Hummingbot Community, 29 Jun 2023

Previous Meeting

Next Meeting

⏯

youtube image

►

From YouTube: Hummingbot Community Call - v1.17, Dashboard Demo, Vote in Epoch 5 Polls, Kujira Community Demo

Description

Community Call Agenda:
• v1.17 Release Highlights
• Dashboard Demo
• Vote in Epoch 5 Polls
• Kujira Community Demo

Time Stamps:
00:00 Intro
01:03 Hummingbot Open Source Repos
2:53 v1.17 Release Highlights - Position Executor
4:21 Real Time Data Collection
5:10 Hummingbot Dashboard
5:32 Candles Feed
6:15 Market Order support
7:14 Demo of Dashboard
8:56 Vote for Epoch 5 Polls
17:32 Kujira Community Demo - Intro
21:25 Quick Start Guide
24:09 Setup Gateway
27:03 Import PMM Strategy example
28:15 Special Thanks
29:28 Q&As

A

So today's agenda uh we're gonna, go through the the release, highlights for uh version 1.17 of hummingbot, uh which just came out yesterday uh then we'll do a demo of the new help me, my dashboard, uh which incorporates uh the the real-time data collection feature uh that was introduced in 1.17, then we'll vote on the epoch, 5 polls, uh which dictate which uh sex connectors and which deck connectors and strategies how about supports every quarter, and this is a really important aspect of helping us allocate our scarce engineering bandwidth so that we can actually focus on improving hunbot.

A

Finally, we do have a new exchange connector that was just approved kujira. um This was built by a community team and they're also going to be maintaining the connector going forward. So Danilo who's been a long time. Member of community so he's in.

B

A

The kujira connector for the rest of us all right. Well, let's get started and uh talk about what's new in 1.17, first and actually before I kind of like go into all these release. Notes, I should actually mention that hummingbot is actually much more than just the single open source python repository now, and the reason is because you know today: hummingbot has grown into a collection of five different, uh open source, repos uh they're, each kind of working alongside each other, to allow people to customize exactly how they deploy the trading box.

A

For example, you know if you're, a new user running out for the first time, uh you'll, probably just use our deploy, examples, repo um and just run. You know the simple Docker compose up command, uh which will basically launch a single version hummingbot and may have a dashboard attached to it. That shows you. You know how that bot is running. On the other hand, a more sophisticated user like a trading firm, they might also use these a different deploy.

A

Example that will instantiate a more complicated configuration one that perhaps is launching 10 different instances, different dashboard, multi-bot performance analysis and some of those Bots might be using the Gateway module to connect to dexes. Others are talking to Sexes and the orchestration module is stitching everything together and allowing the dashboard to see uh overall performance in real time.

A

So obviously, that's you know two very different use cases, but because of how honeybun has kind of evolved for time, I think we're actually very capable of serving both and it's kind of why we're really investing into separating these five open source repos and having the community help us maintain so yeah. So I hope that just explanation wise that helps explain where we are today again. The simplest way to install hummingbot if you're, just new is go to the deploy examples. Repo and just run this Docker compose up command.

A

We've added a few Flags so that, if you're updating this will also work. But this should also work for a new user as well. So a couple things that actually new in this release, one big one is: we've significantly refactored the position. Executor component that helps you build directional strategies as an example, let me just show you the code. This is basically a script that creates long and short positions, Upon, A, Certain signal.

A

This is where you're defining the parameters for the strategy, which exchange you're trading on how much leverage they want to employ if you're doing perpetual and once you've created a position. What are the triple barrier parameters for that position? In other words, how long do you want to render for what is your take profit level? And what is your stop loss level decision? A secure class is a self-encapsulated piece of logic that basically just manages all that, for you, it'll automatically take profit, stop loss and, at the very end of the time limit.

A

If none are hit close, that's how you define how you want the position to be run, and then the only other really factor is really how you open and close Physicians, and so here the example is you're. Reading the RSI and then if RSI is greater than 70 you're going to sell, which is -1 and then, if RSI is, you know, less than 30 you're going to buy or return a one signal, and, and that's really up there- everything else is really handled uh by the abstraction in the class.

