Fluence Labs Fluence workshop: deploying services, 22 Mar 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Fluence workshop: developing services

Description

00:00 Intro
01:34 Setting up environment
02:58 Setting up Fluence tools
09:46 Run local Fluence node
12:10 Check the local node work
17:03 Working with services using AIR. Deploy a service
25:48 Call service API
32:00 Compose services
36:52 Useful resources

To follow this workshop: https://www.notion.so/Fluence-workshop-developing-services-b1bfbf9f76d94135b56394e83fc0a8e9

Previous workshop: https://medium.com/encode-club/hack-the-system-fluence-workshop-1-video-documentation-d0a2052e8f7e

Fluence challenges https://medium.com/encode-club/hack-the-system-announcing-fluence-87fa8dae59?source=friends_link&sk=18b7d34bdf4436d4193f69090aae9157
Get started with Fluence https://fluence.dev/docs
Protocol Paper https://github.com/fluencelabs/rfcs/blob/main/0-overview.md
Developer Hub https://dash.fluence.dev/
Github https://github.com/fluencelabs
Youtube channel https://www.youtube.com/channel/UC3b5eFyKRFlEMwSJ1BTjpbw
Fluence Home https://fluence.network/

Original video: https://www.youtube.com/watch?v=MgdVahMsvpQ&feature=emb_imp_woyt

A

Hello, can you please allow me to share the screen.

B

A

B

B

Should awesome.

A

Okay, great, thank you again for introducing fluence. So what's uh gonna happen today, uh on the previous workshop, I went in a really really details of how to develop services, how to deploy them to the network.

A

Today, I'm gonna focus on how to develop services locally, how to run your own node to connect to the network. So you can see the logs, so you can see the output.

A

You could be missing from the other tools, other means and that's basically it it's going to be pretty sure, but I I hope you will enjoy and again for more detailed explanation.

B

A

How to develop services? You can refer to the previous workshop, it's available down here below in the links.

A

Let me scroll here the recording of it and the scenario here in the notion too, so you can follow okay, scrolling up so we'll start with uh setting up the environment you I hope you already have npm or yarn docker and git installed.

A

um Thus I would not cover uh what I will cover is setting up the rust and cargo and stuff and then setting up fluence tools. So, let's start with setting up the rust, it's pretty easy. You just have to copy this command and paste into your terminal, and then it will ask you and you will just press one and it will do in a few minutes that will work that will install you rust, nightly tool chain if you've ability to compile rust to webassembly right away so yeah after installation is done.

A

uh You would need to either reopen your terminal or or execute defaulting command on your terminal, and you can check then that the cargo is available by well executing cargo minus minus version, so I hope you're starting doing that. In the background um and in another tab, we can begin set down the fluence tools.

A

First, one is flgs: it's a tool to interact with the flint's network, to deploy services, to inspect them to execute air scripts, and this is your go-to uh cli tool to work with the fluence network that you should use to manage your services and inspect them and stuff. So installation is pretty simple. You just need to call npm install install globally on this package. So you just the same.

A

You paste it to your terminal uh hit enter and it will be installed, and then you can check that it has been installed via full disk minus manager, minus minus help. You will see the different options.

B

A

Let me make it bigger: it's really a lot of them here, so you can do a lot of things with the network. You can upload modules, you can inspect modules on the net on the nodes of the network. You can get interfaces of your services, you can create services and that's basically- and you can uh of course execute scripts with this.

A

I will show you how to do that, but first we need it installed. Okay,.

A

So another two tools we need to have installed to follow. The workshop are fcle and frapple.

A

This will allow you to build web assembly in the correct way, with correct types and different metadata built in into webassembly modules, so you can deploy it to the fluence network. Also, you will need a frapple.

A

This is you will need for this current workshop, but overall, it's a useful tool, so you can develop your services locally without deploying them anywhere it's as easy as uh previous one. You just copy paste it in your terminal and it it installs. You can check then that it's installed by printing fc version and it should be fine. You can know that fcle and fc are.

B

A

That's how it works so mind that, and the final step uh we need to do is to copy the examples repository you can do it like. You usually do copy the git repositories.

A

It's pretty simple, so you copy whoops either this line, if you're after authorized in github through https or this one, if you're using ssh keys for authorization.

A

Okay, I will give you a few minute minutes to be sure that you can follow.

C

We have a question in chat alexi if you want to take that in the in the meantime,.

C

Sergey is asked how long till the fluence blockchain layer is ready, and does that mean that until then um applications won't be decentralized.

A

uh Well uh eugene, do you want to answer that one.

B

um Yeah, it is already in the chat, so the idea is that, um because there is no incentivization, yet the network works similarly to how ipfs works and how ipfs provides you guarantees.

