examples/aqua-examples/near-integration

NEAR + Fluence + Aqua Integrations

The build and deploy instructions as well as some of the Aqua code are outdated and the deployed services are no longer available. An updated version will be available soon.

Overview

We provide integration examples for both a Fluence JS peer based on the NEAR API JS and distributed Wasm services wrapping the NEAR RPC API. A NEAR CLI integration is planned for the near future.

In our examples we've been using the Aqua CLI aqua and Marine tooling (the Marine REPL mrepl and Marine CLI marine).

For the purpose of this tutorial, we'll be using Fluence's new fluence CLI tool, which wraps the CLIs you have already been using, e.g., the aqua CLI and Marine tooling CLIs (marine and mrepl), and brings additional features such as project template generation, wrapper generation for deployed services, project dependencies install. See the Fluence CLI docs for more information.

Prerequisites

Please make sure you have the latest Fluence CLI installed by running the following command:

npm install -g @fluencelabs/cli

Fluence JS NEAR Signing Peer

Signing transactions and messages is a critical operation both on- and off-chain and an integral part of most Web3 workflows. In Fluence's open, permissionless peer-to-peer network maintaining data privacy is a challenge. For example, passing the password for a keyfile or the private key itself is quite risky: while a peer-to-peer communication channel is end-to-end encrypted, the "end" of the channel is the node hosting the target service. Hence, a node can easily eavesdrop on decrypted traffic and abscond with your password or key and presumably, your funds. Of course, you can run your own node to eliminate such exploits. Rather than run a full-fledged Rust node for a limited scope requirement, a more advantageous solution might be to implement a Fluence Signing Node (FSN) with Node JS and Fluence JS, which is exactly what we have done for this example. While a Fluence JS peer does not allow for the hosting of arbitrary services at this point, it does allow to easily wrap the NEAR JS SDK and expose whatever interfaces you want to be used/composed with Aqua.

Implementing a Peer Node With Fluence JS and Aqua

As discussed in the documentation, we can use Fluence JS in a Node JS application resulting in a viable peer node of the Fluence p2p network. If you haven't, have a look at the documentation before you continue. To follow along the code below:

cd near-signing-node

In order to create our signing node, we wrap the NEAR JS SDK and, for a minimally viable experiment, expose the sendMoney and a couple non-signing functions.

In order to be able to expose sendMoney as an addressable service to Aqua, we need to implement the sendMoney interface and function in Aqua:

-- near_signing_node.aqua
func send_money(network_id:string, account_id:string, receiver_id:string, amount:string, password: string, node:string, relay:string) -> string:
    on node via relay:
        res <- NearSignerApi.send_money(network_id, account_id, receiver_id, amount)
    <- res

Note that we added additional parameters to our sendMoney implementation: near-api-js specifies the sendMoney function with two parameters -- the receiver id and amount. Since sendMoney is associated with Account, we need to add the from wallet address as well as the network_id to be able to activate the appropriate account on the desired NEAR network. In addition, our exposed sendMoney service runs on a peer-to-peer node and in order to be able to locate and execute the service, we need to provide the node's peer id and relay id. Finally, we guard our service with stylized authentication for which we use the password parameter. That is, on the peer <peer_id>, which we reach via the relay <relay_id> and assuming the checks out, we eventually execute:

sendMoney(receiverId: string, amount: BN)

Once we compile Aqua with the npm run compile aqua command, which writes the Typescript output into the /src/_aqua dir, we can then use the generated code, see src/_aqua/near_signing_node.ts, to implement our sendMoney service and any of the other interfaces specified in Fluence JS, which essentially follows the NEAR example:

// index.ts
async send_money(network_id: string, account_id: string, receiver_id: string, amount: string, password: string): Promise<any> {
        if (!this.password_checker(password)) {
            return Promise.resolve("Not Authorized")
        }
        const config = get_config(network_id, this._keyStore);
        const near = await network_connect(config);
        let account = await near.account(account_id);
        let tx_receipt = await account.sendMoney(receiver_id, amount);
        let result = Promise.resolve(tx_receipt);
        return result;
    }

We implement the send_money method for the class NearSigner implements NearSignerApiDef class, where NearSignerApiDef is generated code from the Aqua compilation and which we register (as an exposed service) in async main like so:

// index.ts
async function main() {
    // ...
    registerNearSignerApi("near", new NearSigner());
    // ...