A

So great job fade a this really makes it easy for anyone who's doing directional strategies. The other big change was the addition of real-time data collection, so this has actually been a really popular request and I think this is going to increase uh the number of use cases for hummingbird that people will have out there. So the way this works is now there's three parameters in the global config file.

A

Market data collection enabled this turns it on and off the interval, which basically defines how you know how often you you collect as well as depth, how deep into the order book uh do you want to take a snapshot so basically, every interval, the bot will store this data in a new sqlite table. uh The Exchange trading pair timestamp prices, yeah and the older we've already Incorporated this into the new version of the dashboard. So you can actually run a bot and have a dashboard track your data coming from the bot in real time.

A

The dashboard is a new community project. It's probably the long-awaited GUI that everyone in the community has. You know wanted for a long time for more information, we're actually doing bi-weekly public Sprint planning calls for dashboard with a small group of community members come next week same time same place. Next Wednesday we're going to have the the dashboard kind of public Sprint planning meeting a few more new things for this release. Part of what we're doing right now is we're adding different uh features to the top exchanges that we support So for binance.

A

That's our Gold Exchange. We can that's where we added uh the candles, feed and all the other new things recently, and then we recently created bounties to allow community members to replicate those changes for our silver tier connect. Please announce that this release marks those editions of candle seed for gate, gate Perpetual as well as kucoin. Unfortunately, kucoin Perpetual does not support I think they don't have the historical K lines API right now, but when it does, we can also add candle Philosophy for Google and Perpetual as well.

A

In addition, we also added Market order, support for for gate, I. Think kucoin already. Has it and there's already bounties out there to add it for sendex, so this is kind of like how we really really do maintenance, where the connector voting system allows us to really move fast and improve the top exchanges, uh which is the gold one later on we'll create bounties to support the silver add those features of silver ones, and then the bronze ones are non-supported. Hopefully the community can can kind of copy those changes onto the bronze ones.

A

So we did see a new, a connector in this release as well. Pokedex is one of the top decentral exchanges in the polka dot ecosystem. It was built by coin Alpha, our original parent company. So thank you very much for the contribution. Finally, we also removed some connection strategies related to the last quarter's polls. We have many community members, who've contributed documentation, changes, uh bug, updates and fixes. So thank you, everyone, alright!

A

So let's, let's move on to the next phase of the call a demo of the new hold on dashboard I actually already have this running. Let me show you first, the bot just running a simple paper trading, two coin bot the code is simple: pmm example, except that I have lowered the spreads by a lot in order to generate lots of Trades.

A

Just a dummy example been running this bot for about probably half an hour now, so it's done 200 trades, but what's cool is that when I ran this I also am running the latest version of the dashboard. This dashboard is actually tracking the performance of this bot in in real time, and the reason you can do that is because we've added the capability in 1.17 to record. uh What's going on with the bot in real time, this is basically showing from the time that I started running the bot.

A

All the trades that my bot has done. I can also see how my inventory has changed, as well as a full list of a trade trades orders and.

B

Basically, everything in.

A

The the orders table- this is very much a beta version. This is why we're having you know, bi-weekly planning groups, to flesh out and to iterate on each of kind of the sample dashboards that we have here, but I think this is already, in my opinion, pretty useful when you're actually running a bot a lot of times, you're, really trying to figure out is like did I actually buy low, that I sell, especially since you can drill down onto specific time frames like this.

A

uh So let's say you ran the bot for you know a couple days and you want to see you know how how the bot's doing you know like maybe on a subset of that I. Think these kind of tools really help you so uh try running the dashboard. It's already open source out there, okay, so uh next up uh we're gonna vote, the epoch five polls um were just created uh every quarter. uh These are the ones that dictate kind of like how we maintain Helen bond. This overall there's been really a good change.

A

It's really allowed us to spend a lot less of our time, maintaining connectors, adding tests before we're testing every connector versus every strategy. uh Today we really focus on testing the gold and silver connectors and only running the core strategy on them. So so this is actually a really effective process, because this is kind of what what's allowed us to focus out more on things like dashboard. That I think help everyone. The first one I'm going to vote on. These are the proposed governance changes that we would like to propose for Epoch.

