Add a vocabulary core.#103
Conversation
tropicrainforest
left a comment
There was a problem hiding this comment.
Thanks for the design. One design question that came to mind: how do we want to handle broad or vague user queries?
For example, if a user searches for temperature, should the vocab core return tas, ta, tasmax, tasmin? Or, if a user searches for flux, should it return precipitation flux, radiation flux, and other flux-related variables?
I’m thinking whether we need an additional way to rank or distinguish these results so the resolver does not silently pick a weak match. For example, curated mappings could be returned with higher priority, while non-curated or crawler-observed matches are returned as lower-priority candidates. I'm not sure yet about a proper solution for this
| instdir="${pair#*:}" | ||
| copy_conf_files "$CONFIG_DIR" "$instdir/conf" | ||
| if [ "$core" = "vocab" ];then | ||
| copy_conf_files "$CONFIG_DIR/freva_vocab" "$instdir/conf" |
There was a problem hiding this comment.
I tried this locally, and it looks like the vocab core only picks up its dedicated managed-schema.xml if solr/freva_vocab/ is copied into the image at build time. Otherwise, it falls back to the default managed schema. I think we should copy it in the Dockerfile first.
@tropicrainforest could you have a look this would be my proposal for a vocabulary core.