feat: Integrate WaterCrawl.dev as a new knowledge base provider (#16396)

Co-authored-by: crazywoola <427733928@qq.com>
This commit is contained in:
Amir Mohsen Asaran
2025-04-07 06:43:23 +02:00
committed by GitHub
parent 0afad94378
commit f54905e685
24 changed files with 1102 additions and 55 deletions

View File

@@ -0,0 +1,161 @@
'use client'
import type { FC } from 'react'
import React, { useCallback, useState } from 'react'
import { useTranslation } from 'react-i18next'
import {
PortalToFollowElem,
PortalToFollowElemContent,
} from '@/app/components/base/portal-to-follow-elem'
import { Lock01 } from '@/app/components/base/icons/src/vender/solid/security'
import Button from '@/app/components/base/button'
import type { WatercrawlConfig } from '@/models/common'
import Field from '@/app/components/datasets/create/website/base/field'
import Toast from '@/app/components/base/toast'
import { createDataSourceApiKeyBinding } from '@/service/datasets'
import { LinkExternal02 } from '@/app/components/base/icons/src/vender/line/general'
type Props = {
onCancel: () => void
onSaved: () => void
}
const I18N_PREFIX = 'datasetCreation.watercrawl'
const DEFAULT_BASE_URL = 'https://app.watercrawl.dev'
const ConfigWatercrawlModal: FC<Props> = ({
onCancel,
onSaved,
}) => {
const { t } = useTranslation()
const [isSaving, setIsSaving] = useState(false)
const [config, setConfig] = useState<WatercrawlConfig>({
api_key: '',
base_url: '',
})
const handleConfigChange = useCallback((key: string) => {
return (value: string | number) => {
setConfig(prev => ({ ...prev, [key]: value as string }))
}
}, [])
const handleSave = useCallback(async () => {
if (isSaving)
return
let errorMsg = ''
if (config.base_url && !((config.base_url.startsWith('http://') || config.base_url.startsWith('https://'))))
errorMsg = t('common.errorMsg.urlError')
if (!errorMsg) {
if (!config.api_key) {
errorMsg = t('common.errorMsg.fieldRequired', {
field: 'API Key',
})
}
}
if (errorMsg) {
Toast.notify({
type: 'error',
message: errorMsg,
})
return
}
const postData = {
category: 'website',
provider: 'watercrawl',
credentials: {
auth_type: 'x-api-key',
config: {
api_key: config.api_key,
base_url: config.base_url || DEFAULT_BASE_URL,
},
},
}
try {
setIsSaving(true)
await createDataSourceApiKeyBinding(postData)
Toast.notify({
type: 'success',
message: t('common.api.success'),
})
}
finally {
setIsSaving(false)
}
onSaved()
}, [config.api_key, config.base_url, onSaved, t, isSaving])
return (
<PortalToFollowElem open>
<PortalToFollowElemContent className='w-full h-full z-[60]'>
<div className='fixed inset-0 flex items-center justify-center bg-background-overlay'>
<div className='mx-2 w-[640px] max-h-[calc(100vh-120px)] bg-components-panel-bg shadow-xl rounded-2xl overflow-y-auto'>
<div className='px-8 pt-8'>
<div className='flex justify-between items-center mb-4'>
<div className='system-xl-semibold text-text-primary'>{t(`${I18N_PREFIX}.configWatercrawl`)}</div>
</div>
<div className='space-y-4'>
<Field
label='API Key'
labelClassName='!text-sm'
isRequired
value={config.api_key}
onChange={handleConfigChange('api_key')}
placeholder={t(`${I18N_PREFIX}.apiKeyPlaceholder`)!}
/>
<Field
label='Base URL'
labelClassName='!text-sm'
value={config.base_url}
onChange={handleConfigChange('base_url')}
placeholder={DEFAULT_BASE_URL}
/>
</div>
<div className='my-8 flex justify-between items-center h-8'>
<a className='flex items-center space-x-1 leading-[18px] text-xs font-normal text-text-accent' target='_blank' href='https://app.watercrawl.dev/'>
<span>{t(`${I18N_PREFIX}.getApiKeyLinkText`)}</span>
<LinkExternal02 className='w-3 h-3' />
</a>
<div className='flex'>
<Button
size='large'
className='mr-2'
onClick={onCancel}
>
{t('common.operation.cancel')}
</Button>
<Button
variant='primary'
size='large'
onClick={handleSave}
loading={isSaving}
>
{t('common.operation.save')}
</Button>
</div>
</div>
</div>
<div className='border-t-[0.5px] border-t-divider-regular'>
<div className='flex justify-center items-center py-3 bg-background-section-burn text-xs text-text-tertiary'>
<Lock01 className='mr-1 w-3 h-3 text-text-tertiary' />
{t('common.modelProvider.encrypted.front')}
<a
className='text-text-accent mx-1'
target='_blank' rel='noopener noreferrer'
href='https://pycryptodome.readthedocs.io/en/latest/src/cipher/oaep.html'
>
PKCS1_OAEP
</a>
{t('common.modelProvider.encrypted.back')}
</div>
</div>
</div>
</div>
</PortalToFollowElemContent>
</PortalToFollowElem>
)
}
export default React.memo(ConfigWatercrawlModal)

