Sophie NEWS 📢
Hello! We're currently testing @betaSophieBot, which boasts a brand new architecture. For the technical aspect, there have been significant changes that will be elaborated on and made open-source at a later stage. We would greatly appreciate it if you could…
I want to share some inside information about the current state of Sophie's development.
Sophie's development has taken a different direction. It's not exactly a replacement for Sophie anymore, even though that was the initial plan. As I worked on the code, I realized that it's not possible to replace Sophie in one go.
Instead, I came up with an idea of that both bots works alongside. So I came up with ProjectQ. It acts like a proxy between the two bots. We have a 'master' bot and a 'slave' bot. Every update we receive from Telegram goes to the 'master' bot, which then decides whether to pass it on to the 'slave' bot. The 'master' bot is called otty (its codename), and the 'slave' bot is Sophie.
Another big change is the change of the database. We're now using FerretDB for Sophie. It's a project that uses Postgres but works like MongoDB, so our existing code still works.
In the future, my plan is to gradually replace Sophie's features in otty while migrating all the data, without any extra effort or even the users noticing.
Sophie's development has taken a different direction. It's not exactly a replacement for Sophie anymore, even though that was the initial plan. As I worked on the code, I realized that it's not possible to replace Sophie in one go.
Instead, I came up with an idea of that both bots works alongside. So I came up with ProjectQ. It acts like a proxy between the two bots. We have a 'master' bot and a 'slave' bot. Every update we receive from Telegram goes to the 'master' bot, which then decides whether to pass it on to the 'slave' bot. The 'master' bot is called otty (its codename), and the 'slave' bot is Sophie.
Another big change is the change of the database. We're now using FerretDB for Sophie. It's a project that uses Postgres but works like MongoDB, so our existing code still works.
In the future, my plan is to gradually replace Sophie's features in otty while migrating all the data, without any extra effort or even the users noticing.
❤1⚡1👎1🔥1🤡1
I would like to hear your thought whatever they would be about improvements or shitpost
In the past few days, I tried to import the current Sophie database into the Project Q architecture using Sophie Beta, but unfortunately, it didn't work.
The problem is with Ferretdb, which is not optimized. It selects the entire collection to filter the data, which means that no matter how much RAM I have, it won't be enough with Sophie's large database.
The only solution is to switch back to using MongoDB as our database.
So now, we have three databases that need to be maintained: Postgres, MongoDB, and Redis.
Anyway, I'm happy to say that my tests were successful, and we are now ready to migrate to a stable version of Sophie.
The problem is with Ferretdb, which is not optimized. It selects the entire collection to filter the data, which means that no matter how much RAM I have, it won't be enough with Sophie's large database.
The only solution is to switch back to using MongoDB as our database.
So now, we have three databases that need to be maintained: Postgres, MongoDB, and Redis.
Anyway, I'm happy to say that my tests were successful, and we are now ready to migrate to a stable version of Sophie.
👎1💩1
The question is now how they can achieve Red Hat sourced dstro to update every 15 minutes?