Fluence Labs Community, 10 Mar 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Developer Community Call #11

Description

00:00 Intro
00:48 Latest updates by Dmitry Kurinsky
04:49 Work in progress project presentation by Joera Mulders. Aragon.IPFS.Fluence. Scripts
20:15 Q&A
23:23 Hackathons info, bounties

Useful Links
https://fluence.network/
https://twitter.com/fluence_project
https://doc.fluence.dev/docs/
https://doc.fluence.dev/aqua-book/
https://github.com/fluencelabs/aqua
https://github.com/fluencelabs/aquavm
https://fluencenetwork.notion.site/fluencenetwork/Fluence-Developer-Community-Calls-2733beddb3ff4e4986c4fcc830fc796d
https://github.com/fluencelabs/marine-rs-sdk
https://t.me/fluencedev
https://fluence.chat

A

All right: well, let's get started, welcome everybody for another affluence, community uh edition and march so time, time's flying by and today.

A

And today I think we well, as usual, we start off with michael with some updates go through the team. I think juera has a presentation for us on his progress and I don't know dmitry. Do you have something at the end?

A

Nothing all right, so it'll be a fast meeting. Anybody has questions. We are ready to answer. Well, dmitry, it's all yours, so.

B

During the last, during the last two weeks, uh we were working on different stuff, but delivered mainly some things around aqua, and I would point your attention on two particular features. The first one is what we call recursive streams.

B

It was delivered to arcway vm, and then we found some bugs and we fixed them, so they are finally delivered and recursive streams is one of the holy grails of aqua. Actually because it it means that you can have a recursive or iterative distributed, algorithms coded in aqua, for example, uh cademlia or a gossip, or uh back pressure uh with pushpull model, or something like that.

B

So that's that's quite nice and we've finally updated the network uh to do the latest version with the other bugs fixed today and going to implement, to use it and uh implement new distributed algorithms in a aqua library during the next I don't know weeks.

B

The second update is the collection creation syntax in aqua. uh We should be very useful. Actually it's just a sugar uh but uh should be very convenient.

B

uh Now you can create a race, uh optional values and streams uh with the syntax sugar in aqua and introducing this syntax is a part of the consequences of the huge refactoring that we had for uh for aqua in january and beginning of february, and more changes are about to to come, but let's announce them next next time.

B

So this is what we delivered more updates to come into weeks. I hope- and uh I'm good.

A

All right thanks, be three questions: anybody recursive streams, back pressure, async mailboxes.

A

ah That's sad all right. Well, then, on to anybody else from the team wants to uh present an update, give an update.

A

One of the many mics.

A

I might I might just randomly click on one and put him on the spot, see what happens.

A

All right, well, george, mike it's all yours.

C

um Okay, can you hear me very well also mike okay, that's interesting. um Can I do um share screen? Okay,.

C

So it's mostly people from the team uh right there now right, so l with all those mics, oh yeah, we don't know who the mics are. It's true. um Are you seeing a screen.

C

Okay, so um um I've been working with fluence now for a couple of months started off with the hackathon.

C

um What I really like is that you know for like any web 3 thing, you're thinking of blockchains and ipfs, and then you have these computation networks like fluence, which can like connect everything together, and you can actually start programming and why I, like specifically fluence, is because it's like sort of an opinionated and I can find my own way in it. um So what I'm trying to do is that um I like to uh aggregate data um in a decentralized manner, so, for example, from.

C

Votes from daois, that's the actual example I'm working on so here we have an aragon.

C

Interface and I have a vote, and I can if, when I'm connected, no I'm not, um I can change my vote or create a new vote and um the code on fluence will listening to it and when it records like a new vote or a new vote, a vote on a vote, um it will create a new data structure in on ipfs, which looks like this.

C

So it's like this ipld sort of tree, and it starts off with the that. This is a data model for governance, because we're talking about these dowels and then um this pair dow example. Two, that's like the one dow that subscribes to it. There's only one dollar yet uh subscribe to it. It's got votes, and here there are like uh 23 votes recorded, and then there is like the different um data, for example. Here you can see like it should have recorded.

C

Oh that's in a file format, but it should not be. If I do this one, oh yeah, so here you can see that the data has been uh stored on the ipfs um and what it also does is that the fluence is now creating a widget, which is an example uh here.

