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

feat(annotation): 添加分段标注统计和进度跟踪功能

Browse Source 
- 新增 SegmentStats 类型定义用于分段统计
- 实现分段标注进度计算和缓存机制
- 添加标注任务状态判断逻辑支持分段模式
- 集成分段统计数据显示到任务列表界面
- 实现分段总数自动计算和验证功能
- 扩展标注状态枚举支持进行中标注状态
- 优化任务选择逻辑基于分段完成状态
- 添加分段统计数据预加载和同步机制
This commit is contained in:
Jerry Yan
2026-01-31 15:42:04 +08:00
parent 3e0a15ac8e
commit 33cf65c9f8
5 changed files with 265 additions and 31 deletions
Download Patch File Download Diff File Expand all files Collapse all files

173
frontend/src/pages/DataAnnotation/Annotate/LabelStudioTextEditor.tsx
View File

@@ -28,6 +28,7 @@ type EditorTaskListItem = {
hasAnnotation: boolean;
annotationUpdatedAt?: string | null;
annotationStatus?: AnnotationResultStatus | null;
segmentStats?: SegmentStats;
};
type LsfMessage = {
@@ -45,6 +46,11 @@ type SegmentInfo = {
chunkIndex: number;
};
type SegmentStats = {
done: number;
total: number;
};
type ApiResponse<T> = {
code?: number;
message?: string;
@@ -136,6 +142,16 @@ const isAnnotationResultEmpty = (annotation?: Record<string, unknown>) => {
};
const resolveTaskStatusMeta = (item: EditorTaskListItem) => {
const segmentSummary = resolveSegmentSummary(item);
if (segmentSummary) {
if (segmentSummary.done >= segmentSummary.total) {
return { text: "已标注", type: "success" as const };
}
if (segmentSummary.done > 0) {
return { text: "标注中", type: "warning" as const };
}
return { text: "未标注", type: "secondary" as const };
}
if (!item.hasAnnotation) {
return { text: "未标注", type: "secondary" as const };
}
@@ -145,6 +161,9 @@ const resolveTaskStatusMeta = (item: EditorTaskListItem) => {
if (item.annotationStatus === AnnotationResultStatus.NOT_APPLICABLE) {
return { text: NOT_APPLICABLE_LABEL, type: "warning" as const };
}
if (item.annotationStatus === AnnotationResultStatus.IN_PROGRESS) {
return { text: "标注中", type: "warning" as const };
}
return { text: "已标注", type: "success" as const };
};
@@ -184,6 +203,25 @@ const buildAnnotationSnapshot = (annotation?: Record<string, unknown>) => {
const buildSnapshotKey = (fileId: string, segmentIndex?: number) =>
`${fileId}::${segmentIndex ?? "full"}`;
const buildSegmentStats = (segmentList?: SegmentInfo[] | null): SegmentStats | null => {
if (!Array.isArray(segmentList) || segmentList.length === 0) return null;
const total = segmentList.length;
const done = segmentList.reduce((count, seg) => count + (seg.hasAnnotation ? 1 : 0), 0);
return { done, total };
};
const normalizeSegmentStats = (stats?: SegmentStats | null): SegmentStats | null => {
if (!stats) return null;
const total = Number(stats.total);
const done = Number(stats.done);
if (!Number.isFinite(total) || total <= 0) return null;
const safeDone = Math.min(Math.max(done, 0), total);
return { done: safeDone, total };
};
const resolveSegmentSummary = (item: EditorTaskListItem) =>
normalizeSegmentStats(item.segmentStats);
const mergeTaskItems = (base: EditorTaskListItem[], next: EditorTaskListItem[]) => {
if (next.length === 0) return base;
const seen = new Set(base.map((item) => item.fileId));
@@ -234,6 +272,9 @@ export default function LabelStudioTextEditor() {
const exportCheckSeqRef = useRef(0);
const savedSnapshotsRef = useRef<Record<string, string>>({});
const pendingAutoAdvanceRef = useRef(false);
const segmentStatsCacheRef = useRef<Record<string, SegmentStats>>({});
const segmentStatsSeqRef = useRef(0);
const segmentStatsLoadingRef = useRef<Set<string>>(new Set());
const [loadingProject, setLoadingProject] = useState(true);
const [loadingTasks, setLoadingTasks] = useState(false);
@@ -276,6 +317,70 @@ export default function LabelStudioTextEditor() {
