YG-Soft/YG_FT
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(data-process): 启用智能文档清理预览

Browse Source
This commit is contained in:
caoxiaozhu
2026-07-24 15:06:01 +08:00
parent 3266a6fc09
commit 4544483fc5
3 changed files with 4 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
frontend/scripts/regression-data-process-wizard.mjs
View File

@@ -110,7 +110,7 @@ assert.match(
/const \{ buildPreviewsByFile \} = useDataProcessPreviewBuild\(\)/, /const \{ buildPreviewsByFile \} = useDataProcessPreviewBuild\(\)/,
'父页面没有通过独立 composable 构建逐文件预览', '父页面没有通过独立 composable 构建逐文件预览',
) )
assert.match(viewSource, /backend-pipeline-v3/, '切分管线版本未升级，旧预览缓存可能被误用') assert.match(viewSource, /backend-pipeline-v4/, '切分管线版本未升级，旧预览缓存可能被误用')
assert.match(viewSource, /getDataProcessPreview\(taskId\.value,\s*\{ page:\s*1, page_size:\s*500 \}\)/, '预览构建后没有分页读取后端数据') assert.match(viewSource, /getDataProcessPreview\(taskId\.value,\s*\{ page:\s*1, page_size:\s*500 \}\)/, '预览构建后没有分页读取后端数据')
assert.doesNotMatch(viewSource, /buildPreviewItems\(/, '创建向导仍在本地构建集成预览数据') assert.doesNotMatch(viewSource, /buildPreviewItems\(/, '创建向导仍在本地构建集成预览数据')
@@ -501,7 +501,7 @@ const smartOptionValues = [...smartOptionsSource.matchAll(/^\s*'([^']+)',?$/gm)]
assert.deepEqual(smartOptionValues, expectedSmartPreprocessOptions, '智能预处理内部值与后端语义不一致') assert.deepEqual(smartOptionValues, expectedSmartPreprocessOptions, '智能预处理内部值与后端语义不一致')
assert.equal(new Set(smartOptionValues).size, smartOptionValues.length, '非结构化智能预处理 value 必须唯一') assert.equal(new Set(smartOptionValues).size, smartOptionValues.length, '非结构化智能预处理 value 必须唯一')
for (const descriptionPart of [ for (const descriptionPart of [
'清理无效内容', '清理页眉页脚、页码、目录和无效内容',
'感知文档结构', '感知文档结构',
'合并短块', '合并短块',
'预过滤低质量内容', '预过滤低质量内容',

2
frontend/src/views/data-process/DataProcessCreateView.vue
View File

@@ -57,7 +57,7 @@ const generationModels = computed(() => modelList.value)
const taskSetupRef = ref<InstanceType<typeof TaskSetupStep>>() const taskSetupRef = ref<InstanceType<typeof TaskSetupStep>>()
const modelSelectionRef = ref<InstanceType<typeof ModelSelectionStep>>() const modelSelectionRef = ref<InstanceType<typeof ModelSelectionStep>>()
const confirmDialogRef = ref<InstanceType<typeof AppConfirmDialog>>() const confirmDialogRef = ref<InstanceType<typeof AppConfirmDialog>>()
const PREVIEW_MODEL_VERSION = 'backend-pipeline-v3' const PREVIEW_MODEL_VERSION = 'backend-pipeline-v4'
const WIZARD_STEPS = [ const WIZARD_STEPS = [
{ id: 'create', title: '创建任务', desc: '填写任务信息与处理配置' }, { id: 'create', title: '创建任务', desc: '填写任务信息与处理配置' },
{ id: 'model', title: '大模型选择', desc: '选择生成模型并设置输出要求' }, { id: 'model', title: '大模型选择', desc: '选择生成模型并设置输出要求' },

2
frontend/src/views/data-process/create/UnstructuredOptionsPanel.vue
View File

@@ -128,7 +128,7 @@ defineExpose({ revealValidation })
/> />
<span class="preprocess-option-copy"> <span class="preprocess-option-copy">
<strong>智能预处理</strong> <strong>智能预处理</strong>
<small>清理无效内容、感知文档结构、合并短块、预过滤低质量内容、近重复去重，并通过重叠保护上下文</small> <small>清理页眉页脚、页码、目录和无效内容，感知文档结构、合并短块、预过滤低质量内容、近重复去重，并通过重叠保护上下文</small>
</span> </span>
</label> </label>