feat: knowledge pipeline (#25360)

Signed-off-by: -LAN- <laipz8200@outlook.com>
Co-authored-by: twwu <twwu@dify.ai>
Co-authored-by: crazywoola <100913391+crazywoola@users.noreply.github.com>
Co-authored-by: jyong <718720800@qq.com>
Co-authored-by: Wu Tianwei <30284043+WTW0313@users.noreply.github.com>
Co-authored-by: QuantumGhost <obelisk.reg+git@gmail.com>
Co-authored-by: lyzno1 <yuanyouhuilyz@gmail.com>
Co-authored-by: quicksand <quicksandzn@gmail.com>
Co-authored-by: Jyong <76649700+JohnJyong@users.noreply.github.com>
Co-authored-by: lyzno1 <92089059+lyzno1@users.noreply.github.com>
Co-authored-by: zxhlyh <jasonapring2015@outlook.com>
Co-authored-by: Yongtao Huang <yongtaoh2022@gmail.com>
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: Joel <iamjoel007@gmail.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: nite-knite <nkCoding@gmail.com>
Co-authored-by: Hanqing Zhao <sherry9277@gmail.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Harry <xh001x@hotmail.com>
This commit is contained in:
-LAN-
2025-09-18 12:49:10 +08:00
committed by GitHub
parent 7dadb33003
commit 85cda47c70
1772 changed files with 102407 additions and 31710 deletions

View File

@@ -0,0 +1,61 @@
import React from 'react'
import Divider from '@/app/components/base/divider'
import Button from '@/app/components/base/button'
import cn from '@/utils/classnames'
import { RiBookOpenLine, RiEqualizer2Line } from '@remixicon/react'
type HeaderProps = {
isInPipeline?: boolean
onClickConfiguration?: () => void
title: string
buttonText?: string
docTitle: string
docLink: string
}
const Header = ({
isInPipeline = false,
onClickConfiguration,
title,
buttonText,
docTitle,
docLink,
}: HeaderProps) => {
return (
<div className='flex items-center gap-x-2'>
<div className='flex shrink-0 grow items-center gap-x-1'>
<div className={cn(
'text-text-secondary',
isInPipeline ? 'system-sm-semibold' : 'system-md-semibold',
)}>
{title}
</div>
<Divider type='vertical' className='mx-1 h-3.5' />
<Button
variant='secondary'
size='small'
className={cn(isInPipeline ? 'size-6 px-1' : 'gap-x-0.5 px-1.5')}
onClick={onClickConfiguration}
>
<RiEqualizer2Line className='size-4' />
{!isInPipeline && (
<span className='system-xs-medium'>
{buttonText}
</span>
)}
</Button>
</div>
<a
className='system-xs-medium flex items-center gap-x-1 overflow-hidden text-text-accent'
href={docLink}
target='_blank'
rel='noopener noreferrer'
>
<RiBookOpenLine className='size-3.5 shrink-0' />
<span className='grow truncate' title={docTitle}>{docTitle}</span>
</a>
</div>
)
}
export default React.memo(Header)

View File

@@ -21,7 +21,7 @@ const Input: FC<Props> = ({
const value = e.target.value
if (isNumber) {
let numberValue = Number.parseInt(value, 10) // integer only
if (isNaN(numberValue)) {
if (Number.isNaN(numberValue)) {
onChange('')
return
}

View File

@@ -3,19 +3,19 @@ import type { CrawlResultItem } from '@/models/datasets'
const result: CrawlResultItem[] = [
{
title: 'Start the frontend Docker container separately',
markdown: 'Markdown 1',
content: 'Markdown 1',
description: 'Description 1',
source_url: 'https://example.com/1',
},
{
title: 'Advanced Tool Integration',
markdown: 'Markdown 2',
content: 'Markdown 2',
description: 'Description 2',
source_url: 'https://example.com/2',
},
{
title: 'Local Source Code Start | English | Dify',
markdown: 'Markdown 3',
content: 'Markdown 3',
description: 'Description 3',
source_url: 'https://example.com/3',
},