win.postMessage({ type, payload }, origin);
}, [origin]);
const applySegmentStats = useCallback((fileId: string, stats: SegmentStats | null) => {
if (!fileId) return;
const normalized = normalizeSegmentStats(stats);
setTasks((prev) =>
prev.map((item) =>
item.fileId === fileId
? { ...item, segmentStats: normalized || undefined }
: item
)
);
}, []);
const updateSegmentStatsCache = useCallback((fileId: string, stats: SegmentStats | null) => {
if (!fileId) return;
const normalized = normalizeSegmentStats(stats);
if (normalized) {
segmentStatsCacheRef.current[fileId] = normalized;
} else {
delete segmentStatsCacheRef.current[fileId];
}
applySegmentStats(fileId, normalized);
}, [applySegmentStats]);
const fetchSegmentStatsForFile = useCallback(async (fileId: string, seq: number) => {
if (!projectId || !fileId) return;
if (segmentStatsCacheRef.current[fileId] || segmentStatsLoadingRef.current.has(fileId)) return;
segmentStatsLoadingRef.current.add(fileId);
try {
const resp = (await getEditorTaskUsingGet(projectId, fileId, {
segmentIndex: 0,
})) as ApiResponse<EditorTaskResponse>;
if (segmentStatsSeqRef.current !== seq) return;
const data = resp?.data;
if (!data?.segmented) return;
const stats = buildSegmentStats(data.segments);
if (!stats) return;
segmentStatsCacheRef.current[fileId] = stats;
applySegmentStats(fileId, stats);
} catch (e) {
console.error(e);
} finally {
segmentStatsLoadingRef.current.delete(fileId);
}
}, [applySegmentStats, projectId]);
const prefetchSegmentStats = useCallback((items: EditorTaskListItem[]) => {
if (!projectId) return;
const fileIds = items
.map((item) => item.fileId)
.filter((fileId) => fileId && !segmentStatsCacheRef.current[fileId]);
if (fileIds.length === 0) return;
const seq = segmentStatsSeqRef.current;
let cursor = 0;
const workerCount = Math.min(3, fileIds.length);
const runWorker = async () => {
while (cursor < fileIds.length && segmentStatsSeqRef.current === seq) {
const fileId = fileIds[cursor];
cursor += 1;
await fetchSegmentStatsForFile(fileId, seq);
}
};
void Promise.all(Array.from({ length: workerCount }, () => runWorker()));
}, [fetchSegmentStatsForFile, projectId]);
const confirmEmptyAnnotationStatus = useCallback(() => {
return new Promise<AnnotationResultStatus | null>((resolve) => {
let resolved = false;
@@ -327,8 +432,13 @@ export default function LabelStudioTextEditor() {
}, [message, projectId]);
const updateTaskSelection = useCallback((items: EditorTaskListItem[]) => {
const isCompleted = (item: EditorTaskListItem) => {
const summary = resolveSegmentSummary(item);
if (summary) return summary.done >= summary.total;
return item.hasAnnotation;
};
const defaultFileId =
items.find((item) => !item.hasAnnotation)?.fileId || items[0]?.fileId || "";
items.find((item) => !isCompleted(item))?.fileId || items[0]?.fileId || "";
setSelectedFileId((prev) => {
if (prev && items.some((item) => item.fileId === prev)) return prev;
return defaultFileId;
@@ -385,6 +495,9 @@ export default function LabelStudioTextEditor() {
if (mode === "reset") {
prefetchSeqRef.current += 1;
setPrefetching(false);
segmentStatsSeqRef.current += 1;
segmentStatsCacheRef.current = {};
segmentStatsLoadingRef.current = new Set();
}
if (mode === "append") {
setLoadingMore(true);
@@ -469,13 +582,16 @@ export default function LabelStudioTextEditor() {
? resolveSegmentIndex(data.currentSegmentIndex) ?? 0
: undefined;
if (data?.segmented) {
const stats = buildSegmentStats(data.segments);
setSegmented(true);
setSegments(data.segments || []);
setCurrentSegmentIndex(segmentIndex ?? 0);
updateSegmentStatsCache(fileId, stats);
} else {
setSegmented(false);
setSegments([]);
setCurrentSegmentIndex(0);
