christinach
commented
2 years ago For the existing fields lc_subject_display, subject_unstem_search, subject_topic_facet :
In production: lc_subject_display: [ "Refugees—Drama", "Africans—Europe—Drama", "Undocumented immigrants—Europe—Drama", "Africa—Emigration and immigration—Drama" ],
Proposed: lc_subject_display: [ "Refugees—Drama", "Africans—Europe—Drama", "Africa—Emigration and immigration—Drama"]
In production: "subject_unstem_search": [ "Refugees—Drama", "Africans—Europe—Drama", "Illegal aliens—Europe—Drama", "Africa—Emigration and immigration—Drama", "Undocumented immigrants—Europe—Drama" ],
Proposed: subject_unstem_search: lc_subjects + local_subjects subject_unstem_search: [ "Refugees—Drama", "Africans—Europe—Drama", "Illegal aliens—Europe—Drama", "Africa—Emigration and immigration—Drama", "Undocumented immigrants—Europe—Drama" ]
In production: "subject_topic_facet": [ "Refugees", "Africans", "Europe", "Undocumented immigrants", "Africa", "Emigration and immigration" ],
Proposed "subject_topic_facet": [ "Refugees", "Africans", "Europe", "Undocumented immigrants", "Africa", "Emigration and immigration" ]
- The catalog will need to use subject_display for 'Subjects' and not lc_subject_display, after a full reindex; So that all the records have the same fields updated with the correct terms. This will have a downtime in the search for some records that they get updated thorugh the incremental job, since we don't have a separate traject_config to use during indexing. see https://github.com/pulibrary/bibdata/issues/1626
- If we want to implement this now we would need to use a different name for the 'lc_subject_display' , for example lc_subject_display_new, so that we dont break the catalog which is looking lc_subject_display. Then when the time comes to run a full reindex we will do the adjustment to the catalog to look subject_display.
- In a second full reindex we can update the lc_subject_display and remove the lc_subject_display_new without any downtime in the catalog.
We can discuss this further in our meeting; if we want to close this issue in this workcycle, we will follow @mzelesky's proposal but we will call it 'lc_subject_display_new'. When we decide to run a full reindex:
- we create the new catalog index collection;
- we swap the index;
- we have an orangelight PR ready to look for 'subject_display' instead of 'lc_subject_display' for 'Subjects' in the record page. (open a new issue for 3. when we decide how to proceed with this issue )
lc_subjects are the LC subject headings found in the record. lc_subject_display is the filtered list of LC subject headings. local_subjects is the new subjects. subject_display is lc_subject_display and local_subjects
work in progress branch https://github.com/pulibrary/bibdata/commit/61e387715c5ffeddb8bf54a2ffdc781d77a24888
We will also have to change the Orangelight fields in the codebase to look subject_display and not lc_subject_display