  // 过滤任务
+  // 获取任务列表
-  const filteredTasks = tasks.filter((task) => {
+  const loadTasks = async () => {
-    const matchesSearch =
+    setLoading(true);
-      task.name.toLowerCase().includes(searchQuery.toLowerCase()) ||
+    try {
-      task.template.toLowerCase().includes(searchQuery.toLowerCase());
+      const params = {
-    const matchesStatus =
+        page: page,
-      filterStatus === "all" || task.status === filterStatus;
+        page_size: pageSize,
-    return matchesSearch && matchesStatus;
+      } as {
-  });
+        page?: number;
-
+        page_size?: number;
-  // 排序任务
+        synthesis_type?: string;
-  const sortedTasks = [...filteredTasks].sort((a, b) => {
+        status?: string;
-    if (sortBy === "createdAt") {
+        name?: string;
-      const dateA = new Date(a.createdAt).getTime();
+      };
-      const dateB = new Date(b.createdAt).getTime();
+      if (searchQuery) params.name = searchQuery;
-      return sortOrder === "asc" ? dateA - dateB : dateB - dateA;
+      if (filterStatus !== "all") params.synthesis_type = filterStatus;
-    } else if (sortBy === "name") {
+      const res = await querySynthesisTasksUsingGet(params);
-      return sortOrder === "asc"
+      setTasks(res?.data?.content || []);
-        ? a.name.localeCompare(b.name)
+      setTotal(res?.data?.totalElements || 0);
-        : b.name.localeCompare(a.name);
+    } catch {
+      setTasks([]);
+      setTotal(0);
+    } finally {
+      setLoading(false);
     }
-    return 0;
-  });
-  const handleTaskAction = (taskId: number, action: string) => {
-    setTasks((prev) =>
-      prev.map((task) => {
-        if (task.id === taskId) {
-          switch (action) {
-            case "pause":
-              return { ...task, status: "paused" as const };
-            case "resume":
-              return { ...task, status: "running" as const };
-            case "stop":
-              return {
-                ...task,
-                status: "failed" as const,
-                progress: task.progress,
-              };
-            default:
-              return task;
-          }
-        }
-        return task;
-      })
-    );
   };
+
+  useEffect(() => {
+    loadTasks();
+    // eslint-disable-next-line
+  }, [searchQuery, filterStatus, page, pageSize]);
+
   // 状态徽章
   const getStatusBadge = (status: string) => {
-    const statusConfig = {
+    const statusConfig: Record<string, { label: string; color: string; icon: ElementType }> = {
-      pending: {
+      pending: { label: "等待中", color: "#F59E0B", icon: Pause },
-        label: "等待中",
+      running: { label: "运行中", color: "#3B82F6", icon: Play },
-        color: "#F59E0B",
+      completed: { label: "已完成", color: "#10B981", icon: CheckCircle },
-        icon: Pause,
+      failed: { label: "失败", color: "#EF4444", icon: Pause },
-      },
+      paused: { label: "已暂停", color: "#E5E7EB", icon: Pause },
-      running: {
-        label: "运行中",
-        color: "#3B82F6",
-        icon: Play,
-      },
-      completed: {
-        label: "已完成",
-        color: "#10B981",
-        icon: CheckCircle,
-      },
-      failed: {
-        label: "失败",
-        color: "#EF4444",
-        icon: Pause,
-      },
-      paused: {
-        label: "已暂停",
-        color: "#E5E7EB",
-        icon: Pause,
-      },
     };
-    return (
+    return statusConfig[status] ?? statusConfig["pending"];
-      statusConfig[status as keyof typeof statusConfig] || statusConfig.pending
-    );
   };
 
