feat(website-crawl): add jina reader as additional alternative for website crawling (#8761)

This commit is contained in:
Zhaofeng Miao
2024-09-30 09:57:19 +08:00
committed by GitHub
parent fb49413a41
commit 369e1e6f58
38 changed files with 927 additions and 75 deletions

View File

@@ -2,11 +2,12 @@
import type { FC } from 'react'
import React, { useCallback, useEffect, useState } from 'react'
import { useTranslation } from 'react-i18next'
import { useBoolean } from 'ahooks'
import Panel from '../panel'
import { DataSourceType } from '../panel/types'
import ConfigFirecrawlModal from './config-firecrawl-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 {
@@ -19,9 +20,11 @@ import {
} from '@/models/common'
import Toast from '@/app/components/base/toast'
type Props = {}
type Props = {
provider: DataSourceProvider
}
const DataSourceWebsite: FC<Props> = () => {
const DataSourceWebsite: FC<Props> = ({ provider }) => {
const { t } = useTranslation()
const { isCurrentWorkspaceManager } = useAppContext()
const [sources, setSources] = useState<DataSourceItem[]>([])
@@ -36,22 +39,26 @@ const DataSourceWebsite: FC<Props> = () => {
// eslint-disable-next-line react-hooks/exhaustive-deps
}, [])
const [isShowConfig, {
setTrue: showConfig,
setFalse: hideConfig,
}] = useBoolean(false)
const [configTarget, setConfigTarget] = useState<DataSourceProvider | null>(null)
const showConfig = useCallback((provider: DataSourceProvider) => {
setConfigTarget(provider)
}, [setConfigTarget])
const hideConfig = useCallback(() => {
setConfigTarget(null)
}, [setConfigTarget])
const handleAdded = useCallback(() => {
checkSetApiKey()
hideConfig()
}, [checkSetApiKey, hideConfig])
const getIdByProvider = (provider: string): string | undefined => {
const getIdByProvider = (provider: DataSourceProvider): string | undefined => {
const source = sources.find(item => item.provider === provider)
return source?.id
}
const handleRemove = useCallback((provider: string) => {
const handleRemove = useCallback((provider: DataSourceProvider) => {
return async () => {
const dataSourceId = getIdByProvider(provider)
if (dataSourceId) {
@@ -69,22 +76,34 @@ const DataSourceWebsite: FC<Props> = () => {
<>
<Panel
type={DataSourceType.website}
isConfigured={sources.length > 0}
onConfigure={showConfig}
provider={provider}
isConfigured={sources.find(item => item.provider === provider) !== undefined}
onConfigure={() => showConfig(provider)}
readOnly={!isCurrentWorkspaceManager}
configuredList={sources.map(item => ({
configuredList={sources.filter(item => item.provider === provider).map(item => ({
id: item.id,
logo: ({ className }: { className: string }) => (
<div className={cn(className, 'flex items-center justify-center w-5 h-5 bg-white border border-gray-100 text-xs font-medium text-gray-500 rounded ml-3')}>🔥</div>
item.provider === DataSourceProvider.fireCrawl
? (
<div className={cn(className, 'flex items-center justify-center w-5 h-5 bg-white border border-gray-100 text-xs font-medium text-gray-500 rounded ml-3')}>🔥</div>
)
: (
<div className={cn(className, 'flex items-center justify-center w-5 h-5 bg-white border border-gray-100 text-xs font-medium text-gray-500 rounded ml-3')}>
<span className={s.jinaLogo} />
</div>
)
),
name: 'Firecrawl',
name: item.provider === DataSourceProvider.fireCrawl ? 'Firecrawl' : 'Jina Reader',
isActive: true,
}))}
onRemove={handleRemove(DataSourceProvider.fireCrawl)}
onRemove={handleRemove(provider)}
/>
{isShowConfig && (
{configTarget === DataSourceProvider.fireCrawl && (
<ConfigFirecrawlModal onSaved={handleAdded} onCancel={hideConfig} />
)}
{configTarget === DataSourceProvider.jinaReader && (
<ConfigJinaReaderModal onSaved={handleAdded} onCancel={hideConfig} />
)}
</>
)