For the complete implementation details, see src/index.ts. Before we test our code, please note that in this implementation the wallet credentials are presumed to be in the ~/.near-credentials directory of the machine/system that runs the Fluence Signing Node.

For testnet wallets, see https://wallet.testnet.near.org/ and https://docs.near.org/docs/develop/basics/create-account, to get started.

If you haven't setup Near locally, go to the Near documentation, install the near CLI as described in its github repo:

npm install -g near-cli

Login using the near CLI with the following command:

near login

You'll get the output:

Please authorize at least one account at the URL above.

Which account did you authorize for use with NEAR CLI?
Enter it here (if not redirected automatically):
Logged in as [ <near-account>.testnet ] with public key [ ed25519:<your-key... ] successfully

Upon a successful login you should have a local credentials:

ls ~/.near-credentials/testnet
<near-account>.testnet.json

Note the implementations of account_state and get_balance, which follow the same implementation pattern discussed above but actually do not require account or wallet access.

Running And Interacting With The Fluence Peer

Open two terminal windows in the ~/near-examples/near-signing-node/ directory to launch the peer and a client peer, respectively. Please note that you can use the peer with a local Fluence node or the testnet. For our purposes, we will be using Fluence's krasnodar testnet.

Install the dependencies with:

# setup the node
npm i

Then compile Aqua:

# compile aqua
npm run compile-aqua

Which produces output similar to:

> near-signing-node@0.1.0 compile-aqua
> fluence aqua -i aqua/ -o src/_aqua

Compiling
... done
2022.08.12 14:11:24:024 main INFO aqua.AquaCli.main:133
    Aqua Compiler 0.7.4-332
Result <path-to-examples>/aqua-examples/near-integration/near-signing-node/src/_aqua/near_signing_node.ts: compilation OK (4 functions, 2 services)

You can check the generated Typescript and AIR code in the src/_aqua directory. With our setup complete, let's start the peer:

# start the node
npm start

Which produces output similar to:

> near-signing-node@0.1.0 start
> node -r ts-node/register src/index.ts

PeerId:  12D3KooWRfvdqfDXw4yLnYLpHLrMm56M3G1nAbti4fDdEhgr5gp2
Relay id:  12D3KooWFEwNWcHqi9rtsmDhsYcDbRUCDXH84RC4FW6UfsFWaoHi
ctrl-c to exit

Please take note of the relay id and peer id for use in your client peer. In order to call the account_state method, open a new terminal window and navigate to the ~/examples/aqua-examples/near-integration/near-signing-node directory and execute:

fluence run \
    -i aqua -f 'account_state("testnet", "<near-account>", "lame_password", "<your-peer-id>", "<your-relay-id>")'

Replace <near-account> with your Near testnet account and <your-peer-id> and <your-relay-id> with the values provided by your peer output as discussed above. Once you've done that, the output should be similar to:

Running:
  function: account_state("testnet", "<near-account>", "lame_password", "12D3KooWRfvdqfDXw4yLnYLpHLrMm56M3G1nAbti4fDdEhgr5gp2", "12D3KooWFEwNWcHqi9rtsmDhsYcDbRUCDXH84RC4FW6UfsFWaoHi")
  relay: /dns4/kras-06.fluence.dev/tcp/19001/wss/p2p/12D3KooWDUszU2NeWyUVjCXhGEt1MoZrhvdmaQQwtZUriuGN1jTr
... done

Result:

{
  "amount": "199999827797124999999980000",
  "block_hash": "6MKaFMkDMqcZrG8toTdAhoqLXrxMJL1JmHWQDnshcstF",
  "block_height": 97232435,
  "code_hash": "11111111111111111111111111111111",
  "locked": "0",
  "storage_paid_at": 0,
  "storage_usage": 346
}

In the output above listed the called function (account_state) with its arguments as well as the relay used for the call. So, you can observe the context of the function call. And there's a result of the call, of course. In our case it displays the basic information for our Near account described in AccountView Interface.

