arXiv import no longer works by either arXiv ID or DOI

[axel-loewe]

Description

Importing arXiv papers into iLibrarian no longer works.

This fails in both cases:

1. when providing the arXiv ID
2. when providing the DOI

This has been broken for a while and appears to affect arXiv papers generally, not just a single record.

Expected behavior

Entering a valid arXiv ID or a valid arXiv DOI should fetch the paper metadata and allow adding the paper to the library.

Actual behavior

No paper is imported / metadata lookup fails when using either:

• an arXiv ID
• an arXiv DOI such as 10.48550/arXiv.xxxxx.xxxxx

Likely cause

A likely reason is that arXiv DOIs are now under the 10.48550 prefix and are registered via DataCite, while iLibrarian appears to rely on Crossref for DOI lookup. As a result, valid arXiv DOIs may not resolve through the current lookup path.

For arXiv ID lookup, the importer may also rely on an outdated arXiv metadata endpoint or parser.

Examples

Examples that should work:

• arXiv ID: 1706.03762
• DOI: 10.48550/arXiv.1706.03762

[mkucej]

Thanks for your report. There are two issues here.

1. arxiv ID import fails using Import UID

This issue is caused by a change in arXiv’s API policy. They appear to have disabled CORS support, which prevents the browser-based "Import UID" tool from connecting. I have reached out to their team for clarification but have not yet received a response. In the meantime, please use Import > arXiv and search by the ID field. This method uses a server-side PHP and is unaffected by the CORS issue.

2. arxiv DOI import fails using Import UID

This is currently not supported. We'll consider this enhancement for a future update.

[axel-loewe]

Thanks, the workaround with Import > arXiv works 👍

[mkucej]

Arxiv replied that they aren't going to fix the CORS access.

TODO: rewrite arxiv UID import to server-side.