View File

@@ -1,43 +0,0 @@
'use client'
import type { FC } from 'react'
import React from 'react'
import { useTranslation } from 'react-i18next'
import { RiBookOpenLine, RiEqualizer2Line } from '@remixicon/react'
import Button from '@/app/components/base/button'
const I18N_PREFIX = 'datasetCreation.stepOne.website'
type Props = {
onSetting: () => void
}
const Header: FC<Props> = ({
onSetting,
}) => {
const { t } = useTranslation()
return (
<div className='flex h-6 items-center justify-between'>
<div className='flex items-center'>
<div className='text-base font-medium text-text-secondary'>{t(`${I18N_PREFIX}.firecrawlTitle`)}</div>
<div className='ml-2 mr-2 h-3.5 w-px bg-divider-regular' />
<Button className='flex h-6 items-center gap-x-[1px] px-1.5' onClick={onSetting}>
<RiEqualizer2Line className='h-3.5 w-3.5 text-components-button-secondary-text' />
<span className='px-[3px] text-xs font-medium text-components-button-secondary-text'>
{t(`${I18N_PREFIX}.configureFirecrawl`)}
</span>
</Button>
</div>
<a
href='https://docs.firecrawl.dev/introduction'
target='_blank'
rel='noopener noreferrer'
className='inline-flex items-center gap-x-1 text-xs font-medium text-text-accent'
>
<RiBookOpenLine className='h-3.5 w-3.5 text-text-accent' />
<span>{t(`${I18N_PREFIX}.firecrawlDoc`)}</span>
</a>
</div>
)
}
export default React.memo(Header)

View File

@@ -7,13 +7,13 @@ import OptionsWrap from '../base/options-wrap'
import CrawledResult from '../base/crawled-result'
import Crawling from '../base/crawling'
import ErrorMessage from '../base/error-message'
import Header from './header'
import Options from './options'
import { useModalContext } from '@/context/modal-context'
import { useModalContextSelector } from '@/context/modal-context'
import type { CrawlOptions, CrawlResultItem } from '@/models/datasets'
import Toast from '@/app/components/base/toast'
import { checkFirecrawlTaskStatus, createFirecrawlTask } from '@/service/datasets'
import { sleep } from '@/utils'
import Header from '../base/header'
const ERROR_I18N_PREFIX = 'common.errorMsg'
const I18N_PREFIX = 'datasetCreation.stepOne.website'
@@ -48,7 +48,7 @@ const FireCrawl: FC<Props> = ({
if (step !== Step.init)
setControlFoldOptions(Date.now())
}, [step])
const { setShowAccountSettingModal } = useModalContext()
const setShowAccountSettingModal = useModalContextSelector(s => s.setShowAccountSettingModal)
const handleSetting = useCallback(() => {
setShowAccountSettingModal({
payload: 'data-source',
@@ -112,6 +112,10 @@ const FireCrawl: FC<Props> = ({
},
}
}
res.data = res.data.map((item: any) => ({
...item,
content: item.markdown,
}))
// update the progress
setCrawlResult({
...res,
@@ -131,7 +135,7 @@ const FireCrawl: FC<Props> = ({
},
}
}
}, [crawlOptions.limit])
}, [crawlOptions.limit, onCheckedCrawlResultChange])
const handleRun = useCallback(async (url: string) => {
const { isValid, errorMsg } = checkValid(url)
@@ -161,6 +165,10 @@ const FireCrawl: FC<Props> = ({
setCrawlErrorMessage(errorMessage || t(`${I18N_PREFIX}.unknownError`))
}
else {
data.data = data.data.map((item: any) => ({
...item,
content: item.markdown,
}))
setCrawlResult(data)
onCheckedCrawlResultChange(data.data || []) // default select the crawl result
setCrawlErrorMessage('')
@@ -173,11 +181,17 @@ const FireCrawl: FC<Props> = ({
finally {
setStep(Step.finished)
}
}, [checkValid, crawlOptions, onJobIdChange, t, waitForCrawlFinished])
}, [checkValid, crawlOptions, onJobIdChange, t, waitForCrawlFinished, onCheckedCrawlResultChange])
return (
<div>
<Header onSetting={handleSetting} />
<Header
onClickConfiguration={handleSetting}
title={t(`${I18N_PREFIX}.firecrawlTitle`)}
buttonText={t(`${I18N_PREFIX}.configureFirecrawl`)}
docTitle={t(`${I18N_PREFIX}.firecrawlDoc`)}
docLink={'https://docs.firecrawl.dev/introduction'}
/>
<div className='mt-2 rounded-xl border border-components-panel-border bg-background-default-subtle p-4 pb-0'>
<UrlInput onRun={handleRun} isRunning={isRunning} />
<OptionsWrap

