173 lines
6.4 KiB
TypeScript
173 lines
6.4 KiB
TypeScript
import { INode, INodeData, INodeParams } from '../../../src/Interface'
|
|
import { TextSplitter } from 'langchain/text_splitter'
|
|
import { CheerioWebBaseLoader, WebBaseLoaderParams } from 'langchain/document_loaders/web/cheerio'
|
|
import { test } from 'linkifyjs'
|
|
import { parse } from 'css-what'
|
|
import { webCrawl, xmlScrape } from '../../../src'
|
|
import { SelectorType } from 'cheerio'
|
|
|
|
class Cheerio_DocumentLoaders implements INode {
|
|
label: string
|
|
name: string
|
|
version: number
|
|
description: string
|
|
type: string
|
|
icon: string
|
|
category: string
|
|
baseClasses: string[]
|
|
inputs: INodeParams[]
|
|
|
|
constructor() {
|
|
this.label = 'Cheerio Web Scraper'
|
|
this.name = 'cheerioWebScraper'
|
|
this.version = 1.1
|
|
this.type = 'Document'
|
|
this.icon = 'cheerio.svg'
|
|
this.category = 'Document Loaders'
|
|
this.description = `Load data from webpages`
|
|
this.baseClasses = [this.type]
|
|
this.inputs = [
|
|
{
|
|
label: 'URL',
|
|
name: 'url',
|
|
type: 'string'
|
|
},
|
|
{
|
|
label: 'Text Splitter',
|
|
name: 'textSplitter',
|
|
type: 'TextSplitter',
|
|
optional: true
|
|
},
|
|
{
|
|
label: 'Get Relative Links Method',
|
|
name: 'relativeLinksMethod',
|
|
type: 'options',
|
|
description: 'Select a method to retrieve relative links',
|
|
options: [
|
|
{
|
|
label: 'Web Crawl',
|
|
name: 'webCrawl',
|
|
description: 'Crawl relative links from HTML URL'
|
|
},
|
|
{
|
|
label: 'Scrape XML Sitemap',
|
|
name: 'scrapeXMLSitemap',
|
|
description: 'Scrape relative links from XML sitemap URL'
|
|
}
|
|
],
|
|
optional: true,
|
|
additionalParams: true
|
|
},
|
|
{
|
|
label: 'Get Relative Links Limit',
|
|
name: 'limit',
|
|
type: 'number',
|
|
optional: true,
|
|
default: '10',
|
|
additionalParams: true,
|
|
description:
|
|
'Only used when "Get Relative Links Method" is selected. Set 0 to retrieve all relative links, default limit is 10.',
|
|
warning: `Retrieving all links might take long time, and all links will be upserted again if the flow's state changed (eg: different URL, chunk size, etc)`
|
|
},
|
|
{
|
|
label: 'Selector (CSS)',
|
|
name: 'selector',
|
|
type: 'string',
|
|
description: 'Specify a CSS selector to select the content to be extracted',
|
|
optional: true,
|
|
additionalParams: true
|
|
},
|
|
{
|
|
label: 'Metadata',
|
|
name: 'metadata',
|
|
type: 'json',
|
|
optional: true,
|
|
additionalParams: true
|
|
}
|
|
]
|
|
}
|
|
|
|
async init(nodeData: INodeData): Promise<any> {
|
|
const textSplitter = nodeData.inputs?.textSplitter as TextSplitter
|
|
const metadata = nodeData.inputs?.metadata
|
|
const relativeLinksMethod = nodeData.inputs?.relativeLinksMethod as string
|
|
const selectedLinks = nodeData.inputs?.selectedLinks as string[]
|
|
let limit = parseInt(nodeData.inputs?.limit as string)
|
|
|
|
let url = nodeData.inputs?.url as string
|
|
url = url.trim()
|
|
if (!test(url)) {
|
|
throw new Error('Invalid URL')
|
|
}
|
|
|
|
const selector: SelectorType = nodeData.inputs?.selector as SelectorType
|
|
|
|
let params: WebBaseLoaderParams = {}
|
|
if (selector) {
|
|
parse(selector) // comes with cheerio - will throw error if invalid
|
|
params['selector'] = selector
|
|
}
|
|
|
|
async function cheerioLoader(url: string): Promise<any> {
|
|
try {
|
|
let docs = []
|
|
const loader = new CheerioWebBaseLoader(url, params)
|
|
if (textSplitter) {
|
|
docs = await loader.loadAndSplit(textSplitter)
|
|
} else {
|
|
docs = await loader.load()
|
|
}
|
|
return docs
|
|
} catch (err) {
|
|
if (process.env.DEBUG === 'true') console.error(`error in CheerioWebBaseLoader: ${err.message}, on page: ${url}`)
|
|
}
|
|
}
|
|
|
|
let docs = []
|
|
if (relativeLinksMethod) {
|
|
if (process.env.DEBUG === 'true') console.info(`Start ${relativeLinksMethod}`)
|
|
if (!limit) limit = 10
|
|
else if (limit < 0) throw new Error('Limit cannot be less than 0')
|
|
const pages: string[] =
|
|
selectedLinks && selectedLinks.length > 0
|
|
? selectedLinks.slice(0, limit)
|
|
: relativeLinksMethod === 'webCrawl'
|
|
? await webCrawl(url, limit)
|
|
: await xmlScrape(url, limit)
|
|
if (process.env.DEBUG === 'true') console.info(`pages: ${JSON.stringify(pages)}, length: ${pages.length}`)
|
|
if (!pages || pages.length === 0) throw new Error('No relative links found')
|
|
for (const page of pages) {
|
|
docs.push(...(await cheerioLoader(page)))
|
|
}
|
|
if (process.env.DEBUG === 'true') console.info(`Finish ${relativeLinksMethod}`)
|
|
} else if (selectedLinks && selectedLinks.length > 0) {
|
|
if (process.env.DEBUG === 'true') console.info(`pages: ${JSON.stringify(selectedLinks)}, length: ${selectedLinks.length}`)
|
|
for (const page of selectedLinks) {
|
|
docs.push(...(await cheerioLoader(page)))
|
|
}
|
|
} else {
|
|
docs = await cheerioLoader(url)
|
|
}
|
|
|
|
if (metadata) {
|
|
const parsedMetadata = typeof metadata === 'object' ? metadata : JSON.parse(metadata)
|
|
let finaldocs = []
|
|
for (const doc of docs) {
|
|
const newdoc = {
|
|
...doc,
|
|
metadata: {
|
|
...doc.metadata,
|
|
...parsedMetadata
|
|
}
|
|
}
|
|
finaldocs.push(newdoc)
|
|
}
|
|
return finaldocs
|
|
}
|
|
|
|
return docs
|
|
}
|
|
}
|
|
|
|
module.exports = { nodeClass: Cheerio_DocumentLoaders }
|