feat(text-annotation): improve manual text annotation and enable visibility in Label Studio (#235)

- Optimize manual text annotation workflow
- Make text content visible and editable within Label Studio

Author: o0Shark0o

frontend/src/components/business/DatasetFileTransfer.tsx

@@ -51,7 +51,7 @@ const DatasetFileTransfer: React.FC<DatasetFileTransferProps> = ({
   selectedFilesMap,
   onSelectedFilesChange,
   onDatasetSelect,
-  datasetTypeFilter = DatasetType.TEXT,
+  datasetTypeFilter,
   ...props
 }) => {
   const [datasets, setDatasets] = React.useState<Dataset[]>([]);
@@ -85,6 +85,7 @@ const DatasetFileTransfer: React.FC<DatasetFileTransferProps> = ({
       page: datasetPagination.current,
       size: datasetPagination.pageSize,
       keyword: datasetSearch,
+      // 仅在显式传入过滤类型时才按类型过滤；否则后端返回所有类型
       type: datasetTypeFilter,
     });
     setDatasets(data.content.map(mapDataset) || []);

frontend/src/pages/DataAnnotation/Create/components/CreateAnnotationTaskDialog.tsx

@@ -183,16 +183,8 @@ export default function CreateAnnotationTask({
       // 手动标注也支持跨数据集、精确到文件的选择
       const selectedFiles = Object.values(selectedFilesMap) as any[];
 
-      const imageExtensions = [".jpg", ".jpeg", ".png", ".bmp", ".gif", ".tiff", ".webp"];
-      const imageFileIds = selectedFiles
-        .filter((file) => {
-          const ext = file.fileName?.toLowerCase().match(/\.[^.]+$/)?.[0] || "";
-          return imageExtensions.includes(ext);
-        })
-        .map((file) => file.id);
-
-      if (imageFileIds.length === 0) {
-        message?.error?.("请至少选择一个图像文件");
+      if (selectedFiles.length === 0) {
+        message?.error?.("请至少选择一个文件");
         setSubmitting(false);
         return;
       }
@@ -213,7 +205,7 @@ export default function CreateAnnotationTask({
         description: values.description,
         datasetId: effectiveDatasetId,
         templateId: values.templateId,
-        fileIds: imageFileIds,
+        fileIds: selectedFiles.map((file) => file.id),
       };
 
       await createAnnotationTaskUsingPost(requestData);
@@ -326,8 +318,8 @@ export default function CreateAnnotationTask({
             label: "手动标注",
             children: (
               <Form form={manualForm} layout="vertical">
-                {/* 选择数据集和图像文件（支持多数据集、多文件） */}
-                <Form.Item label="选择数据集和图像文件" required>
+                {/* 选择数据集和文件（支持多数据集、多文件） */}
+                <Form.Item label="选择数据集和文件" required>
                   <DatasetFileTransfer
                     open
                     selectedFilesMap={selectedFilesMap}
@@ -346,12 +338,11 @@ export default function CreateAnnotationTask({
                         manualForm.setFieldsValue({ name: defaultName });
                       }
                     }}
-                    datasetTypeFilter={DatasetType.IMAGE}
                   />
                   {selectedDataset && (
                     <div className="mt-2 p-2 bg-blue-50 rounded border border-blue-200 text-xs">
                       当前数据集：<span className="font-medium">{selectedDataset.name}</span> - 已选择
-                      <span className="font-medium text-blue-600"> {imageFileCount} </span>个图像文件
+                      <span className="font-medium text-blue-600"> {Object.keys(selectedFilesMap).length} </span>个文件
                     </div>
                   )}
                 </Form.Item>

frontend/src/pages/KnowledgeBase/components/AddDataDialog.tsx

@@ -273,6 +273,7 @@ export default function AddDataDialog({ knowledgeBase, onDataAdded }) {
               open={open}
               selectedFilesMap={selectedFilesMap}
               onSelectedFilesChange={setSelectedFilesMap}
+              datasetTypeFilter={DatasetType.TEXT}
             />
           )}
 

frontend/src/pages/SynthesisTask/CreateTask.tsx

@@ -1,5 +1,6 @@
 import { useEffect, useState } from "react";
 import type { Dataset, DatasetFile } from "@/pages/DataManagement/dataset.model";
+import { DatasetType } from "@/pages/DataManagement/dataset.model";
 import { Steps, Card, Select, Input, Button, Form, message, Tag, Tooltip, InputNumber } from "antd";
 import { Eye, ArrowLeft, ArrowRight, Play, Search, Sparkles, Brain, Layers } from "lucide-react";
 import { Link, useNavigate } from "react-router";
@@ -304,7 +305,7 @@ export default function SynthesisTaskCreate() {
             <DatasetFileTransfer open selectedFilesMap={selectedMap} onSelectedFilesChange={setSelectedMap} onDatasetSelect={(dataset) => {
               setSelectedDataset(dataset);
               form.setFieldsValue({ sourceDataset: dataset?.id ?? "" });
-            }} />
+            }} datasetTypeFilter={DatasetType.TEXT} />
             {selectedDataset && (
               <div className="mt-4 p-3 bg-gray-50 rounded border text-xs text-gray-600">
                 当前数据集：<span className="font-medium text-gray-900">{selectedDataset.name}</span>

runtime/datamate-python/app/module/annotation/service/sync.py

@@ -1,4 +1,6 @@
 from typing import Optional, List, Dict, Any, Tuple, Set
+import os
+
 from app.module.dataset import DatasetManagementService
 from sqlalchemy import update, select
 from app.db.models import DatasetFiles
@@ -52,14 +54,38 @@ def _determine_data_type(self, file_type: str) -> str:
     def _build_task_data(self, file_info: Any, dataset_id: str) -> dict:
         """构建Label Studio任务数据"""
         data_type = self._determine_data_type(file_info.fileType)
-        
-        # 替换文件路径前缀
-        file_path = file_info.filePath.removeprefix(settings.dm_file_path_prefix)
-        file_path = settings.label_studio_file_path_prefix + file_path
-        
+
+        # 默认仍然走 Label Studio 本地文件 URL
+        # 先替换文件路径前缀，构造 /data/local-files/?d=/... 形式
+        relative_path = file_info.filePath.removeprefix(settings.dm_file_path_prefix)
+        ls_file_url = settings.label_studio_file_path_prefix + relative_path
+
+        data_value: Any = ls_file_url
+
+        # 对于纯文本文件（例如 .txt），支持直接把文件内容写入到 data.text，
+        # 这样在 Label Studio 里会直接显示文本内容，而不是 URL。
+        if data_type == "text":
+            try:
+                _, ext = os.path.splitext(file_info.filePath)
+                ext = ext.lower()
+
+                # 目前只对 .txt 做内联，其他如 pdf/doc 仍然使用 URL
+                if ext == ".txt":
+                    with open(file_info.filePath, "r", encoding="utf-8", errors="ignore") as f:
+                        content = f.read()
+                    if content:
+                        data_value = content
+            except Exception as e:
+                # 读取失败时退回到原来的 URL 形式，避免中断同步流程
+                logger.warning(
+                    "Failed to inline text content for file %s: %s",
+                    getattr(file_info, "filePath", "<unknown>"),
+                    str(e),
+                )
+
         return {
             "data": {
-                f"{data_type}": file_path,
+                f"{data_type}": data_value,
                 "file_path": file_info.filePath,
                 "file_id": file_info.id,
                 "original_name": file_info.originalName,