Make search facets configurable

[eaquigley]

---

Author Name: Philip Durbin (@pdurbin)
Original Redmine Issue: 3490, https://redmine.hmdc.harvard.edu/issues/3490
Original Date: 2014-02-03
Original Assignee: Kevin Condon

---

Need:

• new tables to store interesting facets based per dataverse
• persist and retrieve preferred order of facet per dataverse
• have search method on the indexing bean receive a dataverse parameter to retrieve only interesting facets for that dataverse (in the right order)

Facet list for what should be shown at the root level of the Harvard Dataverse as of 2/24/2014 (this should also be the default list of facets if a dataverse admin has not configured a unique list for their dataverse):

• Type (Dataverse, Dataset, File)
• Host Dataverse
• Author
• Affiliation (dataverse "Affiliation" and dataset "Author Affiliation" combined into a single facet)
• Distributor (not "Distributor Name", see Support Large Guestbooks #3609)
• Release Date View File Versions / Difference Summary #3592 (Dataset release date Guestbook "name" field is prefilled to "Guest" #3628 and Dataverse release date UNF Fixes #3629 combined into one facet)
• Keyword
• -Topic Classfication- Per 3/4/14 discussions now named Subject adminEmail field invalid in OAI-PMH Identify response #3619
• File Type

Please note that as of 2014-02-26, the display order of facets that come from datasets (i.e. author, keyword) can be controlled but the display order of facets that come from dataverses and files are hardcoded into SearchServiceBean.

---

Related issue(s): #70, #89, #180, #197
Redmine related issue(s): 3478, 3498, 3592, 3609

---

[eaquigley]

---

Original Redmine Comment
Author Name: Philip Durbin (@pdurbin)
Original Date: 2014-03-04T21:44:52Z

---

Kevin, as we discussed, I'm flipping this ticket over to you with this note to explain the divergence in what was requested by Merce via email on 2014-02-14 (curated by Elizabeth) and what I expect you to see as of commit c88b8da after having run scripts/database/facetpopulate.sql and done an "index all".

Here's the original list (order matters) from Merce:

• Type (Dataverse, Dataset, File)
• Host Dataverse
• Author
• Affiliation
• Distributor
• Release Date
• Keyword
• Topic Classification Per 3/4/14 discussions now named Subject
• File Type

Here's what I expect you to see (order matters):

• Type (Dataverse, Dataset, File)
• Host Dataverse
• Affiliation
• Author (see note 1)
• Distributor Name (see note 2)
• Keyword
• Topic Classification (see note 3)
• Production Date (see note 4)
• Distribution Date (see note 4)
• File Type

Please note the following:

1. Please disregard the values under Author. You'll see FIXMEs for now.
2. Stephen is removing "Name" from "Distributor Name" in https://redmine.hmdc.harvard.edu/issues/3609#note-4 which will probably require an update to the database.
3. "Topic Classification" is supposed to turn into "Subject" in adminEmail field invalid in OAI-PMH Identify response #3619
4. "Production Date" and "Distribution Date" are placeholders for "Release Date" which is not currently actionable due to blocks by Guestbook "name" field is prefilled to "Guest" #3628 and UNF Fixes #3629 per View File Versions / Difference Summary #3592.

[eaquigley]

---

Original Redmine Comment
Author Name: Elda Sotiri (@esotiri)
Original Date: 2014-03-06T20:58:54Z

---

Facets are now read from the database so they are configurable. The issues with the current list are identified in separate tickets. closing

[eaquigley]

---

Original Redmine Comment
Author Name: Philip Durbin (@pdurbin)
Original Date: 2014-03-10T19:36:26Z

---

Philip Durbin wrote:

Kevin, as we discussed, I'm flipping this ticket over to you with this note to explain the divergence in what was requested by Merce via email on 2014-02-14 (curated by Elizabeth) and what I expect you to see as of commit c88b8da after having run scripts/database/facetpopulate.sql and done an "index all".

I just wanted to mention that my scripts/database/facetpopulate.sql script should no longer be run as it uses hard-coded IDs for dataset fields.

The new way to populate the list of facets to be shown is to run root_dataverse_setup.sql from the "[dataverse] Modifications to getting reference data" message sent Fri, 7 Mar 2014 12:28:04 -0500.

We'll be updating the Dev Guide with the new procedure.