Would it be possible to pull PDFs using the oaDOI API? Obviously these should be pulled as second preference to the originals as they may be pre-prints. It would be very useful where one doesn’t have access.
Or we could just pull straight from Sci-Hub… which is what I normally end up doing anyway!
We’ll look into that. We already fall back to open access repositories like PubMed Central when the article is restricted on the publisher’s page.
For obvious reasons, we can’t directly import from Sci-Hub.
Would it be possible to allow the user to customize which APIs are used to retrieve papers?
Getting full text PDFs is complex. it’s not about" customizing APIs" I’m afraid, as there are practically none. It’s a mix of various complicated crawling strategies.
If Paperpile does not get a PDF and you know a way to get it, let us know and we look into it.
Please consider allowing users to specify an alternative URL for PDF fetching.
To be specific, I imagine that paperpile could allow a user to prepend or append a URL like http://localhost:8888/papers/ to a search list. That endpoint would receive a POST’d payload with the full citation info and be expected to return a PDF. (Or, perhaps even a set of PDFs and/or supporting data.)
This feature would enable users to customize paper fetching, caching, sharing, and transport to suit their environments.
I think this is what @slowkow meant.
Thanks for considering.
re-up’n this I would like the ability to add like a proxy but like any site