Similarly, we can call our send_money service with Aqua:

fluence run \
    -i aqua \
    -f 'send_money("testnet", "<near-from-account>", "<near-to-account>", "10000", "lame_password", "<your-peer-id>", "<your-relay-id>")'

Replace the near-from-account and near-to-account account placeholders with your respective testnet wallets and the your-peer-id and your-relay-id with the values provided by your peer. Executing above Aqua statement produces a transaction receipt similar to the one below:

Running:
  function: send_money("testnet", "<near-from-account>", "<near-to-account>", "10000", "lame_password", "12D3KooWRfvdqfDXw4yLnYLpHLrMm56M3G1nAbti4fDdEhgr5gp2", "12D3KooWFEwNWcHqi9rtsmDhsYcDbRUCDXH84RC4FW6UfsFWaoHi")
  relay: /dns4/kras-03.fluence.dev/tcp/19001/wss/p2p/12D3KooWJd3HaMJ1rpLY1kQvcjRPEvnDwcXrH8mJvk7ypcZXqXGE
... done

Result:

{
  "receipts_outcome": [
    {
      "block_hash": "EzB5BiTVyqzJjqbTzRRKfQ2qWdj48F5vArVwtdwDmNaG",
      "id": "FmhBQNgwtvaJUxTEbPhqfsUSjuwjcYw4iLsb1LLsKDDH",
      "outcome": {
        "executor_id": "<near-to-account>",
        "gas_burnt": 223182562500,
        "logs": [],
        "metadata": {
          "gas_profile": [],
          "version": 1
        },
        "receipt_ids": [
          "9Q5mKcBgnoN1cM47nfwBRiN6vUoZa4vPhn6boyXZitNd"
        ],
        "status": {
          "SuccessValue": ""
        },
        "tokens_burnt": "22318256250000000000"
      },
      "proof": [
        {
          "direction": "Left",
          "hash": "EA96udAi8vcAdLHnbKPHTW4qKHnMXhJ4zjD4csETYk9r"
        },
        {
          "direction": "Right",
          "hash": "CNtBo5A3Sma7RrK2J9ntTq7p3v7fxS8zYmdXYWDCfscT"
        }
      ]
    },
    {
      "block_hash": "4WX1AZ9VSJDzjv1j6uHqNcz4W7iqz6XyiCqj7wpGn6h1",
      "id": "9Q5mKcBgnoN1cM47nfwBRiN6vUoZa4vPhn6boyXZitNd",
      "outcome": {
        "executor_id": "<near-from-account>",
        "gas_burnt": 223182562500,
        "logs": [],
        "metadata": {
          "gas_profile": [],
          "version": 1
        },
        "receipt_ids": [],
        "status": {
          "SuccessValue": ""
        },
        "tokens_burnt": "0"
      },
      "proof": []
    }
  ],
  "status": {
    "SuccessValue": ""
  },
  "transaction": {
    "actions": [
      {
        "Transfer": {
          "deposit": "10000"
        }
      }
    ],
    "hash": "2cCxw5RGTqD9UCwqth3Pe3FhRcYkqRimnzyhYWCBKjjA",
    "nonce": 96699860000005,
    "public_key": "ed25519:82CcWWRM9scav5hbqUVL4JZsBwagqFvjZrLDbaoiE9pr",
    "receiver_id": "<near-to-account>",
    "signature": "ed25519:2cmhrzp4PeKPcXE1vUW89krdTcdsApY3h6TT7CshWdrZMBLUjfJQF6pijzYcFUhpwArNQwDmD9GkVep9gYJTb4Hd",
    "signer_id": "<near-from-account>"
  },
  "transaction_outcome": {
    "block_hash": "DhS6KZzK9PdCqot2k4hewfAWkc7nQ9mnZM91XKZdVRkQ",
    "id": "2cCxw5RGTqD9UCwqth3Pe3FhRcYkqRimnzyhYWCBKjjA",
    "outcome": {
      "executor_id": "<near-from-account>",
      "gas_burnt": 223182562500,
      "logs": [],
      "metadata": {
        "gas_profile": null,
        "version": 1
      },
      "receipt_ids": [
        "FmhBQNgwtvaJUxTEbPhqfsUSjuwjcYw4iLsb1LLsKDDH"
      ],
      "status": {
        "SuccessReceiptId": "FmhBQNgwtvaJUxTEbPhqfsUSjuwjcYw4iLsb1LLsKDDH"
      },
      "tokens_burnt": "22318256250000000000"
    },
    "proof": []
  }
}

It's rather convenient to call the account_state and send_money functions. However, there's still a couple of things we need to remember and use such as <your-peer-id> and <your-relay-id>. As a matter of fact we can simplify a method call, and --json-service param of the fluence CLI comes in handy.

