hz/ebiz-dify-ai
1
0
Fork 0
mirror of http://112.124.100.131/huang.ze/ebiz-dify-ai.git synced 2025-12-09 19:06:51 +08:00
Code Issues Packages Projects Releases Wiki Activity

feat: add switches for jina firecrawl watercrawl (#18153)

Browse Source
This commit is contained in:
crazywoola
2025-04-16 15:50:15 +08:00
committed by GitHub
parent b247ef85bf
commit e1455cecd8
10 changed files with 64 additions and 28 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
web/app/components/datasets/create/step-one/index.tsx
View File

@@ -20,7 +20,7 @@ import { useProviderContext } from '@/context/provider-context'
import VectorSpaceFull from '@/app/components/billing/vector-space-full'
import classNames from '@/utils/classnames'
import { Icon3Dots } from '@/app/components/base/icons/src/vender/line/others'
import { ENABLE_WEBSITE_FIRECRAWL, ENABLE_WEBSITE_JINAREADER, ENABLE_WEBSITE_WATERCRAWL } from '@/config'
type IStepOneProps = {
datasetId?: string
dataSourceType?: DataSourceType
@@ -126,9 +126,7 @@ const StepOne = ({
return true
if (files.some(file => !file.file.id))
return true
if (isShowVectorSpaceFull)
return true
return false
return isShowVectorSpaceFull
}, [files, isShowVectorSpaceFull])
return (
@@ -193,7 +191,8 @@ const StepOne = ({
{t('datasetCreation.stepOne.dataSourceType.notion')}
</span>
</div>
<div
{(ENABLE_WEBSITE_FIRECRAWL || ENABLE_WEBSITE_JINAREADER || ENABLE_WEBSITE_WATERCRAWL) && (
<div
className={cn(
s.dataSourceItem,
'system-sm-medium',
@@ -201,7 +200,7 @@ const StepOne = ({
dataSourceTypeDisable && dataSourceType !== DataSourceType.WEB && s.disabled,
)}
onClick={() => changeType(DataSourceType.WEB)}
>
>
<span className={cn(s.datasetIcon, s.web)} />
<span
title={t('datasetCreation.stepOne.dataSourceType.web')}
@@ -209,7 +208,8 @@ const StepOne = ({
>
{t('datasetCreation.stepOne.dataSourceType.web')}
</span>
</div>
</div>
)}
</div>
)
}

13
web/app/components/datasets/create/website/index.tsx
View File

@@ -12,6 +12,7 @@ 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 { ENABLE_WEBSITE_FIRECRAWL, ENABLE_WEBSITE_JINAREADER, ENABLE_WEBSITE_WATERCRAWL } from '@/config'
type Props = {
onPreview: (payload: CrawlResultItem) => void
@@ -84,7 +85,7 @@ const Website: FC<Props> = ({
{t('datasetCreation.stepOne.website.chooseProvider')}
</div>
<div className="flex space-x-2">
<button
{ENABLE_WEBSITE_JINAREADER && <button
className={cn('flex items-center justify-center rounded-lg px-4 py-2',
selectedProvider === DataSourceProvider.jinaReader
? 'system-sm-medium border-[1.5px] border-components-option-card-option-selected-border bg-components-option-card-option-selected-bg text-text-primary'
@@ -95,8 +96,8 @@ const Website: FC<Props> = ({
>
<span className={cn(s.jinaLogo, 'mr-2')}/>
<span>Jina Reader</span>
</button>
<button
</button>}
{ENABLE_WEBSITE_FIRECRAWL && <button
className={cn('rounded-lg px-4 py-2',
selectedProvider === DataSourceProvider.fireCrawl
? 'system-sm-medium border-[1.5px] border-components-option-card-option-selected-border bg-components-option-card-option-selected-bg text-text-primary'
@@ -106,8 +107,8 @@ const Website: FC<Props> = ({
onClick={() => setSelectedProvider(DataSourceProvider.fireCrawl)}
>
🔥 Firecrawl
</button>
<button
</button>}
{ENABLE_WEBSITE_WATERCRAWL && <button
className={cn('flex items-center justify-center rounded-lg px-4 py-2',
selectedProvider === DataSourceProvider.waterCrawl
? 'system-sm-medium border-[1.5px] border-components-option-card-option-selected-border bg-components-option-card-option-selected-bg text-text-primary'
@@ -118,7 +119,7 @@ const Website: FC<Props> = ({
>
<span className={cn(s.watercrawlLogo, 'mr-2')}/>
<span>WaterCrawl</span>
</button>
</button>}
</div>
</div>
{source && selectedProvider === DataSourceProvider.fireCrawl && (