-  // 任务表格列
+  // 类型映射
+  const typeMap: Record<string, string> = {
+    QA: "问答对生成",
+    COT: "链式推理生成",
+  };
+
+  // 表格列
   const taskColumns = [
     {
       title: (
@@ -134,98 +138,77 @@ export default function SynthesisTaskTab() {
       dataIndex: "name",
       key: "name",
       fixed: "left" as const,
-      render: (text: string, task: SynthesisTask) => (
+      render: (_: unknown, task: SynthesisTask) => (
         <div className="flex items-center gap-3">
           <div className="w-8 h-8 bg-blue-500 rounded-lg flex items-center justify-center shadow-sm">
-            {/* 可根据 type 渲染不同图标 */}
             <span className="text-white font-bold text-base">
-              {task.type?.toUpperCase()?.slice(0, 1) || "T"}
+              {task.synthesis_type?.toUpperCase()?.slice(0, 1) || "T"}
             </span>
           </div>
           <div>
             <Link to={`/data/synthesis/task/${task.id}`}>{task.name}</Link>
-            <div className="text-xs text-gray-500">{task.template}</div>
           </div>
         </div>
       ),
     },
     {
       title: "类型",
-      dataIndex: "type",
+      dataIndex: "synthesis_type",
-      key: "type",
+      key: "synthesis_type",
-      render: (type: string) => type.toUpperCase(),
+      render: (type: string) => typeMap[type] || type,
     },
     {
-      title: "状态",
+      title: "文件数",
-      dataIndex: "status",
+      dataIndex: "total_files",
-      key: "status",
+      key: "total_files",
-      render: (status: string) => {
+      render: (num: number, task: SynthesisTask) => <span>{num ?? (task.source_file_id?.length ?? 0)}</span>,
-        const statusConfig = getStatusBadge(status);
-        return <Badge color={statusConfig.color} text={statusConfig.label} />;
-      },
-    },
-    {
-      title: "进度",
-      dataIndex: "progress",
-      key: "progress",
-      width: 150,
-      render: (_: any, task: SynthesisTask) => (
-        <Progress percent={task.progress} size="small" />
-      ),
-    },
-    {
-      title: "源数据集",
-      dataIndex: "sourceDataset",
-      key: "sourceDataset",
-      render: (text: string) => (
-        <div className="text-sm text-gray-900">{text}</div>
-      ),
-    },
-    {
-      title: "生成数量",
-      dataIndex: "generatedCount",
-      key: "generatedCount",
-      render: (_: any, task: SynthesisTask) => (
-        <div className="text-sm font-medium text-gray-900">
-          {task.generatedCount?.toLocaleString?.()} /{" "}
-          {task.targetCount?.toLocaleString?.()}
-        </div>
-      ),
-    },
-    {
-      title: "质量评分",
-      dataIndex: "quality",
-      key: "quality",
-      render: (quality: number) => (quality ? `${quality}%` : "-"),
     },
     {
       title: "创建时间",
-      dataIndex: "createdAt",
+      dataIndex: "created_at",
-      key: "createdAt",
+      key: "created_at",
-      render: formatDateTime,
+      render: (val: string) => formatDateTime(val),
     },
     {
       title: "操作",
       key: "actions",
       fixed: "right" as const,
-      render: (_: any, task: SynthesisTask) => (
+      render: (_: unknown, task: SynthesisTask) => (
         <div className="flex items-center justify-center gap-1">
-          {task.status === "running" && (
+          <Tooltip title="查看详情">
             <Button
-              onClick={() => handleTaskAction(task.id, "pause")}
+              onClick={() => navigate(`/data/synthesis/task/${task.id}`)}
-              className="hover:bg-orange-50 p-1 h-7 w-7"
+              className="hover:bg-blue-50 p-1 h-7 w-7"
               type="text"
-              icon={<Pause className="w-4 h-4" />}
+              icon={<EyeOutlined />}
-            ></Button>
+            />
-          )}
+          </Tooltip>
-          {task.status === "paused" && (
+          <Tooltip title="删除任务">
             <Button
-              onClick={() => handleTaskAction(task.id, "resume")}
+              danger
-              className="hover:bg-green-50 p-1 h-7 w-7"
               type="text"
-              icon={<Play className="w-4 h-4" />}
+              className="hover:bg-red-50 p-1 h-7 w-7"
-            ></Button>
+              icon={<DeleteOutlined />}
-          )}
+              onClick={() => {
+                Modal.confirm({
+                  title: `确认删除任务？`,
+                  content: `任务名：${task.name}`,
+                  okText: "删除",
+                  okType: "danger",
+                  cancelText: "取消",
+                  onOk: async () => {
+                    try {
+                      await deleteSynthesisTaskByIdUsingDelete(task.id);
+                      message.success("删除成功");
+                      loadTasks();
+                    } catch {
+                      message.error("删除失败");
+                    }
+                  },
+                });
+              }}
+            />
+          </Tooltip>
         </div>
       ),
     },
@@ -237,18 +220,15 @@ export default function SynthesisTaskTab() {
       <SearchControls
         searchTerm={searchQuery}
         onSearchChange={setSearchQuery}
-        searchPlaceholder="搜索任务名称或模板..."
+        searchPlaceholder="搜索任务名称..."
         filters={[
           {
             key: "status",
-            label: "状态",
+            label: "类型",
             options: [
-              { label: "全部状态", value: "all" },
+              { label: "全部类型", value: "all" },
-              { label: "等待中", value: "pending" },
+              { label: "问答对生成", value: "QA" },
-              { label: "运行中", value: "running" },
+              { label: "链式推理生成", value: "COT" },
-              { label: "已完成", value: "completed" },
-              { label: "失败", value: "failed" },
-              { label: "已暂停", value: "paused" },
             ],
           },
         ]}
@@ -259,13 +239,23 @@ export default function SynthesisTaskTab() {
         showFilters
         showViewToggle={false}
       />
-
       {/* 任务表格 */}
       <Card>
         <Table
           columns={taskColumns}
-          dataSource={sortedTasks}
+          dataSource={tasks}
           rowKey="id"
+          loading={loading}
+          pagination={{
+            current: page,
+            pageSize: pageSize,
+            total: total,
+            onChange: (p, ps) => {
+              setPage(p);
+              setPageSize(ps);
+            },
+            showSizeChanger: true,
+          }}
           scroll={{ x: "max-content" }}
           locale={{
             emptyText: (