C

So uh this widget is like created uh with fluence, so the data comes in and then it renders html and it makes the folder with ipfs um so that, if you go to this um address, you can actually see the widget.

C

And so every time you take a vote, the widget will update.

C

That's that's about how it works.

C

How am I doing this with the fluence now um so I have um two scheduled scripts running like this. One uh is on uh kasundar the one and basically it needs to create a filter for ethereum blockchain, and it needs to do that.

C

Every 50 minutes uh because, like normal, like a um sort of ether, client like eaters or what's the other one called web3, they connect, I think, through something like a web socket, so they can have like a continuous sort of filter connection, but because we're confined to working with the curl, it needs to cre, recreate one every 50 minutes, uh and that's so one script is doing that and um within the let's see.

C

So that is this one? What we're looking at so ah that I forgot to say anyway, so this one um is creating a new filter and then it's storing it on the dht of the nodes that want to cooperate.

C

In the future, through one of these registry mechanisms that we heard about yes last week, two weeks ago,.

C

um So yeah, so this this one, this script is basically running after 50 minutes. It creates on basic of the subscriptions, which is now only one dow.

C

It creates a new filter and that filter is then stored on the node of on another node and um on the that other node is not a script, is running, it's a large one, and that's this one and it again starts looping through all descriptions that are there and then it goes to look up the filter id and then it filters and then it starts doing all things to create the widget and the data structure.

C

um Yeah. Can anybody make a sound to confirm that? I'm still here.

A

You're still here.

C

Okay, good, um uh do you have any um questions I can I can go on, um but maybe there are questions.

A

Yeah I got one, did you say you, like reinstalling a filter? Every time you run the script.

C

um The script is running like every 15 minutes, and then it just it creates uh um it. It makes a call to the ethereum blockchain and that returns like an id of a filter and with that.

A

Right right, no, no, I understand that I was just asking so you're reinstalling it every 15 minutes.

C

um Okay, so the part is that what I'm, what I'm doing with all my scripts, is that I'm creating all the surfaces every time it runs. Okay, gotcha, all right and- and I'm not really sure if that is like the the way to go in the future. But my my thinking was that if you want to have something that is, um cannot be tampered with is like truly unstoppable.

C

You cannot like really confine yourself to uh existing scripts that are tied to a specific note, because then you would just have no somebody could like okay, it's that note that it's like creating this website with this terrible misinformation or whatever, let's take, and then you there could be a request to take down that specific note.

C

um So I'm thinking of in a ways of how it can be really like um portable um but yeah, I'm I'm it's so early for me to start figuring out how this works, that it can be really stupid as well. I don't know anyway, but I'm creating all the surfaces and then at the end, I'm removing them, which I thought to be very clean. So what happens with the bigger one?

C

uh Sometimes it starts running in problems, so I had this one running for a couple of days and then suddenly there are like 400 services of each. So at a certain point together, there was like on one note: there were like maybe 2, 000 services running um so then I had to stop the script and clean it up. So there was something happening and it's very hard for me to debug how that's happening, what what sort of caused that issue and also the.

A

What would you need to be able to debug better.

C

um Well, I I was running it on my own notes, but then I could. I only have like the server power to run one node right so that wasn't helping.

C

And maybe the logs uh oh yeah, I was thinking that I was going to write another script. That is basically counting the number of services and and then connect that to like a whatsapp thing or sms thing, and then it would like alert me whenever the actual mistake starts occur. And then I can call you and we can figure out what's going wrong.

C

Gotcha um well with requests. um What I would really like is um yeah.

C

I've said it before that um that if you run a scheduled script, that you can put arguments inside and get things out because then also, I would be able to cut up this bigger one in smaller ones, because you see that, um basically, when, um when I pull successfully for an event, then it starts doing, uh it starts like using the surfaces that are more heavier and actually do stuff, but that could be inside another script so that the one script calls another. So the the recurring script calls um a script that runs only once.

C

But it needs the information from the first script to be able to perform its tasks.

C

And then it could be more economic with creation of services and yeah. Maybe this is just too bloated.

C

And oh yeah, so, and maybe that if, if there is like this script, it has like a sort of pre and post function like uh so that, for example, this the the removal of the surfaces that you can put it in a post function so that even when the script sort of fails it.

C

It would then still do that.