View File

@@ -6,7 +6,7 @@
}
.watercrawlLogo {
@apply w-5 h-5 bg-center bg-no-repeat inline-block;
@apply w-4 h-4 bg-center bg-no-repeat inline-block;
/*background-color: #F5FAFF;*/
background-image: url(../assets/watercrawl.svg);
background-size: 16px;

View File

@@ -1,6 +1,6 @@
'use client'
import type { FC } from 'react'
import React, { useCallback, useEffect, useState } from 'react'
import React, { useCallback, useMemo, useState } from 'react'
import { useTranslation } from 'react-i18next'
import s from './index.module.css'
import NoData from './no-data'
@@ -10,9 +10,9 @@ import JinaReader from './jina-reader'
import cn from '@/utils/classnames'
import { useModalContext } from '@/context/modal-context'
import type { CrawlOptions, CrawlResultItem } from '@/models/datasets'
import { fetchDataSources } from '@/service/datasets'
import { type DataSourceItem, DataSourceProvider } from '@/models/common'
import { DataSourceProvider } from '@/models/common'
import { ENABLE_WEBSITE_FIRECRAWL, ENABLE_WEBSITE_JINAREADER, ENABLE_WEBSITE_WATERCRAWL } from '@/config'
import type { DataSourceAuth } from '@/app/components/header/account-setting/data-source-page-new/types'
type Props = {
onPreview: (payload: CrawlResultItem) => void
@@ -22,6 +22,7 @@ type Props = {
onJobIdChange: (jobId: string) => void
crawlOptions: CrawlOptions
onCrawlOptionsChange: (payload: CrawlOptions) => void
authedDataSourceList: DataSourceAuth[]
}
const Website: FC<Props> = ({
@@ -32,58 +33,34 @@ const Website: FC<Props> = ({
onJobIdChange,
crawlOptions,
onCrawlOptionsChange,
authedDataSourceList,
}) => {
const { t } = useTranslation()
const { setShowAccountSettingModal } = useModalContext()
const [isLoaded, setIsLoaded] = useState(false)
const [selectedProvider, setSelectedProvider] = useState<DataSourceProvider>(DataSourceProvider.jinaReader)
const [sources, setSources] = useState<DataSourceItem[]>([])
useEffect(() => {
onCrawlProviderChange(selectedProvider)
}, [selectedProvider, onCrawlProviderChange])
const availableProviders = useMemo(() => authedDataSourceList.filter((item) => {
return [
DataSourceProvider.jinaReader,
DataSourceProvider.fireCrawl,
DataSourceProvider.waterCrawl].includes(item.provider as DataSourceProvider) && item.credentials_list.length > 0
}), [authedDataSourceList])
const checkSetApiKey = useCallback(async () => {
const res = await fetchDataSources() as any
setSources(res.sources)
// If users have configured one of the providers, select it.
const availableProviders = res.sources.filter((item: DataSourceItem) =>
[
DataSourceProvider.jinaReader,
DataSourceProvider.fireCrawl,
DataSourceProvider.waterCrawl,
].includes(item.provider),
)
if (availableProviders.length > 0)
setSelectedProvider(availableProviders[0].provider)
}, [])
useEffect(() => {
checkSetApiKey().then(() => {
setIsLoaded(true)
})
}, [])
const handleOnConfig = useCallback(() => {
setShowAccountSettingModal({
payload: 'data-source',
onCancelCallback: checkSetApiKey,
})
}, [checkSetApiKey, setShowAccountSettingModal])
}, [setShowAccountSettingModal])
if (!isLoaded)
return null
const source = sources.find(source => source.provider === selectedProvider)
const source = availableProviders.find(source => source.provider === selectedProvider)
return (
<div>
<div className="mb-4">
<div className="system-md-medium mb-2 text-text-secondary">
<div className='mb-4'>
<div className='system-md-medium mb-2 text-text-secondary'>
{t('datasetCreation.stepOne.website.chooseProvider')}
</div>
<div className="flex space-x-2">
<div className='flex space-x-2'>
{ENABLE_WEBSITE_JINAREADER && <button
className={cn('flex items-center justify-center rounded-lg px-4 py-2',
selectedProvider === DataSourceProvider.jinaReader
@@ -91,9 +68,12 @@ const Website: FC<Props> = ({
: `system-sm-regular border border-components-option-card-option-border bg-components-option-card-option-bg text-text-secondary
hover:border-components-option-card-option-border-hover hover:bg-components-option-card-option-bg-hover hover:shadow-xs hover:shadow-shadow-shadow-3`,
)}
onClick={() => setSelectedProvider(DataSourceProvider.jinaReader)}
onClick={() => {
setSelectedProvider(DataSourceProvider.jinaReader)
onCrawlProviderChange(DataSourceProvider.jinaReader)
}}
>
<span className={cn(s.jinaLogo, 'mr-2')}/>
<span className={cn(s.jinaLogo, 'mr-2')} />
<span>Jina Reader</span>
</button>}
{ENABLE_WEBSITE_FIRECRAWL && <button
@@ -103,7 +83,10 @@ const Website: FC<Props> = ({
: `system-sm-regular border border-components-option-card-option-border bg-components-option-card-option-bg text-text-secondary
hover:border-components-option-card-option-border-hover hover:bg-components-option-card-option-bg-hover hover:shadow-xs hover:shadow-shadow-shadow-3`,
)}
onClick={() => setSelectedProvider(DataSourceProvider.fireCrawl)}
onClick={() => {
setSelectedProvider(DataSourceProvider.fireCrawl)
onCrawlProviderChange(DataSourceProvider.fireCrawl)
}}
>
🔥 Firecrawl
</button>}
@@ -114,9 +97,12 @@ const Website: FC<Props> = ({
: `system-sm-regular border border-components-option-card-option-border bg-components-option-card-option-bg text-text-secondary
hover:border-components-option-card-option-border-hover hover:bg-components-option-card-option-bg-hover hover:shadow-xs hover:shadow-shadow-shadow-3`,
)}
onClick={() => setSelectedProvider(DataSourceProvider.waterCrawl)}
onClick={() => {
setSelectedProvider(DataSourceProvider.waterCrawl)
onCrawlProviderChange(DataSourceProvider.waterCrawl)
}}
>
<span className={cn(s.watercrawlLogo, 'mr-2')}/>
<span className={cn(s.watercrawlLogo, 'mr-2')} />
<span>WaterCrawl</span>
</button>}
</div>
@@ -152,7 +138,7 @@ const Website: FC<Props> = ({
/>
)}
{!source && (
<NoData onConfig={handleOnConfig} provider={selectedProvider}/>
<NoData onConfig={handleOnConfig} provider={selectedProvider} />
)}
</div>
)