A

Five homiebot is very much a community code base we're sustained by the when everyone uses honeybot it generates fee share that kind of supports us. So from our standpoint, we were stewards of the the pool of hbot tokens as well as the hummingbot code base, so we think there's two changes that would really improve the system. The first one is. We want to try to be a bit more formal about requesting a certain amount of h-spot that will be allocated toward Dev Bounties in a certain Epoch So So.

A

Currently, uh we have 5.7 million in h-bots allocated for certain bounties that have already been assigned or submitted, and currently there's 3.7 million in the community rewards wallet that we have and there's another 300 plus million the reserve wad, which is controlled by the foundation multi-stig board of directors. So this time we believe that if we request a certain allocation, it'll be easier to track and easier to, you know allocate what that's for so we like to request 7 million for Epoch 5 for the next three quarters or Flex one for next three months.

A

This is basically topping up for the allocated bounties plus 5 million for any new bounties that may be allocated during the pop five. So these new bounties, uh what should they go to bug fixes or generally across the board, uh improvements to the gold and silver connectors uh improvements to the core strategies and also projects like dashboard? The second thing we want to propose is a new connector proposal, type So.

A

Currently, the way that countries get into the code base is that um they're they're submitted by a pull request proposal, just like any other pull request, which only requires one h-spot minimum balance to create.

A

However, every connector needs at least 100 000 ish Bond, every every quarter until the total votes to be to actually get into the code base because of the work entailed and actually reviewing connectors. We think that there should be a higher level of hpod required. Explain a bit more. We do a lot of testing on connectors when a connector hat adds things to the code base. uh It may cause issues for other connectors as an example.

A

We actually spend a bunch of time recently on the fox bit connector, because we're trying to upgrade hummingbot from python 3.8 to 3.10, which should make it faster and kind of fits the latest versions of anaconda, but because of certain tests- and you know things in the code um in the in a fox protector, which was something that so much in the community added. uh Now we have to fix that issue before we can make the upgrade.

A

So so that's why that's just one example of where um you know connectors actually do impose a cost uh that is realized by everyone using hummingbot and so having a higher threshold to add a connector in the code base. We think would make sense uh so yeah obviously I'm for this, for this one. For those of you who are new to voting uh it's one token, uh one vote done by a snapshot. Gas is uh it's gasless, let's go through the the other.

A

Three polls uh and I'll basically place my votes and explain my rationale for voting as well. This one is the sex connections poll, and so we divide contrast into two mainly because how many blocks always we want to support both sexes and dexes Sexes are actually one of the main ways to sustain us. So you know, vote for exchanges are actually paying us fee share, so that's typically related to who's, actually using hummingbot, where are people using bot uh and so essentially over the last couple years of wearing hung out.

A

Foundation we've realized that the users who are using Cloud the most and where they're running their Bots, those are the exchanges uh where you know where, where we're seeing the volume, so, um for example, this is also why we maintain this public dashboard. uh That shows you. The reported volumes from from the aggregate of Boss writing so binance and binance tends to be some of the top exchanges.

A

If people run the boss on um and then and then yeah, we have some small exchanges as well, but kucoin I think Gates uh are the other uh large ones um and not surprisingly, uh if you look at our fee share numbers which I just shared on the Discord thread: Finance it's index, gate and kucoin are our top exchanges in terms here, I'm just gonna vote for those, because, in my opinion, the fewer exchanges officially supports the the easier it is for us to to. Actually, you know serve our overall user base.

A

I would basically vote for binance uh ascendex uh yeah kucoin uh gate um who will be in okx because those exchanges where we have active feature agreements and we see uh revenues actually, okay uh for the Dex poll in the last quarter, uh uni swap was gold. Dydx, quick, Swap and pancake swap uh we're silver. I'm gonna try to use more data to drive my decision here. But to me, when I look at the reported volumes dashboard, um you know the the to me usage should be what was driving.

A

You know um my vote, uh because none of the decks really uh pay us, but we do add. Dex is because we feel like it's an important aspect of support in the community. The top decks by volume is actually dydx. Number two uh would actually be dexilot, but dexlot and and dydx have had have done two plus million in volume um across the board. This is also likely because dexlot has some D Minor campaigns running uh from our parent company coin Alpha um and then, and then, after that kind of going by volume.

