use pdf-parser
This commit is contained in:
parent
02d8284f58
commit
930449f956
|
|
@ -66,7 +66,8 @@ class Pdf_DocumentLoaders implements INode {
|
||||||
const blob = new Blob([bf])
|
const blob = new Blob([bf])
|
||||||
|
|
||||||
if (usage === 'perFile') {
|
if (usage === 'perFile') {
|
||||||
const loader = new PDFLoader(blob, { splitPages: false })
|
// @ts-ignore
|
||||||
|
const loader = new PDFLoader(blob, { splitPages: false, pdfjs: () => import('pdf-parse/lib/pdf.js/v1.10.100/build/pdf.js') })
|
||||||
if (textSplitter) {
|
if (textSplitter) {
|
||||||
const docs = await loader.loadAndSplit(textSplitter)
|
const docs = await loader.loadAndSplit(textSplitter)
|
||||||
return docs
|
return docs
|
||||||
|
|
@ -75,7 +76,8 @@ class Pdf_DocumentLoaders implements INode {
|
||||||
return docs
|
return docs
|
||||||
}
|
}
|
||||||
} else {
|
} else {
|
||||||
const loader = new PDFLoader(blob)
|
// @ts-ignore
|
||||||
|
const loader = new PDFLoader(blob, { pdfjs: () => import('pdf-parse/lib/pdf.js/v1.10.100/build/pdf.js') })
|
||||||
if (textSplitter) {
|
if (textSplitter) {
|
||||||
const docs = await loader.loadAndSplit(textSplitter)
|
const docs = await loader.loadAndSplit(textSplitter)
|
||||||
return docs
|
return docs
|
||||||
|
|
|
||||||
|
|
@ -27,7 +27,7 @@
|
||||||
"langchain": "^0.0.44",
|
"langchain": "^0.0.44",
|
||||||
"moment": "^2.29.3",
|
"moment": "^2.29.3",
|
||||||
"node-fetch": "2",
|
"node-fetch": "2",
|
||||||
"pdfjs-dist": "^3.5.141",
|
"pdf-parse": "^1.1.1",
|
||||||
"ws": "^8.9.0"
|
"ws": "^8.9.0"
|
||||||
},
|
},
|
||||||
"devDependencies": {
|
"devDependencies": {
|
||||||
|
|
|
||||||
Loading…
Reference in New Issue