View File

@@ -1,40 +0,0 @@
'use client'
import type { FC } from 'react'
import React from 'react'
import cn from '@/utils/classnames'
import Checkbox from '@/app/components/base/checkbox'
import Tooltip from '@/app/components/base/tooltip'
type Props = {
className?: string
isChecked: boolean
onChange: (isChecked: boolean) => void
label: string
labelClassName?: string
tooltip?: string
}
const CheckboxWithLabel: FC<Props> = ({
className = '',
isChecked,
onChange,
label,
labelClassName,
tooltip,
}) => {
return (
<label className={cn(className, 'flex h-7 items-center space-x-2')}>
<Checkbox checked={isChecked} onCheck={() => onChange(!isChecked)} />
<div className={cn('text-sm font-normal text-gray-800', labelClassName)}>{label}</div>
{tooltip && (
<Tooltip
popupContent={
<div className='w-[200px]'>{tooltip}</div>
}
triggerClassName='ml-0.5 w-4 h-4'
/>
)}
</label>
)
}
export default React.memo(CheckboxWithLabel)

View File

@@ -1,30 +0,0 @@
'use client'
import type { FC } from 'react'
import React from 'react'
import cn from '@/utils/classnames'
import { AlertTriangle } from '@/app/components/base/icons/src/vender/solid/alertsAndFeedback'
type Props = {
className?: string
title: string
errorMsg?: string
}
const ErrorMessage: FC<Props> = ({
className,
title,
errorMsg,
}) => {
return (
<div className={cn(className, 'border-t border-gray-200 bg-[#FFFAEB] px-4 py-2')}>
<div className='flex h-5 items-center'>
<AlertTriangle className='mr-2 h-4 w-4 text-[#F79009]' />
<div className='text-sm font-medium text-[#DC6803]'>{title}</div>
</div>
{errorMsg && (
<div className='mt-1 pl-6 text-xs font-normal leading-[18px] text-gray-700'>{errorMsg}</div>
)}
</div>
)
}
export default React.memo(ErrorMessage)

