Open Science with Kotlin: https://www.biorxiv.org/content/10.1101/2021.02.28.433265v1

The recording of todays seminar on Gdml, visionforge and jupyter-kotlin: .

@Iaroslav Postovalov just discovered a Kotlin builder for TeX by@khud: https://github.com/kotexio/KoTeX. We've currently started a more ambitious project code-named Snark (as in Lewis Carroll's poem), which is aimed to provide document transformation and generation API as well as builders. I think it would make sense to cooperate on that. @Alex Gladkov is also stuying builders and Android rendering. The snark's code is not public yet, I need to add a few things first.

@roman.belov @Maria Khalusova I wonder if there is a historic data about the number of users in each of this slack's channels. I have a fealing that there are several different "modes" for each kotlin direction. First the channel lies dormant with only a few dedicated users, but at some moment the direction accumulates anough interest to go 💥 exponentially.

Does Multik support a random ndarray generator?

message has been deleted

[JBR Seminar] A work seminar on Controls.kt inner implementation details. Alexander Nozik Description: Inner workings of Magix event loop. Usage of Kotlin SharedFlow as a loop implementation, integration with RSocket and ZMQ. Introduction to Controls.kt Sunday, July 18 · 17:00 – 18:00 Google Meet joining info Video call link: https://meet.google.com/zwe-rmow-wzh

Just found a coroutine-based demo for eclipse-milo by @kevinherron: https://github.com/digitalpetri/opc-ua-demo-server. Really nice to find kotlin in IoT and SCADA already.

My background isn’t really in any kind of science, data or otherwise, but rather SCADA / industrial automation / controls.

I work at Inductive Automation (a SCADA/HMI company) and I’m also the lead for the Eclipse Milo. If you have any questions about OPC UA stuff I can probably help 🙂

@kevinherron I decided to start with client. I am currently working slowly through your Java code and I already see why you took the idea of using kotlin so well. It could help a lot with code duplication and other things. My question is more about opc-ua stack. I currently can't see a way to encode key-value map. Also I can't see a way to interpret custom structure without knowing beforehand what it is. Is it possible to encode a map without using array of custom key-value structure?

A demo of first physics "simulation" visualization in VisionForge: https://www.loom.com/share/fde7375039ba4cb9889f7f31bbdca159 Here is the whole code:

val bouncingSphere = SolidGroup {
            sphere(5.0, "ball") {
                detail = 16
                color("red")
                val h = 100.0
                y = h
                GlobalScope.launch {
                    val g = 10.0
                    val dt = 0.1
                    var time = 0.0
                    var velocity = 0.0
                    while (isActive) {
                        delay(20)
                        time += dt
                        velocity -= g * dt
                        y = y.toDouble() + velocity * dt
                        if (y.toDouble() <= 2.5){
                            velocity = sqrt(2*g*h)
                        }
                    }
                }
            }

            box(200, 5, 200, name = "floor"){
                y = -2.5
            }
        }

@elect I want to move the discussion to a more relevant chat. I've implemented basic plotly map API and will deploy dev version shortly. It does not have the quality you want yet. For it we need to use openstreeetmap API. I will return to it tomorrow.

After a lot of development I finally released a first public beta of roboquant, an open source framework for algo-trading. Of course written in Kotlin, but also with great support for running it interactively on Jupyter Notebooks. You can see it in action here on Datalore It is still in beta, but any feedback it much appreciated. You can find out more about it at GitHub

During a seminar today, I've spontaneously made a nice pure kotlin demo for univariate distributions, point and range estimation: https://datalore.jetbrains.com/view/notebook/wr9C8wqmehowEVRWxtiS9Q. The code is not the best that could ever be, but it demonstrates some kotlin features. Feel free to use it as a demo.

Hello people 🙂 tomorrow at Casus (10am) we are gonna host a presentation followed by a short workshop with @holgerbrandl about Kotlin Automation and Scripting, he's known especially for at KotlinConf '17 and his kscript

Hello, for a side project where I have to find the right dimensions and shapes for physical parts, I started to write SI and geometry types in Kotlin as `value class`es, along with various functions for the various calculations I need to perform. FYI, "SI" means "Système International", which translates to "Internation System of Units", and includes all the science 7 base units like distance (in meters) and time (in seconds), and their derivatives like acceleration (m.s^-2), torque (I don't recall about that one)… It's not feature complete, I often find myself adding types, operators and functions, but I'm wondering if sharing such a library would still be useful. I also know it can never be really fully feature complete since the Kotlin type system isn't as flexible as the SI which allows infinte possibilities of derivative units. Has anyone made such a library yet? Do you have projects where it'd be useful to you?

A demonstration for a hypothesis test from yesterday lecture: https://datalore.jetbrains.com/view/notebook/KHxo8B64QfRV9FrNzSvPIA We are creating two 2D point distribution and checking if they belong to the same distribution via custom statistic. Alexander Svetlichnyi is my deputy. He was present during the lecture and proposed the statistic, so we named it after him. Probably it was invented before with a different name.

