- Add support for adding and updating child objects specified as abstract types in the schema.
- Split indexing into 2 phases.
0.27
- Move embed cache to connection and increase size.
- Fix reporting upgrade errors when error path includes an integer.
0.26
- Indexer: Limit workers to 1 task and scale chunk size based on number of items being indexed.
0.25
- Indexer: Limit workers to 4 tasks to avoid out-of-memory errors.
0.24
- If the schema specifies an explicit `mapping`, use it when building the elasticsearch mapping. This provides an escape valve for edge cases (such as not indexing the layout structure of a page).