View File

@@ -1,54 +0,0 @@
'use client'
import type { FC } from 'react'
import React from 'react'
import Input from './input'
import cn from '@/utils/classnames'
import Tooltip from '@/app/components/base/tooltip'
type Props = {
className?: string
label: string
labelClassName?: string
value: string | number
onChange: (value: string | number) => void
isRequired?: boolean
placeholder?: string
isNumber?: boolean
tooltip?: string
}
const Field: FC<Props> = ({
className,
label,
labelClassName,
value,
onChange,
isRequired = false,
placeholder = '',
isNumber = false,
tooltip,
}) => {
return (
<div className={cn(className)}>
<div className='flex py-[7px]'>
<div className={cn(labelClassName, 'flex h-[18px] items-center text-[13px] font-medium text-gray-900')}>{label} </div>
{isRequired && <span className='ml-0.5 text-xs font-semibold text-[#D92D20]'>*</span>}
{tooltip && (
<Tooltip
popupContent={
<div className='w-[200px]'>{tooltip}</div>
}
triggerClassName='ml-0.5 w-4 h-4'
/>
)}
</div>
<Input
value={value}
onChange={onChange}
placeholder={placeholder}
isNumber={isNumber}
/>
</div>
)
}
export default React.memo(Field)

View File

@@ -1,58 +0,0 @@
'use client'
import type { FC } from 'react'
import React, { useCallback } from 'react'
type Props = {
value: string | number
onChange: (value: string | number) => void
placeholder?: string
isNumber?: boolean
}
const MIN_VALUE = 0
const Input: FC<Props> = ({
value,
onChange,
placeholder = '',
isNumber = false,
}) => {
const handleChange = useCallback((e: React.ChangeEvent<HTMLInputElement>) => {
const value = e.target.value
if (isNumber) {
let numberValue = Number.parseInt(value, 10) // integer only
if (isNaN(numberValue)) {
onChange('')
return
}
if (numberValue < MIN_VALUE)
numberValue = MIN_VALUE
onChange(numberValue)
return
}
onChange(value)
}, [isNumber, onChange])
const otherOption = (() => {
if (isNumber) {
return {
min: MIN_VALUE,
}
}
return {
}
})()
return (
<input
type={isNumber ? 'number' : 'text'}
{...otherOption}
value={value}
onChange={handleChange}
className='flex h-9 w-full rounded-lg bg-gray-100 px-2 py-1 text-xs leading-normal caret-primary-600 placeholder:text-gray-400 hover:bg-gray-100 focus:bg-gray-50 focus:ring-1 focus:ring-inset focus:ring-gray-200 focus-visible:outline-none'
placeholder={placeholder}
/>
)
}
export default React.memo(Input)

View File

@@ -1,54 +0,0 @@
'use client'
import { useBoolean } from 'ahooks'
import type { FC } from 'react'
import React, { useEffect } from 'react'
import { useTranslation } from 'react-i18next'
import cn from '@/utils/classnames'
import { Settings04 } from '@/app/components/base/icons/src/vender/line/general'
import { ChevronRight } from '@/app/components/base/icons/src/vender/line/arrows'
const I18N_PREFIX = 'datasetCreation.stepOne.website'
type Props = {
className?: string
children: React.ReactNode
controlFoldOptions?: number
}
const OptionsWrap: FC<Props> = ({
className = '',
children,
controlFoldOptions,
}) => {
const { t } = useTranslation()
const [fold, {
toggle: foldToggle,
setTrue: foldHide,
}] = useBoolean(false)
useEffect(() => {
if (controlFoldOptions)
foldHide()
}, [controlFoldOptions])
return (
<div className={cn(className, !fold ? 'mb-0' : 'mb-3')}>
<div
className='flex h-[26px] cursor-pointer select-none items-center justify-between py-1'
onClick={foldToggle}
>
<div className='flex items-center text-gray-700'>
<Settings04 className='mr-1 h-4 w-4' />
<div className='text-[13px] font-semibold uppercase text-gray-800'>{t(`${I18N_PREFIX}.options`)}</div>
</div>
<ChevronRight className={cn(!fold && 'rotate-90', 'h-4 w-4 text-gray-500')} />
</div>
{!fold && (
<div className='mb-4'>
{children}
</div>
)}
</div>
)
}
export default React.memo(OptionsWrap)