C

um Oh yeah, and what can I else? Can I say: um oh yeah so- and I have this this- um this is the the other one that's running on on custom.2 now uh or the scheduled script. That's actually doing the things and um I I'm logging from the out of the services, basically all the time to this console.

C

So then I can just see what I'm doing um and it's not working at the moment, and so otherwise I would be demoing it anyway. So here just uh about an hour ago, I had one result and then um um it started doing things and then it ran into a bug.

C

um Oh yeah, and so one other cool thing is that um because I'm recreating all these services every time I'm running the script and the code for those services comes from ipfs.

C

C

I'm working on a a surface, for example this one so you're familiar with this like built a shell script. I think I stole it from from you, but then, on the end, um I'm um I'm uh uploading the uh uh the output to ipfs and the cid I get back to ipfs I store on uh on the blockchain on ethereum somewhere, and um so every time I'm creating a surface, I'm going to the blockchain to check okay.

C

What is the latest cid for the latest version of the of the service and it installs it from there? So I can have like a scheduled script running and then just change the service. While the script remains running.

C

And so this week I've been doing this. I was making this form, so you can, um if you're a dow, and you want to subscribe to the surface it. um It just has a form which asks for the uh uh the information needed and then it it stores the storage it's on ethereum and uh ipfs, and then the scheduled script is like gathering that information to run scripts.

C

Well, that's about it.

A

But I may have missed something here when you said the the the forum gets committed to the blockchain and ipfs, and then the script gets it from there, meaning it's pulling it from the blockchain. The cid.

C

A

From through the filter.

C

um So so every every time.

A

So the registration calls a log emit event and then you have a filter for that and that's how the script catches it right.

C

um So so this one every time the script is running: yeah, okay, one that is creating a filter. It goes to a surface and that surface gets all the subscriptions and their metadata from the blockchain.

C

So yeah, so even the script can continue running and there can be more subscriptions added or they can expire or you know whatever.

A

Gotcha, thank you.

B

Oh, how have you implemented the right, tht function.

C

Are there something strange with that any I I started doing that only a few days ago, because after the uh you talk about the registry, I thought. Okay, I have to start looking into this stuff and then I thought I start with well. The registry wasn't really fit for this, so I I tried this one here it is, can you can you see this on a.

B

Yes, yes, I can see it.

C

B

A

I think he froze his wife. I finally gave out.

A

It's just getting interesting.

B

Well, uh as far as I understand it's awkward, hd, not registry, and not the library but the low level api and I'm curious. Why? Why it's so and would be awesome to maybe to have a call and discuss?

B

How can we switch to to the red stream.

D

A

All right, it's probably gonna reconnect but in the meantime, key up your questions for jurera and uh what else anna do you have any updates?

A

Hackathons upcoming yeah.

E

Actually we have, um we have two hackathons running at the moment, so there is one with the swarm called. We are millions and the bounties, you can still apply and we have two bounties from fluence and there are many more from swarm, so just check it out it's in git coin. If you go to hackathons and they're, um the hackathon called one and you will see the bounties there, so you can apply and try to uh to have that to to solve those uh challenges.

E

And then the uh other one uh that is running at the moment and you can still apply- is east denver. That's a virtual e denver hackathon. There are lots of sponsors and we at fluence. We have seven and a half thousand um usdc in bounties.

E

So please check it out I'll, send the link in the channels and also here in zoom, so yeah. That's um that's the news and we're also getting ready for um 8th amsterdam. We will not be participating in amsterdam like supporting uh like we're supporting, but we are not sponsoring the hackathon itself uh because it's just too short for uh like for fluence just two days of offline hackathon, but uh we want to organize side events, and uh hopefully yuka can help us with that, because he is local and yeah.

E

Let's see, what's uh what's gonna happen it's um end of april. If anybody is planning to be there.

A

Very cool, so people sign up and hack. If you want to make some money, it's there.

A

Are you back umera.

A

I thought I saw a new mic sign up, but uh maybe not all right. uh Anybody else.

A

Well then, three two one all right form close. Well thanks. Everybody thanks for signing in thanks for hanging out and um where, even though you're not here, thanks for this great presentation we'll follow up on the dht, we really want to talk to you and see everybody in two weeks. Thank you.

A

Thank you. Bye.

D

B

C

Thank you. Alright,.

D