updateSegmentStatsCache(fileId, null);
}
const taskData = {
@@ -535,7 +651,7 @@ export default function LabelStudioTextEditor() {
} finally {
if (seq === initSeqRef.current) setLoadingTaskDetail(false);
}
}, [iframeReady, message, postToIframe, project, projectId]);
}, [iframeReady, message, postToIframe, project, projectId, updateSegmentStatsCache]);
const advanceAfterSave = useCallback(async (fileId: string, segmentIndex?: number) => {
if (!fileId) return;
@@ -643,13 +759,13 @@ export default function LabelStudioTextEditor() {
// 分段模式下更新当前段落的标注状态
if (segmented && segmentIndex !== undefined) {
setSegments((prev) =>
prev.map((seg) =>
seg.idx === segmentIndex
? { ...seg, hasAnnotation: true }
: seg
)
const nextSegments = segments.map((seg) =>
seg.idx === segmentIndex
? { ...seg, hasAnnotation: true }
: seg
);
setSegments(nextSegments);
updateSegmentStatsCache(String(fileId), buildSegmentStats(nextSegments));
}
if (options?.autoAdvance) {
await advanceAfterSave(String(fileId), segmentIndex);
@@ -669,8 +785,10 @@ export default function LabelStudioTextEditor() {
message,
projectId,
segmented,
segments,
selectedFileId,
tasks,
updateSegmentStatsCache,
]);
const requestExportForCheck = useCallback(() => {
@@ -834,6 +952,9 @@ export default function LabelStudioTextEditor() {
setSegments([]);
setCurrentSegmentIndex(0);
savedSnapshotsRef.current = {};
segmentStatsSeqRef.current += 1;
segmentStatsCacheRef.current = {};
segmentStatsLoadingRef.current = new Set();
if (exportCheckRef.current?.timer) {
window.clearTimeout(exportCheckRef.current.timer);
}
@@ -847,6 +968,12 @@ export default function LabelStudioTextEditor() {
loadTasks({ mode: "reset" });
}, [project?.supported, loadTasks]);
useEffect(() => {
if (!segmented) return;
if (tasks.length === 0) return;
prefetchSegmentStats(tasks);
}, [prefetchSegmentStats, segmented, tasks]);
useEffect(() => {
if (!selectedFileId) return;
initEditorForFile(selectedFileId);
@@ -1097,6 +1224,7 @@ export default function LabelStudioTextEditor() {
dataSource={tasks}
loadMore={loadMoreNode}
renderItem={(item) => {
const segmentSummary = resolveSegmentSummary(item);
const statusMeta = resolveTaskStatusMeta(item);
return (
<List.Item
@@ -1110,18 +1238,25 @@ export default function LabelStudioTextEditor() {
onClick={() => setSelectedFileId(item.fileId)}
>
<div className="flex flex-col w-full gap-1">
<Typography.Text ellipsis style={{ fontSize: 13 }}>
{item.fileName}
</Typography.Text>
<div className="flex items-center justify-between">
<Typography.Text type={statusMeta.type} style={{ fontSize: 11 }}>
{statusMeta.text}
<Typography.Text ellipsis style={{ fontSize: 13 }}>
{item.fileName}
</Typography.Text>
{item.annotationUpdatedAt && (
<Typography.Text type="secondary" style={{ fontSize: 10 }}>
{item.annotationUpdatedAt}
</Typography.Text>
)}
<div className="flex items-center justify-between">
<div className="flex items-center gap-2">
<Typography.Text type={statusMeta.type} style={{ fontSize: 11 }}>
{statusMeta.text}
</Typography.Text>
{segmentSummary && (
<Typography.Text type="secondary" style={{ fontSize: 10 }}>
已标注 {segmentSummary.done}/{segmentSummary.total}
</Typography.Text>
)}
</div>
{item.annotationUpdatedAt && (
<Typography.Text type="secondary" style={{ fontSize: 10 }}>
{item.annotationUpdatedAt}
</Typography.Text>
)}
</div>
</div>
</List.Item>

1
frontend/src/pages/DataAnnotation/annotation.model.ts
View File

@@ -10,6 +10,7 @@ export enum AnnotationTaskStatus {
export enum AnnotationResultStatus {
ANNOTATED = "ANNOTATED",
IN_PROGRESS = "IN_PROGRESS",
NO_ANNOTATION = "NO_ANNOTATION",
NOT_APPLICABLE = "NOT_APPLICABLE",
}