Better query parsing about fatcat-scholar HOT 3 OPEN

Here is a Google Scholar blog post about detecting reference strings: https://scholar.googleblog.com/2016/01/quickly-lookup-references.html

The jargon-y term for this use case is "known item lookup"

from fatcat-scholar.

An initial version of this has been implemented and is live. Testing and iteration probably needed.

from fatcat-scholar.

Some user queries are getting re-written poorly with the current system:

"journal:" Post Communist Economies "year:" 2021
"Title:" A multi-speed fiscal "Europe?" Fiscal rules and fiscal performance in the EU former communist countries. It appears to be online content from Post Communist Economies 31Jan 2021. "Link:" "https://www-tandfonline-com.libproxy-imf.imf.org/doi/full/10.1080/14631377.2020.1867432"

The original query was probably:

journal: Post Communist Economies year: 2021

Some of this may be due to copy/paste from other sources? Eg, an email or multi-line record on a website.

For one thing, we probably shouldn't return the re-written (quoted) query, we should return the original query string (in the search box). Any time we rewrite/modify the query, should indicate that it happened though, and link to query documentation.

Other possible improvements or work arounds are to have an "advanced search" page, or to have separate search boxes/options for different types of query. I'd like to try a little more to stick with the "one simple box" experience though.

from fatcat-scholar.