frontend/src/pages/SynthesisTask/synthesis-api.ts

@@ -0,0 +1,37 @@
+import { get, post, del } from "@/utils/request";
+
+// 创建数据合成任务
+export function createSynthesisTaskUsingPost(data: unknown) {
+  return post("/api/synthesis/gen/task", data);
+}
+
+// 获取数据合成任务详情
+export function querySynthesisTaskByIdUsingGet(taskId: string) {
+  return get(`/api/synthesis/gen/task/${taskId}`);
+}
+
+// 分页查询数据合成任务列表
+export function querySynthesisTasksUsingGet(params: {
+  page?: number;
+  page_size?: number;
+  synthesis_type?: string;
+  status?: string;
+  name?: string;
+}) {
+  return get(`/api/synthesis/gen/tasks`, params as any);
+}
+
+// 删除整个数据合成任务
+export function deleteSynthesisTaskByIdUsingDelete(taskId: string) {
+  return del(`/api/synthesis/gen/task/${taskId}`);
+}
+
+// 分页查询某个任务下的文件任务列表
+export function querySynthesisFileTasksUsingGet(taskId: string, params: { page?: number; page_size?: number }) {
+  return get(`/api/synthesis/gen/task/${taskId}/files`, params as any);
+}
+
+// 获取不同合成类型对应的 Prompt
+export function getPromptByTypeUsingGet(synthType: string) {
+  return get(`/api/synthesis/gen/prompt`, { synth_type: synthType } as any);
+}

frontend/src/routes/routes.ts

@@ -40,6 +40,7 @@ import { withErrorBoundary } from "@/components/ErrorBoundary";
 import AgentPage from "@/pages/Agent/Agent.tsx";
 import RatioTaskDetail from "@/pages/RatioTask/Detail/RatioTaskDetail";
 import CleansingTemplateDetail from "@/pages/DataCleansing/Detail/TemplateDetail";
+import SynthFileTask from "@/pages/SynthesisTask/SynthFileTask.tsx";
 import EvaluationDetailPage from "@/pages/DataEvaluation/Detail/TaskDetail.tsx";
 
 const router = createBrowserRouter([
@@ -160,6 +161,7 @@ const router = createBrowserRouter([
             path: "create",
             Component: SynthesisTaskCreate,
           },
+            {path: ":id", Component: SynthFileTask},
         ],
       },
       {

runtime/datamate-python/app/module/generation/interface/__init__.py

@@ -1,8 +1,8 @@
 from fastapi import APIRouter
 
 router = APIRouter(
-    prefix="/synth",
+    prefix="/synthesis",
-    tags = ["synth"]
+    tags = ["synthesis"]
 )
 