View File

@@ -2,7 +2,7 @@
import type { FC } from 'react'
import React, { useCallback, useState } from 'react'
import { useTranslation } from 'react-i18next'
import Input from './input'
import Input from '../../base/input'
import Button from '@/app/components/base/button'
import { useDocLink } from '@/context/i18n'

View File

@@ -1,43 +0,0 @@
'use client'
import type { FC } from 'react'
import React from 'react'
import { useTranslation } from 'react-i18next'
import { RiBookOpenLine, RiEqualizer2Line } from '@remixicon/react'
import Button from '@/app/components/base/button'
const I18N_PREFIX = 'datasetCreation.stepOne.website'
type Props = {
onSetting: () => void
}
const Header: FC<Props> = ({
onSetting,
}) => {
const { t } = useTranslation()
return (
<div className='flex h-6 items-center justify-between'>
<div className='flex items-center'>
<div className='text-sm font-semibold text-text-secondary'>{t(`${I18N_PREFIX}.jinaReaderTitle`)}</div>
<div className='ml-2 mr-2 h-3.5 w-px bg-divider-regular' />
<Button className='flex h-6 items-center gap-x-[1px] px-1.5' onClick={onSetting}>
<RiEqualizer2Line className='h-3.5 w-3.5 text-components-button-secondary-text' />
<span className='px-[3px] text-xs font-medium text-components-button-secondary-text'>
{t(`${I18N_PREFIX}.configureJinaReader`)}
</span>
</Button>
</div>
<a
href='https://jina.ai/reader'
target='_blank'
rel='noopener noreferrer'
className='inline-flex items-center gap-x-1 text-xs font-medium text-text-accent'
>
<RiBookOpenLine className='h-3.5 w-3.5 text-text-accent' />
<span>{t(`${I18N_PREFIX}.jinaReaderDoc`)}</span>
</a>
</div>
)
}
export default React.memo(Header)

View File

@@ -7,13 +7,13 @@ import OptionsWrap from '../base/options-wrap'
import CrawledResult from '../base/crawled-result'
import Crawling from '../base/crawling'
import ErrorMessage from '../base/error-message'
import Header from './header'
import Options from './options'
import { useModalContext } from '@/context/modal-context'
import Toast from '@/app/components/base/toast'
import { checkJinaReaderTaskStatus, createJinaReaderTask } from '@/service/datasets'
import { sleep } from '@/utils'
import type { CrawlOptions, CrawlResultItem } from '@/models/datasets'
import Header from '../base/header'
const ERROR_I18N_PREFIX = 'common.errorMsg'
const I18N_PREFIX = 'datasetCreation.stepOne.website'
@@ -150,14 +150,15 @@ const JinaReader: FC<Props> = ({
}) as any
if (res.data) {
const { title, content, description, url } = res.data
const data = {
current: 1,
total: 1,
data: [{
title: res.data.title,
markdown: res.data.content,
description: res.data.description,
source_url: res.data.url,
title,
content,
description,
source_url: url,
}],
time_consuming: (Date.now() - startTime) / 1000,
}
@@ -190,7 +191,13 @@ const JinaReader: FC<Props> = ({
return (
<div>
<Header onSetting={handleSetting} />
<Header
onClickConfiguration={handleSetting}
title={t(`${I18N_PREFIX}.jinaReaderTitle`)}
buttonText={t(`${I18N_PREFIX}.configureJinaReader`)}
docTitle={t(`${I18N_PREFIX}.jinaReaderDoc`)}
docLink={'https://jina.ai/reader'}
/>
<div className='mt-2 rounded-xl border border-components-panel-border bg-background-default-subtle p-4 pb-0'>
<UrlInput onRun={handleRun} isRunning={isRunning} />
<OptionsWrap