I am putting finishing touches on a production release of VisionForge. You can see statically generated demo here: http://npm.mipt.ru/demos/VisionForgeDemo.html (WARNING: 6 MB JS bundle size). The page itself is generated with VisionForge and the code is attached. This is a static demo, but it also could be dynamic (receive updates from backend) and it works in Jupyter! As soon as I finally write an article about it, I will move forward to doing a data-driven report generator. Right now it is already quite easy to do dashboards. Also adding new plugins is quite easy (added tables yesterday). The JS bundling is still a bit of issue, but I think it will be easier in future. Comments are welcome.

message has been deleted

Is somebody interested in Compose map component? I've done a prototype for several internal and external projects. It should be easy to port it to android. It is private right now, but if anybody is interested, I can open it.

Meta just open-sourced an automatic differentiation framework for Kotlin! (https://diffkt.org/). It can compute higher-order and symbolic derivatives, supports user-defined types, and has compile-time shape checking. Examples include machine learning and physical simulations.

When I build my app using

plotly-core

locally, gradle has no problem getting the jar. However, when I try to build it on jitpack.io, I get this:

Could not determine the dependencies of task ':kslides-examples:shadowJar'.
> Could not resolve all files for configuration ':kslides-examples:runtimeClasspath'.
   > Could not download plotlykt-core-jvm-0.5.2-dev-2.jar (space.kscience:plotlykt-core-jvm:0.5.2-dev-2)
      > Could not get resource '<https://repo.kotlin.link/space/kscience/plotlykt-core-jvm/0.5.2-dev-2/plotlykt-core-jvm-0.5.2-dev-2.jar>'.
         > Could not GET '<https://pkg-cdn.eu-1.jetbrains.space/1409141240651199418/maven/5577247194344223755/space/kscience/plotlykt-core-jvm/0.5.2-dev-2/plotlykt-core-jvm-0.5.2-dev-2.jar?v=1ntCif1ntCif&Expires=1657991009&Signature=cgVRG4BK9yp~fnK5lCbGlFHA8ib0MuuqwYYcEAwFnVOLDPsQ-f5c~zSQPd3X2S05mmqSI22-TYn-O-VTbN8X-a0LbLPcLTBRMls3oHZMFlxvPQuAzP1sQ9Ib~bvZX216ctUiTXJWeXdVDsPWaER-LaV~OeHhlyO2RXvSrG5KflgeM1sfyVqY~wJvTZZScGFus9UCE-48SqswH0v-grV-hCDZ61FGDPkTDljIQiDFxthO-f6y3by8X0wd9YlHMACb2iro~mQJCdfrgG4L~MxpngPnCkSW81rQBnQzJZJPW9cps6cK06qWDA6-54tFIWyyLHgzYhsNQIwW1oRRhMlyRg__&Key-Pair-Id=APKAIBRFJQGZXNBTFYIA>'.
            > peer not authenticated

The above uses:

-Dhttps.protocols=SSLv3,TLSv1,TLSv1.1,TLSv1.2,TLSv1.3

If do not specify a

https.protocols

value, I get:

Could not determine the dependencies of task ':kslides-examples:shadowJar'.
> Could not resolve all dependencies for configuration ':kslides-examples:runtimeClasspath'.
   > Could not resolve space.kscience:plotlykt-core:0.5.2-dev-2.
     Required by:
         project :kslides-examples
         project :kslides-examples > project :kslides-core
      > Could not resolve space.kscience:plotlykt-core:0.5.2-dev-2.
         > Could not get resource '<https://repo.kotlin.link/space/kscience/plotlykt-core/0.5.2-dev-2/plotlykt-core-0.5.2-dev-2.pom>'.
            > Could not GET '<https://repo.kotlin.link/space/kscience/plotlykt-core/0.5.2-dev-2/plotlykt-core-0.5.2-dev-2.pom>'.
               > The server does not support the client's requested TLS protocol versions: (TLSv1.2). You may need to configure the client to allow other protocols to be used. See: <https://docs.gradle.org/7.5/userguide/build_environment.html#gradle_system_properties>
                  > Received fatal alert: protocol_version

Any thoughts would be much appreciated.

Yesterday I've done a simple finite steps simulation of mixed linear model evolution for younger people from our lab chat: https://datalore.jetbrains.com/notebook/ptQDfQAcrjNxzIO0AEqovZ/wyWh7qFdXrP9OBWPIL7clW/. You can do population dynamics, stable processes or (like I've done) the decay of dictatorship regime (I won't add comments about that in the code). What I wanted to say is that it is already as easy to do in Kotlin as in Python.

Hello, Topic: Computer Science Q: What is the time complexity of the standard library filter function? Context: I wrote an algorithm for Quick Sort but for large data sets I'm getting Time Limit Exceeded. Algorithm

My github account was "suspended" because I work at MIPT, which is under US sanctions. All our important projects are alive, but I won't be able to maintain issues on the github. We will move to YouTrack instance soon for issues, but meanwhile write to me here.