 # Include sub-routers

runtime/datamate-python/app/module/generation/interface/generation_api.py

@@ -18,7 +18,14 @@ from app.db.session import get_db
 from app.module.generation.schema.generation import (
     CreateSynthesisTaskRequest,
     DataSynthesisTaskItem,
-    PagedDataSynthesisTaskResponse, SynthesisType)
+    PagedDataSynthesisTaskResponse,
+    SynthesisType,
+    DataSynthesisFileTaskItem,
+    PagedDataSynthesisFileTaskResponse,
+    DataSynthesisChunkItem,
+    PagedDataSynthesisChunkResponse,
+    SynthesisDataItem,
+)
 from app.module.generation.service.generation_service import GenerationService
 from app.module.generation.service.prompt import get_prompt
 from app.module.shared.schema import StandardResponse
@@ -219,19 +226,26 @@ async def delete_synthesis_task(
         data=None,
     )
 
+
 @router.delete("/task/{task_id}/{file_id}", response_model=StandardResponse[None])
 async def delete_synthesis_file_task(
     task_id: str,
     file_id: str,
     db: AsyncSession = Depends(get_db)
 ):
-    """删除数据合成任务中的文件任务"""
+    """删除数据合成任务中的文件任务，同时刷新任务表中的文件/切片数量"""
+    # 先获取任务和文件任务记录
+    task = await db.get(DataSynthesisInstance, task_id)
+    if not task:
+        raise HTTPException(status_code=404, detail="Synthesis task not found")
+
     file_task = await db.get(DataSynthesisFileInstance, file_id)
     if not file_task:
         raise HTTPException(status_code=404, detail="Synthesis file task not found")
 
     # 删除 SynthesisData（根据文件任务ID）
-    await db.execute(delete(SynthesisData).where(
+    await db.execute(
+        delete(SynthesisData).where(
             SynthesisData.synthesis_file_instance_id == file_id
         )
     )
@@ -243,11 +257,28 @@ async def delete_synthesis_file_task(
     )
 
     # 删除文件任务记录
-    await db.execute(delete(DataSynthesisFileInstance).where(
+    await db.execute(
+        delete(DataSynthesisFileInstance).where(
             DataSynthesisFileInstance.id == file_id
         )
     )
 
