ad hoc
ad hoc
Not possible, this is tied to the representation on the documents in milli @Kerollmops
TBH it should not matter. JSON are manipulated programatically, and the order is irrelevant. We make assumptions on the order of the fields at indexing, but the [JSON spec](https://www.json.org/json-en.html) specifies...
@qdequele the fieldId doubles as field position, that's why
@dichotommy nope
Like what we did with facets (which are lowercased), we will need to store the synonyms in two different places. On one side the one we currently store need to...
In the end, this is not possible to do it in a straighforward manner without impacting user that use non-latin scripts. This involves work done on the tokenizer. In the...
Still an open issue @qdequele
I have explored multiple potential optimizations, but they have not been fruitful, I'll list them here: - I had noticed that we had a majority of 8bytes writes, so I...
I think one further step from what @k0nserv suggests would be to have an interface like so: ```rust trait TimeGen { type Future: Future + Unpin; // other bounds may...
funny, I spent 3 hours debugging an error caused by just that, where a serializing error occured in a `SerdeJson`...