A

um I would actually pick dydx I. Think in terms of other connectors that people are using a lot. I would say um it seems like deck slot they're using um and then a uniswell uh and I do see kind of pancake there. As.

B

Well, to me out.

A

To be honest, I think the deck side of the equation has always lagged the centralized side, because it's a lot harder to run bots on decks as well. It's more complicated, but, to me like I, would say the the increase in volume we're seeing on some of our order. Book Dex connectors is quite um interesting, and so that's why I think more of the older book uh connectors should be uh gold and silver. So, finally, the last one that we have is the core strategies, pull that dictates. Whichever strategies we uh apply when testing connectors.

A

This one I think is pretty simple. I would say, I think the Three core strides we have right now, which is am ARB uh cross exchange and pre-market, making um are the right ones uh and to be honest for us, like we've, actually been really pushing users to create scripts instead of strategies, uh because you know we found that those who customize and and kind of like you know really, you know tune their own strategies uh tend to have the most success.

A

However, I do think it's important to have a few base strategy templates in the repository, because it helps new new users. Understand um how to run home? Okay, those are the the three polls we used to do a prior to issues poll, but we were the only ones nominating issues. So if you want to become like a certain issue with a certain pull requests that you want us to fix, you can always reach out to us on Discord uh or, if there's a there's, also the pull request proposal method.

A

uh If you want a more formal way of getting your changes into the code base, uh but because of lack of kind of like Community participation in that poll, I think we'll just focus on these three polls on a quarterly basis going forward and not do the monthly priority issues pull anymore. But overall governance is really important aspect of doing this of maintaining home Bond. So please, uh if you have ishpot, please get the vote down. Thank you.

A

I think we're done with this part, we're ready to hand off the call to and bring on board the stage Danilo from the community who built your connector.

B

Hello, everyone, okay! So let's start our demonstration.

B

B

And we are we create this connector, so we were working this project for some time now and we are very glad to be presenting you, uh this new connector to the to the hummingbird community and also this possibility to the kujillo community about Market making with hamimot.

B

First of all, just talking just a little bit about myself and so I'm, a senior software engineer, I have 10 plus years of experience in the field, then I'm also CEO of the Fantastic lab. We are focused on blockchain, Smart contacts, machine learning, web 3, high frequency trading and the code we are working with. Having bought fire I think at least one year very deeply, and we also work with decentralizer autonomous organization aspects and Consulting. We also have our partner roboter.

B

Let's talk just a little bit about kojira, which are very vibrant and engaged Community. If you join, then if you are not dead, yet they have this platform here that you can create your orders, see the graphs and see the Phila door. This Market limit again they have this like a very beautiful dashboard here that we can swap tokens. You can like the cross bridging operations here as well from different chains. If you need they have like the Explorer of them, which we call it finder and they actually have more tools.

B

As you can see, this all can Finn and bow, and all of them is like the same standard, so very well, designed very beautiful and actually very powerful and for the people from the kojira community that doesn't know too much about hummingbot yet so we also have an awesome Community here in honeybot, it's very active, like you, you've seen from Mike. Here we vote for all of the things the community is engaged. You can create pull request proposals also government's proposals.

B

So it's like ready, Community engagement or these sources open source work with high frequency trading with hummingbot. We specialize in Market, making and Arbitrage strategies having about that building with a lot of his scripts and strategies ready to use a lot of them are so cool as and you would definitely need to check it out and we have Hemi bought in two different structures like two main repositories. We call it client and the Gateway.

B

The Gateway is the one responsible for like connecting with um the centralized exchanges or with the other, what we call it connectors. We, the gate, will be exposing this API but, of course, placing an order or cancel all the business kind of like sensitive thing. We need to use certificates for that and we have then the client that will be communicating with the Gateway in a very secure manner, using the certificates and the Gateway is mostly written in typescript and the client, mostly written in Python and inside of the client.