+    # 刷新任务级别统计字段：总文件数、总文本块数、已处理文本块数
+    if task.total_files and task.total_files > 0:
+        task.total_files -= 1
+        if task.total_files < 0:
+            task.total_files = 0
+
+    await db.commit()
+    await db.refresh(task)
+
+    return StandardResponse(
+        code=200,
+        message="success",
+        data=None,
+    )
+
+
 @router.get("/prompt", response_model=StandardResponse[str])
 async def get_prompt_by_type(
     synth_type: SynthesisType,
@@ -258,3 +289,157 @@ async def get_prompt_by_type(
         message="Success",
         data=prompt,
     )
+
+
+@router.get("/task/{task_id}/files", response_model=StandardResponse[PagedDataSynthesisFileTaskResponse])
+async def list_synthesis_file_tasks(
+    task_id: str,
+    page: int = 1,
+    page_size: int = 10,
+    db: AsyncSession = Depends(get_db),
+):
+    """分页获取某个数据合成任务下的文件任务列表"""
+    # 先校验任务是否存在
+    task = await db.get(DataSynthesisInstance, task_id)
+    if not task:
+        raise HTTPException(status_code=404, detail="Synthesis task not found")
+
+    base_query = select(DataSynthesisFileInstance).where(
+        DataSynthesisFileInstance.synthesis_instance_id == task_id
+    )
+
+    count_q = select(func.count()).select_from(base_query.subquery())
+    total = (await db.execute(count_q)).scalar_one()
+
+    if page < 1:
+        page = 1
+    if page_size < 1:
+        page_size = 10
+
+    result = await db.execute(
+        base_query.offset((page - 1) * page_size).limit(page_size)
+    )
+    rows = result.scalars().all()
+
+    file_items = [
+        DataSynthesisFileTaskItem(
+            id=row.id,
+            synthesis_instance_id=row.synthesis_instance_id,
+            file_name=row.file_name,
+            source_file_id=row.source_file_id,
+            target_file_location=row.target_file_location,
+            status=row.status,
+            total_chunks=row.total_chunks,
+            processed_chunks=row.processed_chunks,
+            created_at=row.created_at,
+            updated_at=row.updated_at,
+            created_by=row.created_by,
+            updated_by=row.updated_by,
+        )
+        for row in rows
+    ]
+
+    paged = PagedDataSynthesisFileTaskResponse(
+        content=file_items,
+        totalElements=total,
+        totalPages=(total + page_size - 1) // page_size,
+        page=page,
+        size=page_size,
+    )
+
+    return StandardResponse(
+        code=200,
+        message="Success",
+        data=paged,
+    )
+
+
+@router.get("/file/{file_id}/chunks", response_model=StandardResponse[PagedDataSynthesisChunkResponse])
+async def list_chunks_by_file(
+    file_id: str,
+    page: int = 1,
+    page_size: int = 10,
+    db: AsyncSession = Depends(get_db),
+):
+    """根据文件任务 ID 分页查询 chunk 记录"""
+    # 校验文件任务是否存在
+    file_task = await db.get(DataSynthesisFileInstance, file_id)
+    if not file_task:
+        raise HTTPException(status_code=404, detail="Synthesis file task not found")
+
+    base_query = select(DataSynthesisChunkInstance).where(
+        DataSynthesisChunkInstance.synthesis_file_instance_id == file_id
+    )
+
+    count_q = select(func.count()).select_from(base_query.subquery())
+    total = (await db.execute(count_q)).scalar_one()
+
+    if page < 1:
+        page = 1
+    if page_size < 1:
+        page_size = 10
+
+    result = await db.execute(
+        base_query.order_by(DataSynthesisChunkInstance.chunk_index.asc())
+        .offset((page - 1) * page_size)
+        .limit(page_size)
+    )
+    rows = result.scalars().all()
+
+    chunk_items = [
+        DataSynthesisChunkItem(
+            id=row.id,
+            synthesis_file_instance_id=row.synthesis_file_instance_id,
+            chunk_index=row.chunk_index,
+            chunk_content=row.chunk_content,
+            chunk_metadata=getattr(row, "chunk_metadata", None),
+        )
+        for row in rows
+    ]
+
+    paged = PagedDataSynthesisChunkResponse(
+        content=chunk_items,
+        totalElements=total,
+        totalPages=(total + page_size - 1) // page_size,
+        page=page,
+        size=page_size,
+    )
+
+    return StandardResponse(
+        code=200,
+        message="Success",
+        data=paged,
+    )
+
+
+@router.get("/chunk/{chunk_id}/data", response_model=StandardResponse[list[SynthesisDataItem]])
+async def list_synthesis_data_by_chunk(
+    chunk_id: str,
+    db: AsyncSession = Depends(get_db),
+):
+    """根据 chunk ID 查询所有合成结果数据"""
+    # 可选：校验 chunk 是否存在
+    chunk = await db.get(DataSynthesisChunkInstance, chunk_id)
+    if not chunk:
+        raise HTTPException(status_code=404, detail="Chunk not found")
+
+    result = await db.execute(
+        select(SynthesisData).where(SynthesisData.chunk_instance_id == chunk_id)
+    )
+    rows = result.scalars().all()
+
+    items = [
+        SynthesisDataItem(
+            id=row.id,
+            data=row.data,
+            synthesis_file_instance_id=row.synthesis_file_instance_id,
+            chunk_instance_id=row.chunk_instance_id,
+        )
+        for row in rows
+    ]
+
+    return StandardResponse(
+        code=200,
+        message="Success",
+        data=items,
+    )

runtime/datamate-python/app/module/generation/schema/generation.py

@@ -70,6 +70,67 @@ class PagedDataSynthesisTaskResponse(BaseModel):
     page: int
     size: int
 
