Skip to content

Commit

Permalink
Hotfix pdf (#3306)
Browse files Browse the repository at this point in the history
* fix: remove blank space between words when use pdf-loader

* fix: add skipBlank option in pdf-loader

* Change PDF loader instance method name, add docs

* Formatting

---------

Co-authored-by: ppxu <[email protected]>
  • Loading branch information
jacoblee93 and ppxu authored Nov 17, 2023
1 parent bb9fb67 commit 6d91724
Show file tree
Hide file tree
Showing 5 changed files with 73 additions and 11 deletions.
Original file line number Diff line number Diff line change
Expand Up @@ -49,6 +49,21 @@ const loader = new PDFLoader("src/document_loaders/example_data/example.pdf", {
});
```

## Eliminating extra spaces

PDFs come in many varieties, which makes reading them a challenge. The loader parses individual text elements and joins them together with a space by default, but
if you are seeing excessive spaces, this may not be the desired behavior. In that case, you can override the separator with an empty string like this:

```typescript
import { PDFLoader } from "langchain/document_loaders/fs/pdf";

const loader = new PDFLoader("src/document_loaders/example_data/example.pdf", {
parsedItemSeparator: "",
});

const docs = await loader.load();
```

## Loading directories

import CodeBlock from "@theme/CodeBlock";
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -29,11 +29,25 @@ npm install pdfjs-dist
```typescript
import { WebPDFLoader } from "langchain/document_loaders/web/pdf";

const loader = new WebPDFLoader(
"src/document_loaders/example_data/example.pdf",
{
// you may need to add `.then(m => m.default)` to the end of the import
pdfjs: () => import("pdfjs-dist/legacy/build/pdf.js"),
}
);
const blob = new Blob(); // e.g. from a file input

const loader = new WebPDFLoader(blob, {
// you may need to add `.then(m => m.default)` to the end of the import
pdfjs: () => import("pdfjs-dist/legacy/build/pdf.js"),
});
```

## Eliminating extra spaces

PDFs come in many varieties, which makes reading them a challenge. The loader parses individual text elements and joins them together with a space by default, but
if you are seeing excessive spaces, this may not be the desired behavior. In that case, you can override the separator with an empty string like this:

```typescript
import { WebPDFLoader } from "langchain/document_loaders/web/pdf";

const blob = new Blob(); // e.g. from a file input

const loader = new WebPDFLoader(blob, {
parsedItemSeparator: "",
});
```
Original file line number Diff line number Diff line change
Expand Up @@ -16,6 +16,10 @@ npm install pdf-parse

```typescript
import { PDFLoader } from "langchain/document_loaders/fs/pdf";
// Or, in web environments:
// import { WebPDFLoader } from "langchain/document_loaders/web/pdf";
// const blob = new Blob(); // e.g. from a file input
// const loader = new WebPDFLoader(blob);

const loader = new PDFLoader("src/document_loaders/example_data/example.pdf");

Expand Down Expand Up @@ -52,3 +56,18 @@ const loader = new PDFLoader("src/document_loaders/example_data/example.pdf", {
pdfjs: () => import("pdfjs-dist/legacy/build/pdf.js"),
});
```

## Eliminating extra spaces

PDFs come in many varieties, which makes reading them a challenge. The loader parses individual text elements and joins them together with a space by default, but
if you are seeing excessive spaces, this may not be the desired behavior. In that case, you can override the separator with an empty string like this:

```typescript
import { PDFLoader } from "langchain/document_loaders/fs/pdf";

const loader = new PDFLoader("src/document_loaders/example_data/example.pdf", {
parsedItemSeparator: "",
});

const docs = await loader.load();
```
11 changes: 9 additions & 2 deletions langchain/src/document_loaders/fs/pdf.ts
Original file line number Diff line number Diff line change
Expand Up @@ -11,13 +11,20 @@ export class PDFLoader extends BufferLoader {

private pdfjs: typeof PDFLoaderImports;

protected parsedItemSeparator: string;

constructor(
filePathOrBlob: string | Blob,
{ splitPages = true, pdfjs = PDFLoaderImports } = {}
{
splitPages = true,
pdfjs = PDFLoaderImports,
parsedItemSeparator = " ",
} = {}
) {
super(filePathOrBlob);
this.splitPages = splitPages;
this.pdfjs = pdfjs;
this.parsedItemSeparator = parsedItemSeparator;
}

/**
Expand Down Expand Up @@ -76,7 +83,7 @@ export class PDFLoader extends BufferLoader {
}
}

const text = textItems.join(" ");
const text = textItems.join(this.parsedItemSeparator);

documents.push(
new Document({
Expand Down
11 changes: 9 additions & 2 deletions langchain/src/document_loaders/web/pdf.ts
Original file line number Diff line number Diff line change
Expand Up @@ -12,14 +12,21 @@ export class WebPDFLoader extends BaseDocumentLoader {

private pdfjs: typeof PDFLoaderImports;

protected parsedItemSeparator: string;

constructor(
blob: Blob,
{ splitPages = true, pdfjs = PDFLoaderImports } = {}
{
splitPages = true,
pdfjs = PDFLoaderImports,
parsedItemSeparator = " ",
} = {}
) {
super();
this.blob = blob;
this.splitPages = splitPages ?? this.splitPages;
this.pdfjs = pdfjs;
this.parsedItemSeparator = parsedItemSeparator;
}

/**
Expand Down Expand Up @@ -61,7 +68,7 @@ export class WebPDFLoader extends BaseDocumentLoader {
lastY = item.transform[5];
}
}
const text = textItems.join(" ");
const text = textItems.join(this.parsedItemSeparator);

documents.push(
new Document({
Expand Down

1 comment on commit 6d91724

@vercel
Copy link

@vercel vercel bot commented on 6d91724 Nov 17, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Successfully deployed to the following URLs:

langchainjs-docs – ./docs/core_docs/

langchainjs-docs-langchain.vercel.app
langchainjs-docs-ruddy.vercel.app
langchainjs-docs-git-main-langchain.vercel.app
js.langchain.com

Please sign in to comment.