PDF Loader does not take into account pages with no text

[The-Sebastian]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

try {
        // Try to load the PDF document
        const pdfLoader = new PDFLoader(new Blob([fileBuffer]), {
          splitPages: true,
        });
        const docs = await pdfLoader.load();
    
        const numberOfPages = docs.length;
    
        // numberOfPages is missing a few pages from the original pdf

       }

Description

• I am trying to use Langchains pdf loader in order to extract the text from my pdf and then upload that text into my vector database with the page number that belongs to it. My issue is that there are some pages that DO NOT have any text and it seems like the loader removes them from the original. So lets say I have a pdf with 23 pages then I get a new doc of only 20 pages if 3 of them don't have any text.

Here is the line of code in langchainjs that removes pages with no content.

How can I make a request so that a new property is added to the constructor that will allow a developer to choose whether to retain blank pages or not? Here is the parse property in the code of langchain.

System Info

"yarn info langchain"
Mac
Node 18.13.0
yarn version 1.22.18