fix: Dependency update for unstructured component

[erichare]

This PR makes two changes to the dependencies of the project:

1. It adds the pdf extra package for unstructured, for processing PDFs
2. It bumps the minimum version for huggingface-hub which is required for Unstructured.

[github-actions]

Pull Request Validation Report

This comment is automatically generated by Conventional PR

Whitelist Report

Whitelist	Active	Result
Pull request is a draft and should be ignored	✅	❌
Pull request is made by a whitelisted user and should be ignored	❌	❌
Pull request is submitted by a bot and should be ignored	✅	❌
Pull request is submitted by administrators and should be ignored	❌	❌

Result

Pull request does not satisfy any enabled whitelist criteria. Pull request will be validated.

Validation Report

Validation	Active	Result
All commits in this pull request has valid messages	❌	✅
Pull request does not introduce too many changes	❌	✅
Pull request has a valid title	✅	✅
Pull request has mentioned issues	❌	✅
Pull request has valid branch name	❌	✅
Pull request should have a non-empty body	✅	✅

Result

Pull request satisfies all enabled pull request rules.

_{Last Modified at 26 Aug 24 16:46 UTC}

[aws-amplify-sa-east-1]

This pull request is automatically being deployed by Amplify Hosting (learn more).

Access this pull request here: https://pr-3552.dmtpw4p5recq1.amplifyapp.com

[ogabrielluiz]

LGTM

[nicoloboschi]

Could you explain how the new script nltk_setup is intended to be used ?

[erichare]

Could you explain how the new script nltk_setup is intended to be used ?

Hi @nicoloboschi

the unstructured component, and specifically the pdf parsing tools, use nltk behind the scenes. But they assume that these three packages, punkt etc, are already available on the host - if they aren’t, the component will error and ask you to download them first.

so the basic idea is just that every time we deploy a release that we are sure those nltk packages are on the machine and accessible for the nltk library - does that make sense? Maybe there’s another way to go about it, I.,e, the dockerfile orchestrates it? But that’s the idea

[nicoloboschi]

@erichare could we make it part of the langflow run script?
and at the bootstrap it verifies/download the needed files

The current format is not that helpful in fix/troubleshoot the problem

[erichare]

feat: allow drive loader to read a folder recursively #3572

@nicoloboschi when you get a chance, could you look at the latest commit? I'm not sure whether i put it exactly in the most logical place, but the utility function for downloading the resources i put in setup.py, then the run script imports that function and calls it at startup. If the packages are already available, it'll skip downloading them.

[nicoloboschi]

LGTM