20
web/app/components/datasets/create/website/no-data.tsx
View File

@@ -6,6 +6,7 @@ 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'
const I18N_PREFIX = 'datasetCreation.stepOne.website'
@@ -16,29 +17,30 @@ type Props = {
const NoData: FC<Props> = ({
onConfig,
provider,
}) => {
const { t } = useTranslation()
const providerConfig = {
[DataSourceProvider.jinaReader]: {
[DataSourceProvider.jinaReader]: ENABLE_WEBSITE_JINAREADER ? {
emoji: <span className={s.jinaLogo} />,
title: t(`${I18N_PREFIX}.jinaReaderNotConfigured`),
description: t(`${I18N_PREFIX}.jinaReaderNotConfiguredDescription`),
},
[DataSourceProvider.fireCrawl]: {
} : null,
[DataSourceProvider.fireCrawl]: ENABLE_WEBSITE_FIRECRAWL ? {
emoji: '🔥',
title: t(`${I18N_PREFIX}.fireCrawlNotConfigured`),
description: t(`${I18N_PREFIX}.fireCrawlNotConfiguredDescription`),
},
[DataSourceProvider.waterCrawl]: {
emoji: <span className={s.watercrawlLogo} />,
} : null,
[DataSourceProvider.waterCrawl]: ENABLE_WEBSITE_WATERCRAWL ? {
emoji: '💧',
title: t(`${I18N_PREFIX}.waterCrawlNotConfigured`),
description: t(`${I18N_PREFIX}.waterCrawlNotConfiguredDescription`),
},
} : null,
}
const currentProvider = providerConfig[provider]
const currentProvider = Object.values(providerConfig).find(provider => provider !== null) || providerConfig[DataSourceProvider.jinaReader]
if (!currentProvider) return null
return (
<>

7
web/app/components/header/account-setting/data-source-page/index.tsx
View File

@@ -3,6 +3,7 @@ import DataSourceNotion from './data-source-notion'
import DataSourceWebsite from './data-source-website'
import { fetchDataSource } from '@/service/common'
import { DataSourceProvider } from '@/models/common'
import { ENABLE_WEBSITE_FIRECRAWL, ENABLE_WEBSITE_JINAREADER, ENABLE_WEBSITE_WATERCRAWL } from '@/config'
export default function DataSourcePage() {
const { data } = useSWR({ url: 'data-source/integrates' }, fetchDataSource)
@@ -11,9 +12,9 @@ export default function DataSourcePage() {
return (
<div className='mb-8'>
<DataSourceNotion workspaces={notionWorkspaces} />
<DataSourceWebsite provider={DataSourceProvider.jinaReader} />
<DataSourceWebsite provider={DataSourceProvider.fireCrawl} />
<DataSourceWebsite provider={DataSourceProvider.waterCrawl} />
{ENABLE_WEBSITE_JINAREADER && <DataSourceWebsite provider={DataSourceProvider.jinaReader} />}
{ENABLE_WEBSITE_FIRECRAWL && <DataSourceWebsite provider={DataSourceProvider.fireCrawl} />}
{ENABLE_WEBSITE_WATERCRAWL && <DataSourceWebsite provider={DataSourceProvider.waterCrawl} />}
</div>
)
}