ualbertalib / discovery

Discovery is the University of Alberta Libraries' catalogue interface, built using Blacklight
http://search.library.ualberta.ca
12 stars 3 forks source link

[Discovery] clicking author name does a keyword search instead of a phrase search #1329

Open theLinkResolver opened 6 years ago

theLinkResolver commented 6 years ago

Describe the bug Clicking author name in a record does a keyword search of the terms instead of a phrase search. Clicking an additional author name produces the correct behaviour.

To Reproduce e.g. https://library.ualberta.ca/catalog/1928299

  1. Look at the URL for Kirkland, Glen - it is a keyword author search
  2. Look at the URL for Davies, Richard... - it is a phrase search (i.e. with quotes around the entire string)

Expected behavior Author names are based on controlled vocabulary and should be searched with precision. Another field we should implement is the series title, and should be searched with precision

seanluyk commented 6 years ago

Thanks @thelinkresolver, one thing I'd wonder about in this approach are roles...do we want the full string searched in quotes?

theLinkResolver commented 6 years ago

@seanluyk Short answer - no. Only the parts of the string that comprise the authorized access point: subfields a,b,c,d,q

The rest of this response is "enhancement" territory.

It might, however, be interesting to offer users the option of clicking the relator term ($e) which would deliver a search of the string plus that term, to find other works where the person has had the same role (a bit like what we do for subject headings - giving the user the option of where in the heading to click).

A great example use case (credit to Brian!) would be Clint Eastwood, where a user could click Eastwood, Clint, 1930- film director in a record and see, first and foremost, other movies he's directed, rather than his acting-only roles.

A note on relator terms: they appear in RDA records but not pre-RDA records (except where manually added, which does happen). So, a search of this nature would not be comprehensive, but "the future is longer than the past" (again borrowing from Brian), and would be increasingly comprehensive over time. A possible compromise would be for the search to rank higher those records that match the name string + relator term, and then to still deliver records that have the name string only but ranked lower in the search. This might happen anyway based on how the search is constructed and the existing relevance ranking.

seanluyk commented 6 years ago

@theLinkResolver thanks for the clarification/specificity, this is something we can get added to a future sprint and the enhancement at a later date. I tend to agree, the desired behaviour for controlled vocabulary should be precision where possible

theLinkResolver commented 5 years ago

@seanluyk This might be a good candidate for getting done before the code freeze.

seanluyk commented 5 years ago

@theLinkResolver I agree, although I don't know if we'll get there in time. Flagging @weiweishi @murny @pgwillia to ask how difficult this would be to implement?

weiweishi commented 5 years ago

I feel this is a very simple change - need to tweak the way on how we construct the solr query. So if I'm not mistaken, we should be able to add this to our backlog before the freeze.

Weiwei ShiDigital Initiative Applications Librarian

2-10L Cameron Library, University of Alberta 780-492-7802 | weiwei.shi@ualberta.ca "The University of Alberta respectfully acknowledges that we are situated on Treaty 6 territory, traditional lands of First Nations and Métis people."

On Tue, Jan 15, 2019 at 3:28 PM Sean Luyk notifications@github.com wrote:

@theLinkResolver https://github.com/theLinkResolver I agree, although I don't know if we'll get there in time. Flagging @weiweishi https://github.com/weiweishi @murny https://github.com/murny @pgwillia https://github.com/pgwillia to ask how difficult this would be to implement?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/ualbertalib/discovery/issues/1329#issuecomment-454576557, or mute the thread https://github.com/notifications/unsubscribe-auth/AB8-ftG3TJdm_mzMgYgTearbQkZbKaEjks5vDlYWgaJpZM4Xk-Q3 .

weiweishi commented 5 years ago

we will also implement this with the series title.

pgwillia commented 5 years ago

There's a bunch of history around this looks like issues #580, #827 and #1068 are involved. To summarize this is where it's at:

label solr field marc mapping type of search
Series title_series_t 440anpv:490av title phrase
Author author_display 100abcdq:110[a-z]:111[a-z] author keyword
Additional authors/performers author_addl_t 700abdcegqu:710abcdegnu:711acdegjnqu general phrase search

My understanding of what is required is that all of these special cases become phrase queries against the specific Solr field that they came from? Similar to how the advanced search works except as phrases instead of keywords?

https://search.library.ualberta.ca/catalog/667528 is an example with a number of these.