Upon the Near signing node start, it creates a JSON file with its peer id and relay id information we can use with the fluence CLI and in our Aqua code.

The js-services.json looks like:

{
  "name": "JsService",
  "serviceId": "JsService",
  "functions": [
    {
      "name": "get",
      "result": {
        "peerId": "12D3KooWRfvdqfDXw4yLnYLpHLrMm56M3G1nAbti4fDdEhgr5gp2",
        "relayPeerId": "12D3KooWFEwNWcHqi9rtsmDhsYcDbRUCDXH84RC4FW6UfsFWaoHi"
      }
    }
  ]
}

And the corresponding Aqua code for transfer_money looks slightly different: instead of using the function signature to pass the node and relay arguments, we're now injecting the values with JsService.get() into the function body:

-- near_signing_node.aqua
func transfer_money(network_id:string, account_id:string, receiver_id:string, amount:string, password: string) -> string:
    services <- JsService.get()  -- step 1
    on services.peerId via services.relayPeerId: -- step 2
        res <- NearSignerApi.send_money(network_id, account_id, receiver_id, amount, password) -- step 3
    <- res

On the step 1 using the defined function get() of the JsService service we get the information about our service (its peer id and relay id) which is wrapped in the Services data structure. On step 2 we specify, that the next block should be executed on our JS peer with the id we got on the 1st step (so we extract it using services.peerId) which we connect to thru a relay with the services.relayPeerId relay id. On the 3rd step we call the Near API service wrapper using NearSignerApi.send_money to call the corresponding NEAR API function sendMoney.

Those parameters values we injected in the function body are available thru the js-services.json and the following definitions in our Aqua code:

-- near_signing_node.aqua
data Services:
    peerId: string
    relayPeerId: string

service JsService("JsService"):
    get: -> Services

Considering all the above, the transfer_money call looks like:

fluence run \
    -i aqua \
    -f 'transfer_money("testnet", "<near-from-account>", "<near-to-account>", "10000", "lame_password")' --json-service js-services.json

The similar approach can be used for the account_state with its updated implementation account_state_view:

-- near_signing_node.aqua
func account_state_view(network_id:string, account_id:string, password: string) -> string:
    services <- JsService.get()
    on services.peerId via services.relayPeerId:
        res <- NearSignerApi.account_state(network_id, account_id, password)
    <- res

And we can call the account_state_view in a similar fashion as we've done for the transfer_money:

fluence run \
  -i aqua \
  -f 'account_state_view("testnet", "<near-account>", "lame_password")' --json-service js-services.json

You can use the Testnet Explorer to further investigate the token transfer you executed.

Summary

In this section, we implemented a basic Fluence peer that outlines an approach to shield our NEAR wallet keys from other network participants and to implement singing related functionality, such as the transfer_money token transfer method. Additional methods, such as the more generic sign transaction and deploy contract can be easily implemented this way and we are looking forward to your pull requests. Also note, that our simple string return can be vastly improved by adding the appropriate interfaces.

In the next section, we briefly discuss how a variety of NEAR methods can be implemented as distributed, hosted services for easy deployment, re-use and scalability.

Fluence Wasm NEAR Services

Operating your own node may not always be desireable for a variety of reasons ranging from costs to reuse to scalability and failover requirements. A core feature of the Fluence peer-to-peer network paradigm, of course, is the deployment of Wasm services to essentially any peer, given some hosting agreement, which allows for high portability as well as easy reuse and scalability as a "deploy and forget", low cost solution. Even if the operation of a node is deemed necessary, as outlined in our Signing Node example above, it still may make sense to split services into a self-operated peer for signing-related activities and some hosted Wasm services otherwise. Of course, Aqua allows you to seamlessly compose any one of the (exposed) services regardless of the deployment approach.

In order to create a NEAR Wasm adapter, we wrap whatever functionality we need from the NEAR RPC API in our Wasm module(s).

