PDFLoader is not working - Cannot find package 'pdf-parse'

[kulaone]

When loading the library the PDFLoader won't work unless runing a npm install inside the library.
This is because the package: "pdf-parse": "^1.1.1", is not found.

Error [ERR_MODULE_NOT_FOUND]: Cannot find package 'pdf-parse' imported from /home/runner/LangChain/node_modules/langchain/dist/document_loaders/pdf.js

Can we add it to the "dependencies" area in package.json?

[nfcampos]

Hi, you can fix this by installing pdf-parse, by running npm i pdf-parse or yarn add pdf-parse in your project. pdf-parse is set as a peer dependency because not everyone using langchain needs it.

[kulaone]

Hi @nfcampos,
Hi, thanks for your response. I understand that pdf-parse is set as a peer dependency, it would be helpful to have that information documented somewhere. As someone who may not be familiar with all the dependencies required for LangChain, it would be helpful to know what I need to install in order for each part of the library to work properly.

[nfcampos]

We're actively working on improving the docs, will make sure to include this, thanks for letting us know

[Lionardo]

I agree with @kulaone this needs to be documented. Just run into this issue myself.

[nfcampos]

Hi, we have now added this to the docs, see https://js.langchain.com/docs/modules/indexes/document_loaders/examples/file_loaders/pdf