View File

@@ -5,19 +5,15 @@ import { useTranslation } from 'react-i18next'
import Panel from '../panel'
import { DataSourceType } from '../panel/types'
import ConfigFirecrawlModal from './config-firecrawl-modal'
import ConfigWatercrawlModal from './config-watercrawl-modal'
import ConfigJinaReaderModal from './config-jina-reader-modal'
import cn from '@/utils/classnames'
import s from '@/app/components/datasets/create/website/index.module.css'
import { fetchDataSources, removeDataSourceApiKeyBinding } from '@/service/datasets'
import type {
DataSourceItem,
} from '@/models/common'
import type { DataSourceItem } from '@/models/common'
import { DataSourceProvider } from '@/models/common'
import { useAppContext } from '@/context/app-context'
import {
DataSourceProvider,
} from '@/models/common'
import Toast from '@/app/components/base/toast'
type Props = {
@@ -58,6 +54,16 @@ const DataSourceWebsite: FC<Props> = ({ provider }) => {
return source?.id
}
const getProviderName = (provider: DataSourceProvider): string => {
if (provider === DataSourceProvider.fireCrawl)
return 'Firecrawl'
if (provider === DataSourceProvider.waterCrawl)
return 'WaterCrawl'
return 'Jina Reader'
}
const handleRemove = useCallback((provider: DataSourceProvider) => {
return async () => {
const dataSourceId = getIdByProvider(provider)
@@ -82,27 +88,42 @@ const DataSourceWebsite: FC<Props> = ({ provider }) => {
readOnly={!isCurrentWorkspaceManager}
configuredList={sources.filter(item => item.provider === provider).map(item => ({
id: item.id,
logo: ({ className }: { className: string }) => (
item.provider === DataSourceProvider.fireCrawl
? (
<div className={cn(className, 'ml-3 flex h-5 w-5 items-center justify-center rounded border border-divider-subtle !bg-background-default text-xs font-medium text-text-tertiary')}>🔥</div>
logo: ({ className }: { className: string }) => {
if (item.provider === DataSourceProvider.fireCrawl) {
return (
<div
className={cn(className, 'ml-3 flex h-5 w-5 items-center justify-center rounded border border-divider-subtle !bg-background-default text-xs font-medium text-text-tertiary')}>🔥</div>
)
: (
<div className={cn(className, 'ml-3 flex h-5 w-5 items-center justify-center rounded border border-divider-subtle !bg-background-default text-xs font-medium text-text-tertiary')}>
<span className={s.jinaLogo} />
}
if (item.provider === DataSourceProvider.waterCrawl) {
return (
<div
className={cn(className, 'ml-3 flex h-5 w-5 items-center justify-center rounded border border-divider-subtle !bg-background-default text-xs font-medium text-text-tertiary')}>
<span className={s.watercrawlLogo}/>
</div>
)
),
name: item.provider === DataSourceProvider.fireCrawl ? 'Firecrawl' : 'Jina Reader',
}
return (
<div
className={cn(className, 'ml-3 flex h-5 w-5 items-center justify-center rounded border border-divider-subtle !bg-background-default text-xs font-medium text-text-tertiary')}>
<span className={s.jinaLogo}/>
</div>
)
},
name: getProviderName(item.provider),
isActive: true,
}))}
onRemove={handleRemove(provider)}
/>
{configTarget === DataSourceProvider.fireCrawl && (
<ConfigFirecrawlModal onSaved={handleAdded} onCancel={hideConfig} />
<ConfigFirecrawlModal onSaved={handleAdded} onCancel={hideConfig}/>
)}
{configTarget === DataSourceProvider.waterCrawl && (
<ConfigWatercrawlModal onSaved={handleAdded} onCancel={hideConfig}/>
)}
{configTarget === DataSourceProvider.jinaReader && (
<ConfigJinaReaderModal onSaved={handleAdded} onCancel={hideConfig} />
<ConfigJinaReaderModal onSaved={handleAdded} onCancel={hideConfig}/>
)}
</>