+
+class DataSynthesisFileTaskItem(BaseModel):
+    """数据合成任务下的文件任务项"""
+    id: str
+    synthesis_instance_id: str
+    file_name: str
+    source_file_id: str
+    target_file_location: str
+    status: Optional[str] = None
+    total_chunks: int
+    processed_chunks: int
+    created_at: Optional[datetime] = None
+    updated_at: Optional[datetime] = None
+    created_by: Optional[str] = None
+    updated_by: Optional[str] = None
+
+    class Config:
+        orm_mode = True
+
+
+class PagedDataSynthesisFileTaskResponse(BaseModel):
+    """分页数据合成任务文件任务响应"""
+    content: List[DataSynthesisFileTaskItem]
+    totalElements: int
+    totalPages: int
+    page: int
+    size: int
+
+
+class DataSynthesisChunkItem(BaseModel):
+    """数据合成文件下的 chunk 记录"""
+    id: str
+    synthesis_file_instance_id: str
+    chunk_index: Optional[int] = None
+    chunk_content: Optional[str] = None
+    chunk_metadata: Optional[Dict[str, Any]] = None
+
+    class Config:
+        orm_mode = True
+
+
+class PagedDataSynthesisChunkResponse(BaseModel):
+    """分页 chunk 列表响应"""
+    content: List[DataSynthesisChunkItem]
+    totalElements: int
+    totalPages: int
+    page: int
+    size: int
+
+
+class SynthesisDataItem(BaseModel):
+    """合成结果数据项"""
+    id: str
+    data: Optional[Dict[str, Any]] = None
+    synthesis_file_instance_id: str
+    chunk_instance_id: str
+
+    class Config:
+        orm_mode = True
+
+
 class ChatRequest(BaseModel):
     """聊天请求参数"""
     model_id: str

runtime/datamate-python/app/module/generation/service/generation_service.py

@@ -168,11 +168,11 @@ class GenerationService:
             self.db.add(chunk_record)
 
         # 更新文件任务的分块数量
-        file_task.chunk_count = len(chunks)
+        file_task.total_chunks = len(chunks)
         file_task.status = "processing"
 
-        await self.db.refresh(file_task)
         await self.db.commit()
+        await self.db.refresh(file_task)
 
     async def _invoke_llm_for_chunks(
         self,

runtime/datamate-python/app/module/generation/service/prompt.py

@@ -1,7 +1,6 @@
 from app.module.generation.schema.generation import SynthesisType
 
-QA_PROMPT="""
+QA_PROMPT="""# 角色
-# 角色
 你是一位专业的AI助手，擅长从给定的文本中提取关键信息并创建用于教学和测试的问答对。
 
 # 任务
@@ -11,7 +10,7 @@ QA_PROMPT="""
 {document}
 
 # 要求与指令
-1.  **问题类型**：生成{synthesis_count - 1}-{synthesis_count + 1}个问答对。问题类型应多样化，包括但不限于：
+1.  **问题类型**：生成 {synthesis_count} 个左右的问答对。问题类型应多样化，包括但不限于：
     *   **事实性**：基于文本中明确提到的事实。
     *   **理解性**：需要理解上下文和概念。
     *   **归纳性**：需要总结或归纳多个信息点。
@@ -30,8 +29,7 @@ QA_PROMPT="""
 """
 
 
-COT_PROMPT="""
+COT_PROMPT="""# 角色
-# 角色
 你是一位专业的数据合成专家，擅长基于给定的原始文档和 COT（Chain of Thought，思维链）逻辑，生成高质量、符合实际应用场景的 COT 数据。COT 数据需包含清晰的问题、逐步推理过程和最终结论，能完整还原解决问题的思考路径。
 
 # 任务
@@ -41,7 +39,7 @@ COT_PROMPT="""
 {document}
 
 # 要求与指令
-1. **数量要求**：生成 {min\_count}-{max\_count} 条 COT 数据（min\_count={synthesis\_count-1}，max\_count={synthesis\_count+1}）。
+1. **数量要求**：生成 {synthesis_count} 条左右的 COT 数据。
 2. **内容要求**：
     * 每条 COT 数据需包含 “问题”“思维链推理”“最终结论” 三部分，逻辑闭环，推理步骤清晰、连贯，不跳跃关键环节。
     * 问题需基于文档中的事实信息、概念关联或逻辑疑问，是读完文档后自然产生的有价值问题（避免无意义或过于简单的问题）。