Media is too big
VIEW IN TELEGRAM
almost done with the serial simulator, it uses posix PTYs to simulate our sensors. It handles commands and broadcasts telemetry continuously (all configurable). The behavior is defined in a TOML file. Added it to a test pipeline for one of the services I maintain and it works pretty well.
Used nix (not that nix) it's a crate of rust bindings for the posix api. This sim doesn't handle binary streams, but other than that it should be able to handle most scenarios. If anyone is building something that uses sensors then they're probably using some type of serial connection (RS-232 etc) and if they don't have the sensor (or it's expensive to keep it in the lab) they can just model the protocol in the toml and use this instead. I will share a sample toml but, any sensor in the world that uses serial and has a public spec can be easily modeled (i hope) you don't even need to buy it.
I'll strip out the device-specific/company-specific stuff and make it generic, then share the repo on GitHub.
Used nix (not that nix) it's a crate of rust bindings for the posix api. This sim doesn't handle binary streams, but other than that it should be able to handle most scenarios. If anyone is building something that uses sensors then they're probably using some type of serial connection (RS-232 etc) and if they don't have the sensor (or it's expensive to keep it in the lab) they can just model the protocol in the toml and use this instead. I will share a sample toml but, any sensor in the world that uses serial and has a public spec can be easily modeled (i hope) you don't even need to buy it.
I'll strip out the device-specific/company-specific stuff and make it generic, then share the repo on GitHub.
π₯9β€6
Oh, and added "replaying". You can basically take sensor output from a device that's in prod and replay it in your dev environment to reproduce certain issues and handle edge cases.
π4
This media is not supported in your browser
VIEW IN TELEGRAM
π₯5π2β€1
Stayed up most of the night, Opus 4.6 is good. Took me about 40-50k tokens for a rust rewrite of an implementation of the NIST CBEFF (Common Biometric Exchange Formats Framework) parser we used to use at NID (it was a Java library). Not too simple, not too hard. Opus had some minor inaccuracies that were fixed, but ran the tests against it and it did great. But it costs so much π
Original implementation: https://github.com/mosip/bio-utils
Original implementation: https://github.com/mosip/bio-utils
GitHub
GitHub - mosip/bio-utils
Contribute to mosip/bio-utils development by creating an account on GitHub.
π₯6β€1
Every generation is shaped by its cultural icons who reflect the eraβs aspirations through their work. Todayβs generation is a little different; not only do they aspire for greatness, but they also understand that art doesnβt flourish if artists canβt be discovered. Beza and Bandamlak just opened a cinema, mostly focusing on short indie films; Sehin and Abinet started Artawi (a gallery mostly focusing on photography); Nafkot (fuabilich) and Zion started Studio 11 (a gallery mostly focusing on paintings, but Iβve seen plenty of amazing photography shows).
We need to support artists, but most of all we need to support artists that create spaces for others to flourish. Years ago I had almost lost hope in Ethiopian cinema. In 2025 I regained all of it, and this weekend I can say that Iβm fully confident (and you can mark my words) that the next two years will be the best years EVER for the art form. I will come back to this post in 2028.
α«α΅αα°α!
We need to support artists, but most of all we need to support artists that create spaces for others to flourish. Years ago I had almost lost hope in Ethiopian cinema. In 2025 I regained all of it, and this weekend I can say that Iβm fully confident (and you can mark my words) that the next two years will be the best years EVER for the art form. I will come back to this post in 2028.
α«α΅αα°α!
β€βπ₯13β€8π1