feat: add ingestion pipeline children delimiters configs (#11979)

### What problem does this PR solve?

Add children delimiters for Ingestion pipeline config

### Type of change

- [x] New Feature (non-breaking change which adds functionality)
This commit is contained in:
Jimmy Ben Klieve
2025-12-17 11:18:54 +08:00
committed by GitHub
parent 30019dab9f
commit 2595644dfd
10 changed files with 218 additions and 3 deletions

View File

@ -2,7 +2,8 @@ import { DelimiterInput } from '@/components/delimiter-form-field';
import { RAGFlowFormItem } from '@/components/ragflow-form';
import { SliderInputFormField } from '@/components/slider-input-form-field';
import { BlockButton, Button } from '@/components/ui/button';
import { Form } from '@/components/ui/form';
import { Form, FormControl, FormField, FormItem } from '@/components/ui/form';
import { Switch } from '@/components/ui/switch';
import { zodResolver } from '@hookform/resolvers/zod';
import { Trash2 } from 'lucide-react';
import { memo } from 'react';
@ -26,6 +27,12 @@ export const FormSchema = z.object({
value: z.string().optional(),
}),
),
enable_children: z.boolean(),
children_delimiters: z.array(
z.object({
value: z.string().optional(),
}),
),
overlapped_percent: z.number(), // 0.0 - 0.3 , 0% - 30%
});
@ -46,6 +53,11 @@ const SplitterForm = ({ node }: INextOperatorForm) => {
control: form.control,
});
const childrenDelimiters = useFieldArray({
name: 'children_delimiters',
control: form.control,
});
useWatchFormChange(node?.id, form);
return (
@ -90,6 +102,59 @@ const SplitterForm = ({ node }: INextOperatorForm) => {
<BlockButton onClick={() => append({ value: '\n' })}>
{t('common.add')}
</BlockButton>
<fieldset>
<div className="mb-2 flex justify-between items-center gap-1">
<span>{t('flow.enableChildrenDelimiters')}</span>
<FormField
control={form.control}
name="enable_children"
render={({ field: { value, onChange, ...restProps } }) => (
<FormItem>
<FormControl>
<Switch
checked={value}
onCheckedChange={onChange}
{...restProps}
/>
</FormControl>
</FormItem>
)}
/>
</div>
{form.getValues('enable_children') && (
<div className="space-y-4">
{childrenDelimiters.fields.map((field, index) => (
<div key={field.id} className="flex items-center gap-2">
<RAGFlowFormItem
name={`children_delimiters.${index}.value`}
label="children_delimiter"
labelClassName="!hidden"
className="flex-auto space-y-0"
>
<DelimiterInput className="!m-0"></DelimiterInput>
</RAGFlowFormItem>
<Button
type="button"
variant="ghost"
onClick={() => childrenDelimiters.remove(index)}
>
<Trash2 />
</Button>
</div>
))}
<BlockButton
onClick={() => childrenDelimiters.append({ value: '\n' })}
>
{t('common.add')}
</BlockButton>
</div>
)}
</fieldset>
</FormWrapper>
<div className="p-5">
<Output list={outputList}></Output>

View File

@ -288,6 +288,11 @@ function transformSplitterParams(params: SplitterFormSchemaType) {
...params,
overlapped_percent: Number(params.overlapped_percent) / 100,
delimiters: transformObjectArrayToPureArray(params.delimiters, 'value'),
// Unset children delimiters if this option is not enabled
children_delimiters: params.enable_children
? transformObjectArrayToPureArray(params.children_delimiters, 'value')
: [],
};
}
@ -713,7 +718,7 @@ export function convertToObjectArray<T extends string | number | boolean>(
/**
* convert the following object into a list
*
*
* {
"product_related": {
"description": "The question is about product usage, appearance and how it works.",

View File

@ -2,6 +2,7 @@ import {
AutoKeywordsFormField,
AutoQuestionsFormField,
} from '@/components/auto-keywords-form-field';
import { ChildrenDelimiterForm } from '@/components/children-delimiter-form';
import { DelimiterFormField } from '@/components/delimiter-form-field';
import { ExcelToHtmlFormField } from '@/components/excel-to-html-form-field';
import { LayoutRecognizeFormField } from '@/components/layout-recognize-form-field';
@ -21,6 +22,7 @@ export function NaiveConfiguration() {
<MinerUOptionsFormField></MinerUOptionsFormField>
<MaxTokenNumberFormField initialValue={512}></MaxTokenNumberFormField>
<DelimiterFormField></DelimiterFormField>
<ChildrenDelimiterForm />
<EnableTocToggle />
<OverlappedPercent />
</ConfigurationFormContainer>

View File

@ -24,6 +24,8 @@ export const formSchema = z
layout_recognize: z.string(),
chunk_token_num: z.number(),
delimiter: z.string(),
enable_children: z.boolean(),
children_delimiter: z.string(),
auto_keywords: z.number().optional(),
auto_questions: z.number().optional(),
html4excel: z.boolean(),

View File

@ -63,6 +63,8 @@ export default function DatasetSettings() {
layout_recognize: DocumentType.DeepDOC,
chunk_token_num: 512,
delimiter: `\n`,
enable_children: false,
children_delimiter: `\n`,
auto_keywords: 0,
auto_questions: 0,
html4excel: false,

View File

@ -67,6 +67,13 @@ export function SavingButton() {
await saveKnowledgeConfiguration({
kb_id,
...values,
parser_config: {
...values.parser_config,
// Unset children delimiter if this option is not enabled
children_delimiter: values.parser_config.enable_children
? values.parser_config.children_delimiter
: null,
},
});
})();
}