Closed sfarnel closed 3 years ago
Thanks very much @jchartrand Not surprising that there are few items with all of this data (UofA has it but it's not yet part of this new set of data); this is why we are holding off on widgets or search fields for these. That will come in time!
For now:
@danydvd will be adding also a field for the link back to the original item in the home repository so we will have this soon as well
If we can start with this we can have a look and see of we like the presentation
Okay, adding the fields to item level page now.
Quick question: Could there be multiple advisors, departments, or disciplines for a single record?
Should I assume that there could be multiples at some point, and so when outputting them, separate them by a semi-colon?
It’s fairly straightforward because the solr fields are all arrays, even when there is a single value. I basically just iterate over all values in the array, whether that is just one value, or many. And then join them together with semi-colons.
On Sep 4, 2020, at 12:11 PM, Sharon Farnel notifications@github.com wrote:
Thanks very much @jchartrand https://github.com/jchartrand Not surprising that there are few items with all of this data (UofA has it but it's not yet part of this new set of data); this is why we are holding off on widgets or search fields for these. That will come in time!
For now:
please include the fields even where there is no data; leave the space blank (that is, do not use n/a or any other info; this is what is currently the case when there are no subjects) please use labels Supervisor (for advisor data), Department or Faculty (for department data) and Discipline (for discipline) they can be placed in that order, after Subjects but before Abstract @danydvd https://github.com/danydvd will be adding also a field for the link back to the original item in the home repository so we will have this soon as well
If we can start with this we can have a look and see of we like the presentation
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/jchartrand/can-link/issues/6#issuecomment-687244395, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAEFSXOQQU4UCGEDDLXSLDDSEEGSZANCNFSM4P7YXHXQ.
@sfarnel @danydvd
I just noticed that the department values look like so:
"['Department of English Language and Literature']"
i.e., they are enclosed in an extra [' and ']
Those bits should probably be stripped out as part of the original data ingest into SOLR (i.e, in Danoosh's script?)
It is possible that there could be multiples. Would it be possible to list them on separate lines rather than separate by semicolons? At least to start and we can see how it looks?
Thanks @jchartrand Yea, agreed re: extraneous punctuation. @danydvd can you fix? Thanks!
@sfarnel @danydvd
New fields have been added - see, for example, the single record that comes back in a search for 'arsenic' with author 'Lori'
The direct link to the record is:
http://206.167.181.124/record/81bd7da7a9d4efcdd1df5ebe2b4f93a2
but the direct link doesn't seem to be working at the moment. I get nginx error:
404 Not Found nginx/1.4.6 (Ubuntu)
Which I thought we'd fixed by redirecting all urls to the index.html file, but it seems not to be working now?
Do you want any of those three fields (dept, discipline, advisor) added to the popup that appears on the search results page?
Thanks @jchartrand this looks great! I think given the inconsistent presence of these, let's leave them for the full item screen for now. @danydvd would you be able to investigate the direct link issue that seems to have popped up again? Thanks!
@jchartrand Danoosh says that the URL to the original item in its home repository is in new data and will be indexed; can we add this to the item page (just above abstract) with the label Link to fulltext that would be great; thanks!
@sfarnel Yes, no problem. @danydvd, just let me know when the new data has been indexed.
@sfarnel @jchartrand just wanted to give you an update on the re-indexing process. The triplestore has north of 120K items (including 1000+ from U of A). The script that generated the JSON for SOLR that used to take a few minutes (when the size of the triplesotre was around 87K) is now running for more than 2 hours. I have been looking at the graphDB logs and it seems that it is hitting a snag at some point which brings down the speed to almost zero. I will try a few more times with different methods. If unsuccessful, I will separate the repositories by institutions (this would work for re-indexing but is not a viable option for the long run).
Thanks for the update Danoosh. I agree; this could work in the near term but we'll have to figure out that piece for the long term.
@jchartrand @sfarnel The issue was actually my query, a single character (url vs uri) sends the query into an infinite loop. The new core has been indexed (CanLink-new-4). The department text is corrected and the link to original item page has been added (field name: link).
Awesome; thanks @danydvd Those pesky little things!
@danydvd I tried switching to CanLink-new-4, but it generates an error when trying to find the University of Alberta entry in the universities mapping file. Alberta doesn't seem to be in there:
http://206.167.181.124/institutions.json
I'll leave it as CanLink-new-3 for now.
@jchartrand My bad. I was adding U of A manually before (as we did not have any data in the new index) and I forgot to do that after sorting. It is there now.
Thanks Danoosh. @jchartrand can you update to the new core? That way we should be able to test various things with the item page and hopefully call it 'done'. Thanks!
Done - updated to CanLink-new-4 and added the link to the record page.
Thanks @jchartrand the links look good. I think we can close this one for now
On search results page, if you click on See Abstract, the 'Link to Full Record' button (bottom right) doesn't work:
When you click on this you should be taken to the item display page