Update on Typesense Feature Development and Priorities

Hi Typesense team, just wanted to know what is the progress regarding https://github.com/typesense/typesense/issues/227 , I am really looking forward for this feature, when do you guys think it can be done realistically?

We prioritized it but then had to put it on the backburner because it requires some intricate changes and we're already making other intricate changes for the other features we are working on now. We don't want to make too many core changes at one time. So it's not on the top-3 things that we have planned for. For now, you have to flatten the document before indexing and unflatten it after fetching it.

If you don't mind me asking, what are the top 3 features you guys are prioritizing right now?

What we have prioritized for 0.22:

a) Performance optimisations that speed up updates and reduce memory usage. This involved a lot of refactor to core datastructures, and is almost done.
b) Overrides using dynamic filtering: if someone searches for e.g. “white dress”, automatically apply white as a color filter then do a text search only for “dress”. About 80% done.
c) Improving multi-field search relevancy. We have fixed almost 4-5 edge cases here, which covers last-mile relevancy issues that some customers have noticed.

Apart from these, we have fixed several other smaller bugs and feature requests (e.g. https://github.com/typesense/typesense/issues/273).

Once 0.22.0 is out, we plan to focus on:

a) Schema updates
b) Phrase matching where you can query for an exact match by wrapping a phrase in quotes, e.g. "foo bar"
c) Other smaller bug fixes and feature requests.

Support for nested documents will probably be scheduled for the release after that.