B

You can operate your Bot, you can import your strategies start and stop it and it's like having what is better tested. It's been used for a very long time with a lot of people, and then I would like to show you very briefly. Our quick start guide. We created this repository here, so we are going to send a message after our presentation, so you guys can like check this out as well, and this is like to make your life very easy I'm going to present now our demo.

B

The same first thing is client, create client.sh I'm, going to you, which call this customizer and just to show you those options just to show you the. How is our repository? It's very simple. You have a client which these scripts to help you to create the client, but actually you don't need it. If you already know hummingbot and you can create the client from other sources, and actually you will recommend that just one is just to make sure have you get started like from now.

B

If you want right and also this one, it's like to create the Gateway and the docker file for it as well, and definitely you can check the code, so you can get used to it and also see what it's doing and okay. So let me proceed uh it's. This is like the image that I would like to use, which handbag version I would like to use the latest one by default. We are using the shared folder.

B

We want to proceed okay, so in about some few seconds we might see the the screen yeah that it is. uh We then create a password now that it is, we are inside of the client, but our Gateway is still offline, so we still need to create our Gateway instead and here I'm going to use this default image, the Gateway is pose a part for us to communicate.

B

It's usually 15 888, so I'm going to choose default and now I need to pass like a certificate here, but I need to generate the certificates first, so I'm going to call here, Gateway generate search, I'm going to put the passphrase I'm, also choosing a simple one, and then, after that the client saved the certificates, and we ask you to save in this folder and, as you can see, we have the certificate here now and I need to push the same passphrase that I use there when generating the certificate, which is different from the password you enter inside of Hannibal and I should see that the Gateway will be up and running.

B

The Gated that was offline is now online. I can also test it with Gateway test connection, for example, and the next stop is, and we need to configure the wallet. The gojiro wallet, so let's do that Gateway connect, Fujita.

B

To choose the chain: we only have this one: okay, I would like to use testnet for the demonstration and, if I use the default none here, we are going to automatically choose the best RPC for you, but you can also like Define a specific RPC and for the the purpose of this demonstration. uh Yes, this guy is here from processary and also the people from Vegeta to help with this one so I'm going to use the specific one.

B

Okay, I'm, going to just wait a little bit. So the Gateway is like refreshing.

B

Perfect just wait now: the Gateway is now online again and now I can get the uh minimonic, which is the private key for kojira, so I will copy it here. If you don't know how to do that in kojira, we don't use metamask or phentral.

B

We usually use Kepler, but there are some other ones that you can use as well, and if you are here, for example, you can go to inside of your wallet and then you can go through the view, recovery phase and you can grab your mnemonic there, which is, will be your private one, private key and after that you can place and paste it here. As you can see, we can place it securely and in quadril it's possible to have multiple accounts, but it's not that common.

B

So if you are unsure you can just put 0 here as well and just wait a little bit. We are going okay. So now we were able from the minimonic and the account number to fetch here. The public key of ours and I can show you here, uh for example, from what I have in my wallet. I have uh 10 000 usk about 1000, kuji and 10 000 demo, so if I call hummingbird balance command, you know, I should be able to see the same thing.

B

So, let's see and as you can see, it works here and now the next step is for you to get started really quickly. We also included here in our quick start. We have here in conf strategies and we have this pmm uh threat, and so we are going to import this one pmm strategy example.

B

Now we can hit start to start Market making, but before that show you what we have here actually in the market, so I'm going to use this squeegee demo market for this demonstration and after he didn't start something should start happening yeah so yeah we're starting so actually initializing. We are canceling the orders there that are open. We are settling the funds. I am going to explain a bit later and after all of this retrieving information from the market, the balances and the other things.

B

And, let's see, oh, that it is so we can see that here in kojira we have Market making now I think this kind of ends our demonstration. But one thing that I would like to show is: when we stop, we are we try to gracefully stop. So please just be a little bit patient. We are canceling the orders, selling the market funds and withdrawing the funds to your main wallet here and and at the end we are going to have like in this end one. It's like finish it now. It's a special thanks.

B

uh Should they call you the community that's attending here. So thanks a lot to the Heming God Community, that's attending here as well thanks a lot too also I would like to thank the Dove and Hans from kojira, so we were working for the in this project for about 15 weeks, so it was a great journey.

