galterlibrary / invenio-subjects-mesh Goto Github PK

See https://docs.github.com/en/code-security/dependabot (and example implementations in our private repos)

Environment: Invenio v9.1

Problem: It is not possible to select MeSH terms that are both headings and subheadings, for example, Blood, Cytology, or Genetics

Steps to Reproduce:

1. Set up InvenioRDM and install invenio-subjects-mesh
2. Log in and start a new upload
3. In the 'Search for a subject by name' box, type 'blood' or 'cytology' or 'genetics'

Results:

• The autocomplete dropdown populates with MeSH headings that use the blood, cytology, or genetics subheadings.
• The list of heading suggestions in the list begin with the letter "H"; the sort order is non-obvious.
• The length of the drop-down is limited, so the main headings for these words never appear and therefore cannot be selected and added to a record.
• These headings can only be found if you type at least four letters of an allowable subheading for the main headings; the main headings cannot be used as standalone headings.

Expected Results:
Main headings should appear first in the drop down, before any use as subheadings. This seems like an index weighting issue.

Suggestions:
Subheadings should perhaps not appear at all. Please consider releasing a version of this vocabulary that only includes main headings. There are ~30,000 MeSH headings, which is plenty for many users' needs, and the inclusion of heading-subheading pairs blows that up to something like ~650,000 terms, which seems excessive and requires more expertise to apply correctly.

This may or may not be with datastreams API. Mostly about terms that change.