B

If there is no um nodes in the network on ipfs who will host your file, then it will just disappear if you don't host it by yourself, so the same thing in in the fluence network, so you can deploy your service on some available nodes, but you don't have any.

B

You know like a fundamental guarantees that this node will will continue to serve your services until you do it yourself, at least at one node, and this is why I I talked about example of community run application, so you have a community. You have a several people who committed to the hosters of of your application. Backhand.

B

You basically can achieve this decentralization goal. um Using this way, uh but after the blockchain layer exists, you will be able to have some. You know you will be able to pay people for hosting it, so you'll have more guarantees.

B

I hope I I hope I answered.

A

Thank you, so I hope the installation more or less complete once again, if you want to follow the workshop and copy paste.

A

Yeah, what kind of cool hack projects have been built previously on premiums? We have a repository of examples.

A

And here you can see all sorts of uh applications and examples you can start with. You can see here.

A

B

There's an example project.

A

So you can kick off your hacking. uh You can see here the integration with ethereum. You can see there an analog of blue filters. uh You can see here how to integrate with ipfs, how to use our webassembly port of sqlite and really there are a lot of things to do on and to learn about fluence.

A

So this is a good start and, as you evgeny mentioned before, we have a fluent path project. This one has a more rich backhand front end, so you can refer to this one too, it's pretty user friendly. So I think this could be also a great example.

A

A

Let's start with running a local fluency node, so we can so we have a node to deploy something to.

A

We can just paste this here. We can also uh pass arguments here, so we can see what uh the possible uh options for this. Let me just update. If you ever find yourself, if you you have used the flint's node before you can always update it by using docker pull, that's a good thing to do, because we deliver releases pretty often and it changes pretty often so if anything doesn't work for some reason, you always start with dr pull first or ask question right away in the discord.

A

So yeah you can see that there is a new uh option available. It's called local. It allows you to run a node, that's that is not connected to the network.

A

Can be done like this, so here.

B

A

Node, you can see it's locks and it shows you that if you have any troubles you can do docker full and go to our discussions or discord to ask questions. A few interesting thing you can see here also is the peer id of the node that we just started. You will need it to deploy services and to manage them and to call them. So it's a really important information.

A

You can see here the version of the air interpreter. That's a webassembly module too. So everything is portable and the very same air interpreter is running in the browser when you're using print.jssdk. So it's the same on the node and in the browser, I think, that's nice. So let's open a new tab and let's return to our workshop, so we have started a local fluence node.

A

Now, let's check that it works first. What we'll do is we will create a simple air script that does really almost nothing besides. It's telling the network to go to the relay node and then to take the result back to the client.

A

uh The role of the client plays fl dist. So foggiest is the client here and relay is our mode. This construction means that we just copy pasting this code to the file under the name script layer.

A

So we just select this paste into the terminal and then we.

A

Select this so the uh what we just done. We, I have created a file script layer. It contains the script with two lines: go to the relay: uh save hello as a variable under name result and then send these result to the client to the init pure id. So it's the it. This is the id of the initiator of the script. So whoever would run this script, he would receive the result back and the result is a variable that contains a simple string, hello.

A

B

A

Example to execute the script we will use fojis run air command.

A

We will need to specify by a path to the script and uh address of the node by default it it uses uh address of the node in the network, but here we want to use address of the node of our local node. That's why we specify here our localhost, and here we specify localhost. Then we specify a port. We say that you should expect websocket on the port four nines, and this way we specify the pure id of the node, so it works as a public key.

A

This way we have our connection to the node, secure and encrypted. The uh peer id can be found here in the logs of the node. So let's copy paste that command up to here.

A

Okay and then we can just insert the peer id.

A

And let's see if it works, so what happened? uh I have executed the command, so I sent this script to the network.

A

You can see that my that I some metadata is printed and you can see the id of the script of the particle with the script, but that doesn't matter what matters is that we have received the result back. So this script has worked.

A

We sent the we asked relay to generate us some data and send it back to us as a result.

A

So here we have our result. You can see that it's wrapped in array like just like here.

A

Okay, and also we can see that- let me repeat that again, so we can see the logs. So here we can see the node that the node received this particle and the executed script. You can see that we have executed the op identity operation. It basically means return. Identity is a mathematical operation of returning back what it received. So if you call identity one, you receive one back. If you call identity hello, you receive hello back, so you can see that our node executed of identity and send particle back to us.

A

A

Now we have checked that we have a working node uh if you haven't cloned the examples repository yet it's the time to do that, because we will need it right now. So, let's clone it. If you haven't and let's go to the folder.

A

Of the examples, once we are in the examples repository you, we should go to the curl underscore template folder. Here we will see a project that contains of backhand and front-end parts right now we will be interested only in the backend part to learn how to deploy backend to our uh local node.