B

uh A lot of good things happening a lot of difficulties as well, but we are now delivering this interesting product to you and also I'd like to think like Mike Fink, for inviting us Carlito for opening this space, one that is from the board of directors and Philly. That was also always supporting us to get everything like ready and done and might have Givens our partner from robot AI. So we, uh you probably are going to hear more about us guys in the future, because we might be coming with more strategies.

B

Scripts, also with some features for hummingbot as well, and maybe some of the new connectors too. Thank.

A

You very much Anila yeah, that's a great presentation. um Actually I did a couple questions actually and I. Think some people in the community had about kajira, so I think I asked a question about um about gas. uh I think someone said basically I asked like: does it cost gas when you cancel an order on kajura and apparently the answer is yes, so uh this is actually I would say.

A

My recommendation here would be actually to start looking at custom scripts or strategies that take into account this Factor so uh like, as in history here, we've actually integrated with a bunch of dexes that have this feature um going back as far back as radar relay uh you know a few years ago, and because we had to relay had this feature uh where you know it was like you had to pay gas to cancel, um it makes a lot of marketing strategies, not work, uh even ones like pure Market making.

A

You know when you're, when you're kind of replacing orders every 30 seconds, let's say um and if even if there's like a you know a small cost to cancel each one, it's going to make your strategy not viable. You know over over like a week a week-long period. You know it's imagine like paying one cent every 30 seconds. You know it's almost makes any viable.

A

Any marketing strategy does not work so so the way that we got around this with um kind of radar relay and some of the initial uh connectors was um a having a lot longer kind of like you know um like uh cancellation times. So it's only if the spread was like. You know like reach a certain point: do you cancel? You don't cancel all the time that was one and the other one was. uh There were specific fact like.

A

There was specific order, types that beta really offered that allow orders to expire, essentially without being canceled, and so we would basically, you know, like I, think one of the original strategies we had at hummingbot would take advantage of expirations.

A

um You know, and it only cancel when it was like completely economical, so I I think actually I would say. That's why I think um we want to encourage uh people like you who are built. These Dex connectors to essentially create strategies that are taking account the specific features of the decks, because I think that's actually what people need uh to.

A

Actually, you know reduce a barrier of entry and start Market making on the taxes, because currently I think that the user just takes a pure marketing strategy and runs on a conjura uh they're, probably gonna lose money, uh and obviously we know none of us want that.

B

Yes, definitely I think this is a very good feedback, so I cannot really say if we rent for a very long time. I know that the kuji refuse are very low, but anyway I think it's important to check the impact for let's say that you keep your Bot running for one month or something like it. How much is going to cost in the end, yeah I.

A

Don't have this.

B

Information but I can check it.

A

Out so I would say uh yeah. If you talk to any professional Market maker, you tell them they're going to be canceling their incur costs to cancel orders. um You know just yeah. You should see how they respond, uh because this is actually honestly the main reason why greater relay never took off because they couldn't attract uh real market makers that actually want to pay. You know fees to cancel. So that's actually why you look at uidx today, uh dydx uh they.

A

They basically absorb the fees themselves, because they're running their own app chain on Cosmos or something like that, uh but I think it's uh yeah. It's one factor uh that that I think yeah is orbug Texas. You should really look at.

B

A

Think the session opportunity as well, because you know it's like because the off-the-shelf strategies, don't work uh I- think there's kind of like you know it's kind of like so this is when I think a customization uh and having like kind of you, know something that actually allows because, like.

B

Because, like I, think exchanges to.

A

Me are always like it's like they're, like three-sided marketplaces. You know you need uh the Traders, you need the market makers and then you also need, like the token projects who want to. You know like create a new token on and put it on your exchange, uh so I think just yeah make anything that makes it easier for especially the market makers to to come and run a bot and not not get killed, and it's gonna be helpful. Well, thanks!

A

Everyone, if you're here next Wednesday same time, uh we're doing the dashboard public Sprint planning meeting uh and then we'll have the you know another. uh You know how to call another Wednesday after that so uh see you.

B

A

And uh you know, keep on running box thanks.