Feature Request: search by arxiv-id about ads

[vsudilov]

Hi,

Does making a fielded query against identifier and/or alternate_bibcode (as documented in https://github.com/adsabs/adsabs-dev-api/blob/master/search.md#fields) work for this use case?

[lzkelley]

alternate_bibcode does not work with an arxiv ID, but identifier does! Sorry I missed this in the docs. Thanks for pointing it out.

[jonnybazookatone]

Just to clarify:

• identifer: will contain all identifiers ever used (in theory) arXiv, ADS arXiv bibcode, ADS bibcode, etc.
• alternate_bibcode: will contain the ADS bibcode that was constructed (but modified) from the original arXiv identifier, e.g., 2014arXiv1409.7699E rather than arXiv:1409.7699

You can also use the field map arxiv like:

ads.SearchQuery(q='arxiv:1409.7699')

but in the end the search engine is just mapping arxiv to identifier under the hood....

[lzkelley]

Interesting. I did try just using q='1409.7699', which didn't seem to work; it needs the 'arxiv:' specification? Does that mean its a field-specific query instead of an regex for all fields?

[jonnybazookatone]

Yes, you need the arxiv field. But like I said, all that happens is the parser turns arxiv into identifier. So arxiv:1409.7699 = identifier:1409.7699, I guess it's only useful for readability in your script....

When you do an 'unfielded' search, i.e., q='1409.7699', the query parser has a default rule for what it is searching for. Currently, it does a search combined on author, title, and abstract, with weighting:

• author^1.5 title^1.4 abstract^1.3 all

You can see this here: https://github.com/romanchyla/montysolr/blob/2d2f8e315f427e6ebd3276ffbfb3ffa6ac1a6b3a/contrib/examples/adsabs/server/solr/collection1/solrconfig.xml#L625

[jonnybazookatone]

And forgot to directly answer your second part. identifier is a stored field, and so you're doing a fielded search when you specify arxiv.

You can see a list of the stored and virtual fields here: https://adsabs.github.io/help/search/comprehensive-solr-term-list

[lzkelley]

Very helpful, thanks for the explanations @jonnybazookatone!