What Elixir related stuff are you doing?

just wrote a

fly.io/phoenix-files/streaming-uploads-with-liveview/

Followed this blog. and implemented S3writer. Been in need for that for along time!

gist: remote_upload.ex · GitHub

1 Like

5 Likes

Nice job, any chance you want to open source your project?
I’ll be soon installing a grid at home and, although I won’t be using the same inverter as you, this could serve as inspiration.

Cheers

1 Like

Sure, ill clean it up and post. :innocent:

1 Like

I have been engaged in the development of rocketmq-related technologies recently, mainly using the Go language. However, the quality of the Go language SDK for rocketmq is very poor, so I am currently rewriting an SDK using Elixir. At present, I have already completed the producer part.

1 Like

You might want to take a look at this here: GitHub - benvp/toniefy: 🎵 Listen Spotify® on your Toniebox.

Edit: hm, that’s exactly what I wrote a few years ago: What Elixir related stuff are you doing? - #573 by benlime

Wrote a quick demo with Phoenix LiveView & PubSub for my buddy, who is mostly C# and Java dev. Good enough to show the basics, I think.
https://github.com/holobeat/broadcast-demo

I’m beginner here. Currently just learning interactively by using elixir koans but i expect i could create my personal blog using elixir with phoenix.

4 Likes

I’m building a pure Elixir syslog logging solution. I say pure because most of the existing solutions just use a NIF or a port driver to bring in the native API.

My approach is to introduce a socket server which implements the Erlang IO protocol and use it as a device with logger_std_h. It supports UDS, UDP, and TCP transports.

I’ve also implemented a higher-order formatter, wrapping other formatters so you can serialize the message part of the larger syslog message the way you want.

2 Likes

Building https://researchmate.ai with LiveView.

It’s a tool that will do a web research and write a comprehensive, up-to-date article with references to used sources. Try it out, it’s just released in beta and free!

Using LiveView for streaming AI responses and building interactive interfaces feels like cheating :smiling_face:

3 Likes

I’ve been building a personal site for saving and sharing travel ideas.

Gen 1: Phoenix + JSON + ReScript + React
Gen 2: Phoenix + Absinthe + ReScript + React
Current: Phoenix + Absinthe + LiveView (Absinthe still supporting mobile app)

5 Likes

At this time, I’m looking at leveraging cloud-native tooling (i.e. Docker, Kubernetes, and so on) with the Elixir Programming Language.

Im creating a POC with elixir for people detection using:

  • membrane
  • kvs
  • sagemaker

I’ve been working on Ex4j a library to abstract Neo4j access:
https://hexdocs.pm/ex4j/Ex4j.html

1 Like

I’m still polishing my Elixir skills, but not working on anything particularly interesting… Just hanging out on the forums and trying not to give out too much bad advice!

4 Likes

Well I’m still learning, Looking forward to work on elixir projects. I built an Elixir + Python project. This one talks about getting data from Jenkins builds and displaying a dashboard. I used Python to run few ML code to train on estimated and execution duration of the jobs and predict anomalies (simple stuff). Used highcharts for the graph.

Project: GitHub - ltenzil/build_anomatron at master

I’m currently migrating a Ruby, Python, and Javascript app to Phoenix LiveView.

3 Likes

I thought that the modern font format is vastly under-explored. Modern OpenType specs support colored vector graphics, is programmable / quasi-Turing complete, is in principle application-agnostic, cross-platform, and have decades-long longevity. For the past two years, I’ve been working on these areas, most of the energy focussed on a font that provides colored, context-sensitive pronunciations for Cantonese.

v1 of the font was released May 2023 (www.visual-fonts.com), and made its way to HN front-page: Cantonese Font with Pronunciation | Hacker News

v2 of the font had been successfully built and will be released in May 2024. It is vastly more ambitious than v1: it includes every reading of every readable character in every known contexts.

Early testers have used it for teaching materials,

sing along to Canto pop:

and read books on Kindles.


Backing this is a thorough, state-of-the-art understanding of the language, and mechanisms for data to drive font production. Elixir is used exclusively for both. Right now there is nothing very specific to Elixir that (in theory) could not have been done in other languages (I wrote no GenServers outside of porting out), but the immutability (thus functional) approach was exactly the discipline I needed.

I also loved the progression from IEx → Livebook → Mix projects; it gives a smooth progression from write one-time code, that gets re-factored and eventually packaged into a dependency. I think Livebook apps are vastly understated: I had once hacked together, during lunch hour, a prototype for a speech therapist application, that does real-time “chat” where the client do voice-recording + ASR for Cantonese using Whisper-v3 to generate a transcript. And I consider myself a n00b hobbyist.

v2 of the font will be accompanied by some web-apps (for legacy systems that don’t support one or more aspects of modern font rendering). Those will run off the same libraries, and have Phoenix / LiveView for an interactive experience (that requires minimum JS from me).

2 Likes