A

uh If you wish to learn how to connect front-end to the back-end, you can refer to the previous work, okay, so the interesting. uh So what will happen now? We need to deploy the service how that works. Let's see, let's here we have a deploy.sh script. That does all the work, but if we don't know what's inside, that's not interesting right. We need to know how it works to understand how the deployment process works. Interference network, so what's happened, what's happening here.

A

First, we have some checks uh that everything is installed like fc and the full beast. Then what we do we go to the backend folder and we call fc build minus minus release. What did what this tool does? Is? It builds the rust project and outputs a webassembly file. Let's do that manually.

A

First, let's go to the back end and we can see that it's a simple rust project.

A

It contains cargotome with legal dependencies. Let me show you: it depends on the fluence rust sdk and we have a new version for it and that's it and we can take a look at what's inside of this service.

A

It has only one file on the main file, because it's a pretty simple service that does what it does. It goes to the url and returns the result, so it queries a given url and returns. You back a result as a string and that's basically so um yeah back to deploy sage.

A

Here we have common to build the backend. You can paste it here and it will build. We can find here.

A

A

A

Here it is uh that's built and uh what for what happens? Next, we check here in the script that this wasn't really exist or we print error, and then we call a flgs new service common to deploy to deploy the service on the network.

A

Here you can see that we are also referring to some. We can config json, let's review it, so we can understand what's happening there here we can see. This is the configuration file for the service we specify here the name. It can be anything and we also specify uh mounted binary. So the what, when we saw this rust core, what it actually did.

A

It called the binary on the host in our case in the fluence container, and it simply called the curl comment, so it was just almost the same as if you do it by hand. So if you do something like this.

A

It will return the results and that's exactly what's happening here in the rust code, except that it it's uh in a sandbox and the only way to get access to the curl command is to uh explicitly give permission to call the curl binary here under the name curl. So what this thing in the config does this section? It allows service to call the curl.

A

Binary once we have a.

B

A

File that was module and this json configuration file. We can deploy the service, but here, if we just execute this command, uh it will deploy our service not to the local node but to the actual network.

A

So we need to add uh node address there to to switch to the local node.

A

So, let's see what we have used.

A

Previously so we was, we were using this address, it hasn't changed because we haven't restarted, no so yeah, so we just simply add this parameter. Here I put a slash backslash here, so I can show you uh on the next line, but you can just go without it. You can just append to the end.

A

So, let's now execute the deploy sh file.

A

Go back to the root uh of the curl template project, so here we in the examples, slash curl template and we execute the deploy dot sh.

A

So we deploy the service to the node. You can see. uh Several calls were executed here uh of, add module and blueprint. Blueprint is a recipe for the service.

A

That's gets created under the hood and finally, you can see that srv create was executed to create the service. It's all happening inside air scripts.

A

So this really can be easily repeated manually without default list, but default list uh makes it easier and you don't need to remember all the names for a certain create distance blueprint and all the stuff, so it just automates the easy stuff for you.

A

So uh we have executed, deploy sh script, which in turn executed fl this new service. Like this and after service creation, we got back service id this. This id is very important. You will need it to. You will need to know it to call services on the network.

A

If you want to call a service, you need to know where it's deployed on the in the network and its service id. So you can call it.

B

A

So, let's just call the let's call our service, but first, let's review the api we're going to call on the service. As I said before, it's a really simple service that all it does. It provides a request, function that accepts a single string, parameter and returns string result.

A

Of the url request that was performed.

A

So I, since it's only one function, that's available there. We will call it let's to call a service. We need a an air script, it's almost the same as before, except that, instead of identity, we now have service request.

A

Here you can see that service is a variable. uh Everything, that's not in the double quotes or person. Signs is a variable or instruction.

A

So here you can see that service is a variable, so we will need to set it, and also url is a variable, so we will need to somehow provide it to the script, to set it somehow.

A

So what this script does it says: go to relay call function, request on this service and pass this url as a as an argument to the function and once you get the result, put it to the result, variable and return to the claim.

A

This line is the same as before. It just says that we should return the result to the client, so let's create that script. The same way as before.

A

So it worked, and now we need to call that script first. What we need to do is to change the service id here to the one that were was created when we created the service after the deploy sh call, so we copy the service id. We paste it here.

A

You can see that our argument, the variable url- is set to this url. It will search through that that go for definition of condensed milk and also we need to change the period here of node. Let me do.

A

B

A

Are we have the full command we just copied.

A

And what happened so we executed this command. We passed passed, correct, node address, we passed argument variables as json object and we provided a path to the script to execute.

