Hey guys, do you know what the release schedule is like and when is 0.21 expected? I’m busy writing a typesense search driver for Statamic CMS and wondering how long it might take to close a feature require like #252… Statamic is very content heavy with lots of markdown files so this would need to added before the driver could be used in production. Also keeping an eye on #266 which will make things a lot easier and expected to be fixed in 0.21.

We usually release the next version when we feel that the current RC is stable and has enough useful features to make a meaningful release. We do make stable RC builds available for customers who want a feature earlier but that's not yet part of an official release build.

#266

is fixed except for the observations in the last comment on the issue. We currently treat an optional field with

null

value as if it was not part of the document at all, so there is no way to query them or sort on them, They will they be part of the document in the search response for the same reason. As for

#252

I will be happy to add that feature in the next couple of weeks.

Great, yeah regarding #266 I found that in my driver code I have to filter each document and remove the null values before uploading them otherwise I get an error… so being able to upload null values and have typesense ignore them (on optional fields) would mean I can skip that sanitisation step in my own code. I’m only expecting to release my code in the beginning of August so a few weeks time for #252 is no problem ✨ and I’ll put the driver on the marketplace so other people can use it and help maintain it.

Cool, I will be sure to update that GH issue thread when a RC build with that feature is available.

One more question I had regarding the driver, Statamic uses UUIDs instead of integers, so a document would look like this:

---
title: 'Follow the leader'
type: blog
updated_at: 1578924579
id: 4b1aa095-e0a1-4138-92e0-0dfc3e81697b
---

In the driver when I create the schema I just add my own

id

field as a string, and I assume this overrides the default int ID that typesense uses? Because when I delete an entry on typesense I have to use the

id

field… but it works when I use my string UUID too. Just wondering if this might cause any problems or performance issues? Or if I need to find a way to create an int id for each entry.

I assume this overrides the default int ID that typesense uses?

Yes, that's correct and perfectly valid. Even though Typesense generates auto incrementing ID when an

id

is not present, that ID is also a stringified number.

Hey @Kishore Nallan I tried this today and it seems like it’s working great! I was able to get the highlighted content field only. ✨ So just to double check… even though I use

exclude_fields=content

and

highlight_fields=content

together in the request — it will still search and rank the results the same (with the content field) but will only exclude the field in the returned response. Using

exclude_fields

doesn’t affect the search engine or ranking?

Awesome. And, yes

exclude_fields

and

highlight_fields

are both applied only in the final presentation layer.