Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Relevance of search results #46

Closed
eaquigley opened this issue Jul 9, 2014 · 3 comments
Closed

Relevance of search results #46

eaquigley opened this issue Jul 9, 2014 · 3 comments
Assignees

Comments

@eaquigley
Copy link
Contributor


Author Name: Elizabeth Quigley (@eaquigley)
Original Redmine Issue: 3453, https://redmine.hmdc.harvard.edu/issues/3453

Original Assignee: Elda Sotiri


Relevance

tl;dr: Should this return any results? http://dvn-build.hmdc.harvard.edu/search.xhtml?q=data+thisTermDoesNotAppear

This stems from previous discussions on what should be highlighted in the search results and what in the result was hit by the search term. During usability testing, users mentioned they were slightly confused as to why some results appeared based on their search. Am attaching Mike's search results card sketch from before.

During usability testing, a user attempted to narrow a search by adding an additional term. Adding the term did put the dataset she was thinking of at the top of her search results but she received more search results, not fewer. Internally, we've noticed this as well... if you add "thisTermDoesNotAppear" to the end of a search query that does match something (i.e. a search for "data"), you still get results. On a Trello card we were calling this the "something rather than nothing" feature: https://trello.com/c/MAr4d5Ow/17-prototype-search-browse

Stop words are a potential issue. Should they be stripped?

If a dataverse is called "Friday" (Friday Dataverse) and a search for "friday" is done, should it match everything in the "Friday" dataverse, including datasets and files? What about child dataverses? Should all datasets and files in the child dataverse match as well?


Highlighing

As of this writing, highlighting only occurs on the "description" field. We probably want highlighting on more fields.

We've received feedback that the yellow highlighting is too yellow.

3/26/2014 note: This is still something that needs to be worked out. Which fields can a user search on that would pull a result? How do we show those fields in the search results card?

4/4/2014 note: The more I think about this, I think highlighting should only be appearing in the description field. If the search query hits the title or the citation, I don't think there is a need for highlighting. Additionally, since the citation is in a yellow box already, I don't think yellow highlighting on a yellow background would be that appealing to look at. This brings me to another thought, please see the section: Citation. (EQ)

4/14/2014 note: Highlighting has its own ticket at #3737 and has been moved to QA.


Citation

If a user's search query hits a field within the citation, do we need a visual cue to show that is what caused the result to appear? If it's author that is the field hit, then wouldn't the author facet also be showing this? If so, then that would be three places (facet, search query in bar, and in the citation) so do we need an additional thing showing the user the relationship between the result and the query? Something to test. (EQ)

4/15/2014: for beta, we won't be highlighting/bolding anything from the search query that is found in the citation. (EQ)


Showing Relevancy

We need to implement a way for a user to understand why their search query brought back the result if it isn't the title, in the description, or one of the citation fields. The simplest thing to do would have the metadata field (without the search query/term next to it) appear in the results card so I think we should implement that and see how users like it.

This is handled in the highlighting ticket at #3737.


Using a singular term and not getting plurals of it in the results

If a user searches for tree, they only see results for tree when they should be seeing results for tree and trees.


Related issue(s): #392, #484
Redmine related issue(s): 3430, 3807, 3899


@eaquigley
Copy link
Contributor Author


Original Redmine Comment
Author Name: Philip Durbin (@pdurbin)
Original Date: 2014-04-14T18:04:58Z


Liz, as we discussed, can you please split this into subtickets that can be developed, tested, and prioritized separately? Thanks! The highlighting part already has its own ticket: #3737.

@eaquigley eaquigley added this to the Dataverse 4.0: Beta 1 milestone Jul 9, 2014
@eaquigley
Copy link
Contributor Author


Original Redmine Comment
Author Name: Gustavo Durand (@scolapasta)
Original Date: 2014-04-23T16:40:26Z


Checked with Liz, she says she broke it up and put the related tickets in the related issues section.

Can you please verify, and if it is all covered, can we then close this one?

@eaquigley
Copy link
Contributor Author


Original Redmine Comment
Author Name: Elda Sotiri (@esotiri)
Original Date: 2014-04-29T14:10:22Z


Added Feature #3899: should archived version of datasets be searchable. All seem to be covered so I'm closing this ticket.

kcondon pushed a commit that referenced this issue Mar 24, 2020
Update from develop IQSS
vicding-mi added a commit to vicding-mi/dataverse that referenced this issue Mar 19, 2021
* set default when no config is set for signposting
* modification according to reviews
* move long json string from code to bunddle
* allow empty config on the level 2 profile
* revision based on Herbert feedback
* coding style cleanup SignpostingResources
* remove leading comma
* fix capitalize with header name
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

2 participants