A

We send the effort, this send the script to the node and node returned as a result. So it's a simple request response topology here.

A

B

A

We were returning string result structure from the re request, function, string.

C

A

Structure has the following: well structure: it has the following fields: error, a return code, stdr and std out, since there was no errors during the execution, we have only std out field and we have return code code 0. That means success can return your own structure. Any structure that you want, except enough, can be returned from the function and you can manipulate it directly in the script.

A

I will show you how in a moment, so here you can see that we have uh exactly what has been returned by the curl, uh but as a string here in std out. So you can tell that it's as simple as working with binaries and it allows you to access external network resources easily.

A

Okay, so maybe some any questions so far.

A

A

So basically, we just executed a curl binary uh that queried the following url and returned its result. It's a simple, really really simple example, but uh since it's a general purpose uh platform that where you can program and express any logic, you want, you can build any service uh in rust and you can build many of them and compose all of them with air scripts. You can compose here. You can see that the calls are sequential, but it's possible to have them in parallel, it's possible to iterate over arrays or lists of results.

A

So really almost any logic can be expressed with the air scripts and rust.

A

Just to show you a simple example of composing services: um let's well compose the single service by calling it twice so we'll we'll do the same thing as before, except the script. Now is a little bit more complex. We have a nested seek instruction. That means that we start sequential execution. Then we start another sequential execution that takes only two instructions and we execute everything one instruction after another.

A

Here we we have a language reference in the documentation. Documentation is linked in the resources section below. I will show it to you, so what we have we have here a call to the duckduckgo about the same condensed milk, and we have a call to doctor go about definition of beetroot and here, as you remember, we have not just a string as a result, but but a structure that contains std out field, sdr and else and other field other fields.

A

So here we can manipulate the data fields. We can access them directly in the script, so this makes it possible to have your data manipulated and changed without any central authority.

A

Since this script is attached to the packet and sent along along the particles and executed by every node, it means it meets. You can have this logic executed on the network, not on some central authority.

A

Okay, let's paste this script to the file.

A

Let's check that it worked okay, great and let's run it except let's change the those two things. First, we will need a mod address from our previous call.

A

Let me zoom out a little, so it doesn't jump all around.

A

So I'm using here the service id from our deployment and the node address of the node that we run locally in the docker container.

A

Something went wrong.

A

A

Okay, so here we have a lot of output, but the essence is the same. We have our script executed and we have two calls there to our service on the node and we have returned two results this time. Looking out, so it's easily seen that we have two abstracts here: first about condensed milk and second about beetroot, and here you can see that all that this data was produced by std out fields.

A

This is actually a security metadata. I was talking in detail about it and how to secure your applications with that metadata on the previous workshop. So please refer there to know more about how to build, authorized applications.

A

So we have a composition uh here, it's a single service, but it can be different service. It can be tens of services. You can compose different ethereum oracles and calculate prices between them and take the one you trust the most and send it to your application. For example, you can really do anything, you can imagine, that's it. It's a short one this time. So um let me just review resources.

A

So, as eugenie already said, we have a dashboard. So every time you deploy the services to the node, they will appear here.

A

They will appear in this list of services, and here, if you run your node without the minus minus local at the end, and you have external access to your ip address, your node will appear here too.

A

We have a documentation that the documents how to use air- we didn't go much detail this time about it, but here you can read about all the instructions, uh all the possibilities for examples: how to handle errors when they happen and how to iterate through the arrays in your scripts.

A

um It's pretty nice structured. So you can learn a lot about films from here. We can so yeah and please refer to the previous workshop.

A

It was deeper in details and we have a recording with the examples uh repository you already know, and if you have any questions you can go to the discord uh in encode discord or to our github discussions here and we have more uh resources in our documentation here. uh Different youtube talks overviews. You can see our protocol paper on github, it's a really nice reading.

A

So yeah, that's it! Thank you.

C

Awesome, thank you very much alexi uh for that fantastic presentation and kind of workshop there uh just to put in the the chat. um This is a link to the the video and the resources from the the last workshop, which was um went into a bit more detail.

C

So, if you'd like to go and look at that, and perhaps follow that along and feel free to follow that uh in terms of the encode club, discord I'll send a link to that in chat now as well, and if someone from the fluence team would like to put the influence discord as well. I'm sure they'll be more than welcome to take any questions there, also um and in terms of uh hack, the system uh upcoming events wise.

C

We have one next week with polygon, uh who previously were called matic and they are a layer, two solution on ethereum. So if you're thinking of integrating ethereum uh but we're kind of scared of the high gas fees feel free to come along to that workshop uh next tuesday.

C

um But thanks again, everyone for coming today, thanks to the fluence team and particularly alexi and evgeny, for their presentations, and I hope you all have a great evening.