feat: additional metadata added to Spider tool (#2923)

* feat: additional metadata added to Spider tool

* console.log error instead of throwing to keep flow

* add omit

* pnpm lint
This commit is contained in:
William Espegren
2024-08-02 23:48:48 +02:00
committed by GitHub
parent 64d9de205f
commit 99ef9214c7
@@ -1,3 +1,4 @@
import { omit } from 'lodash'
import { TextSplitter } from 'langchain/text_splitter'
import { Document, DocumentInterface } from '@langchain/core/documents'
import { BaseDocumentLoader } from 'langchain/document_loaders/base'
@@ -10,6 +11,7 @@ interface SpiderLoaderParameters {
apiKey?: string
mode?: 'crawl' | 'scrape'
limit?: number
additionalMetadata?: Record<string, unknown>
params?: Record<string, unknown>
}
@@ -18,11 +20,12 @@ class SpiderLoader extends BaseDocumentLoader {
private url: string
private mode: 'crawl' | 'scrape'
private limit?: number
private additionalMetadata?: Record<string, unknown>
private params?: Record<string, unknown>
constructor(loaderParams: SpiderLoaderParameters) {
super()
const { apiKey, url, mode = 'crawl', limit, params } = loaderParams
const { apiKey, url, mode = 'crawl', limit, additionalMetadata, params } = loaderParams
if (!apiKey) {
throw new Error('Spider API key not set. You can set it as SPIDER_API_KEY in your .env file, or pass it to Spider.')
}
@@ -31,6 +34,7 @@ class SpiderLoader extends BaseDocumentLoader {
this.url = url
this.mode = mode
this.limit = Number(limit)
this.additionalMetadata = additionalMetadata
this.params = params
}
@@ -61,7 +65,10 @@ class SpiderLoader extends BaseDocumentLoader {
(doc) =>
new Document({
pageContent: doc.content || '',
metadata: { source: doc.url }
metadata: {
...(this.additionalMetadata || {}),
source: doc.url
}
})
)
}
@@ -125,6 +132,14 @@ class Spider_DocumentLoaders implements INode {
type: 'number',
default: 25
},
{
label: 'Additional Metadata',
name: 'additional_metadata',
type: 'json',
description: 'Additional metadata to be added to the extracted documents',
optional: true,
additionalParams: true
},
{
label: 'Additional Parameters',
name: 'params',
@@ -134,6 +149,17 @@ class Spider_DocumentLoaders implements INode {
placeholder: '{ "anti_bot": true }',
type: 'json',
optional: true
},
{
label: 'Omit Metadata Keys',
name: 'omitMetadataKeys',
type: 'string',
rows: 4,
description:
'Each document loader comes with a default set of metadata keys that are extracted from the document. You can use this field to omit some of the default metadata keys. The value should be a list of keys, seperated by comma. Use * to omit all metadata keys execept the ones you specify in the Additional Metadata field',
placeholder: 'key1, key2, key3.nestedKey1',
optional: true,
additionalParams: true
}
]
this.credential = {
@@ -149,18 +175,39 @@ class Spider_DocumentLoaders implements INode {
const url = nodeData.inputs?.url as string
const mode = nodeData.inputs?.mode as 'crawl' | 'scrape'
const limit = nodeData.inputs?.limit as number
let additionalMetadata = nodeData.inputs?.additional_metadata
let params = nodeData.inputs?.params || {}
const credentialData = await getCredentialData(nodeData.credential ?? '', options)
const spiderApiKey = getCredentialParam('spiderApiKey', credentialData, nodeData)
const _omitMetadataKeys = nodeData.inputs?.omitMetadataKeys as string
let omitMetadataKeys: string[] = []
if (_omitMetadataKeys) {
omitMetadataKeys = _omitMetadataKeys.split(',').map((key) => key.trim())
}
if (typeof params === 'string') {
try {
params = JSON.parse(params)
} catch (e) {
throw new Error('Invalid JSON string provided for params')
console.error('Invalid JSON string provided for params')
}
}
if (additionalMetadata) {
if (typeof additionalMetadata === 'string') {
try {
additionalMetadata = JSON.parse(additionalMetadata)
} catch (e) {
console.error('Invalid JSON string provided for additional metadata')
}
} else if (typeof additionalMetadata !== 'object') {
console.error('Additional metadata must be a valid JSON object')
}
} else {
additionalMetadata = {}
}
// Ensure return_format is set to markdown
params.return_format = 'markdown'
@@ -169,6 +216,7 @@ class Spider_DocumentLoaders implements INode {
mode: mode as 'crawl' | 'scrape',
apiKey: spiderApiKey,
limit: limit as number,
additionalMetadata: additionalMetadata as Record<string, unknown>,
params: params as Record<string, unknown>
}
@@ -182,6 +230,20 @@ class Spider_DocumentLoaders implements INode {
docs = await loader.load()
}
docs = docs.map((doc: DocumentInterface) => ({
...doc,
metadata:
_omitMetadataKeys === '*'
? additionalMetadata
: omit(
{
...doc.metadata,
...additionalMetadata
},
omitMetadataKeys
)
}))
return docs
}
}