A first step of getting opensearch working is to create a json document with the search data.
I think first steps are:
[x] Create class or object handle open search document creation
[x] Start by including all the current fields being added to search in the database
[x] Convert multiple items to JSON arrays instead of munging together (where appropriate)
[x] Add appropriate python tests for document creation
For the future, we may want to think about some additional tasks that would enhance search, but I think it's best to get a working search first and then enhance as a separate stage.
Stuff like adding specific fields for many more values or displaying search facets/filters and other more modern search techniques. However, these things may come at a later time and it's not nearly as straightforward to implement in EZID as in an application like Dryad because the metadata requirements in EZID are more lax (for example records that only have "ERC" metadata with who/what/when) and a lot of identifiers may not have a robust set of all the modern DataCite metadata (I think).
A first step of getting opensearch working is to create a json document with the search data.
I think first steps are:
For the future, we may want to think about some additional tasks that would enhance search, but I think it's best to get a working search first and then enhance as a separate stage.
Stuff like adding specific fields for many more values or displaying search facets/filters and other more modern search techniques. However, these things may come at a later time and it's not nearly as straightforward to implement in EZID as in an application like Dryad because the metadata requirements in EZID are more lax (for example records that only have "ERC" metadata with who/what/when) and a lot of identifiers may not have a robust set of all the modern DataCite metadata (I think).