View File

@@ -5,8 +5,8 @@ import { useTranslation } from 'react-i18next'
import s from './index.module.css'
import { Icon3Dots } from '@/app/components/base/icons/src/vender/line/others'
import Button from '@/app/components/base/button'
import { DataSourceProvider } from '@/models/common'
import { ENABLE_WEBSITE_FIRECRAWL, ENABLE_WEBSITE_JINAREADER, ENABLE_WEBSITE_WATERCRAWL } from '@/config'
import { DataSourceProvider } from '@/models/common'
const I18N_PREFIX = 'datasetCreation.stepOne.website'
@@ -21,7 +21,11 @@ const NoData: FC<Props> = ({
}) => {
const { t } = useTranslation()
const providerConfig = {
const providerConfig: Record<DataSourceProvider, {
emoji: React.ReactNode
title: string
description: string
} | null> = {
[DataSourceProvider.jinaReader]: ENABLE_WEBSITE_JINAREADER ? {
emoji: <span className={s.jinaLogo} />,
title: t(`${I18N_PREFIX}.jinaReaderNotConfigured`),
@@ -39,7 +43,7 @@ const NoData: FC<Props> = ({
} : null,
}
const currentProvider = providerConfig[provider] || providerConfig[DataSourceProvider.jinaReader]
const currentProvider = providerConfig[provider] || providerConfig.jinareader
if (!currentProvider) return null

View File

@@ -32,7 +32,7 @@ const WebsitePreview = ({
<div className='system-xs-medium truncate text-text-tertiary' title={payload.source_url}>{payload.source_url}</div>
</div>
<div className={cn(s.previewContent, 'body-md-regular')}>
<div className={cn(s.fileContent)}>{payload.markdown}</div>
<div className={cn(s.fileContent)}>{payload.content}</div>
</div>
</div>
)

View File

@@ -1,43 +0,0 @@
'use client'
import type { FC } from 'react'
import React from 'react'
import { useTranslation } from 'react-i18next'
import { RiBookOpenLine, RiEqualizer2Line } from '@remixicon/react'
import Button from '@/app/components/base/button'
const I18N_PREFIX = 'datasetCreation.stepOne.website'
type Props = {
onSetting: () => void
}
const Header: FC<Props> = ({
onSetting,
}) => {
const { t } = useTranslation()
return (
<div className='flex h-6 items-center justify-between'>
<div className='flex items-center'>
<div className='text-base font-medium text-text-secondary'>{t(`${I18N_PREFIX}.watercrawlTitle`)}</div>
<div className='ml-2 mr-2 h-3.5 w-px bg-divider-regular' />
<Button className='flex h-6 items-center gap-x-[1px] px-1.5' onClick={onSetting}>
<RiEqualizer2Line className='h-3.5 w-3.5 text-components-button-secondary-text' />
<span className='px-[3px] text-xs font-medium text-components-button-secondary-text'>
{t(`${I18N_PREFIX}.configureWatercrawl`)}
</span>
</Button>
</div>
<a
href='https://docs.watercrawl.dev/'
target='_blank'
rel='noopener noreferrer'
className='inline-flex items-center gap-x-1 text-xs font-medium text-text-accent'
>
<RiBookOpenLine className='h-3.5 w-3.5 text-text-accent' />
<span>{t(`${I18N_PREFIX}.watercrawlDoc`)}</span>
</a>
</div>
)
}
export default React.memo(Header)

View File

@@ -7,13 +7,13 @@ import OptionsWrap from '../base/options-wrap'
import CrawledResult from '../base/crawled-result'
import Crawling from '../base/crawling'
import ErrorMessage from '../base/error-message'
import Header from './header'
import Options from './options'
import { useModalContext } from '@/context/modal-context'
import type { CrawlOptions, CrawlResultItem } from '@/models/datasets'
import Toast from '@/app/components/base/toast'
import { checkWatercrawlTaskStatus, createWatercrawlTask } from '@/service/datasets'
import { sleep } from '@/utils'
import Header from '../base/header'
const ERROR_I18N_PREFIX = 'common.errorMsg'
const I18N_PREFIX = 'datasetCreation.stepOne.website'
@@ -177,7 +177,13 @@ const WaterCrawl: FC<Props> = ({
return (
<div>
<Header onSetting={handleSetting} />
<Header
onClickConfiguration={handleSetting}
title={t(`${I18N_PREFIX}.watercrawlTitle`)}
buttonText={t(`${I18N_PREFIX}.configureWatercrawl`)}
docTitle={t(`${I18N_PREFIX}.watercrawlDoc`)}
docLink={'https://docs.watercrawl.dev/'}
/>
<div className='mt-2 rounded-xl border border-components-panel-border bg-background-default-subtle p-4 pb-0'>
<UrlInput onRun={handleRun} isRunning={isRunning} />
<OptionsWrap