Creating And Deploying NEAR Wasm Services

In the services/near-adapter/modules directory, you find a minimal WASM adapter near-rpc-services for NEAR RPC API to get you started. Since we are connecting to on-chain resources via JSON-RPC, we need our service module to have access to cUrl, which we provide with the cUrl adapter:

// src/main.rs
#[marine]
#[link(wasm_import_module = "curl_adapter")]
extern "C" {
    pub fn curl_request(cmd: Vec<String>) -> MountedBinaryResult;
}

Let's have a look at the implementation of the network status method, which provides a fairly extensive snapshot of the network at the time in inquiry. Our adapter, or wrapper, implementation needs to envelope the RPC status endpoint and requires only one parameter: the network_id', e.g., testnet:

// src.main.rs
// <snip>
#[marine]
pub struct Result {
    pub stderr: String,
    pub stdout: String,
}

#[marine]
pub fn node_status(network_id: String) -> Result {
    let method = "status".to_string();
    let url = url_maker(network_id);
    let params = "[]".to_string();
    let curl_params: Vec<String> = rpc_maker(url, method, params);
    let response = curl_request(curl_params);
    Result {
        stderr: String::from_utf8(response.stderr).unwrap(),
        stdout: String::from_utf8(response.stdout).unwrap(),
    }
}

Note that we use the Result struct to capture the curl response.

We can interact with the node_status in REPL. Please make sure you are in the near-integration/services directory. Open the REPL with fluence service repl nearAdapter and get the following output:

fluence service repl nearAdapter
Making sure service and modules are downloaded and built... done
Welcome to the Marine REPL (version 0.18.0)
Minimal supported versions
  sdk: 0.6.0
  interface-types: 0.20.0

app service was created with service id = e5ae9c3e-60f8-4ae7-b434-6f8085246c1d
elapsed time 326.464043ms

1> call near_rpc_services node_status ["testnet"]
result: Object({"stderr": String(""), "stdout": String("{\"jsonrpc\":\"2.0\",\"result\":{\"chain_id\":\"testnet\",\"latest_protocol_version\":55,\"protocol_version\":54,\"rpc_addr\":\"0.0.0.0:4040\",\"sync_info\":{\"earliest_block_hash\":\"87rXaRN96eVGijmoxXMvKm9XAas1RedpHgh5ifaMfQne\",\"earliest_block_height\":96939089,\"earliest_block_time\":\"2022-08-07T05:20:29.139629926Z\",\"epoch_id\":\"9fvV3KdWb71CtFj6shiFrAgBfq4Zqk16reWBXSGRhgZy\",\"epoch_start_height\":97111890,\"latest_block_hash\":\"Dn6Q8cPogGmGj3m15t2e1c6hgbrBpmd8yrjegX5GU2Nb\",\"latest_block_height\":97136216,\"latest_block_time\":\"2022-08-09T10:12:28.971640387Z\",\"latest_state_root\":\"HBRDv6dEYwv1WEiwTDwPHzXBLmFG5c19YVqHGADH5gM7\",\"syncing\":false},\"validator_account_id\":null,\"validators\":[{\"account_id\":\"legends.pool.f863973.m0\",\"is_slashed\":false},
...
...
...
{\"account_id\":\"kuutamo.pool.f863973.m0\",\"is_slashed\":false},{\"account_id\":\"gargoyle.pool.f863973.m0\",\"is_slashed\":false}],\"version\":{\"build\":\"crates-0.14.0-148-g5228fb106\",\"rustc_version\":\"1.61.0\",\"version\":\"1.28.0-rc.3\"}},\"id\":\"dontcare\"}")})
 elapsed time: 539.529855ms

2>
...

As you can see, this is a straight mapping of the RPC response to the Result struct introduced above, which we can process in Aqua like so:

-- some example aqua file
data Result:
    stderr: string
    stdout: string

service NearRpcServices:
    node_status(network_id: string) -> Result

func rpc_foo(network_id: string, node: string, service_id: string) -> string: 
    on node:
        NearRpcServices service_id
        res <- NearRpcServices.node_status(network_id)
        if res.stderr:
            result <<- "call failed"
        else:
            result <<- res.stdout
        <- result

Before we can use our Fluence NEAR adapter, we need to deploy our Wasm modules to one or more host peers. We can do that with Fluence CLI:

fluence deploy

The fluence CLI will make sure that all required services and modules are in place, can be either downloaded or built. It gives us the deployment confirmation:

Making sure all services are downloaded... done
Making sure all modules are downloaded and built... done

Going to deploy services described in <path-to-examples>/examples/aqua-examples/near-integration/services/fluence.yaml:

nearAdapter:
  get: ./near-adapter
  deploy:
    - deployId: default


? Do you want to deploy all of these services? Yes
Deploying:
  service: nearAdapter
  deployId: default
  on: 12D3KooWCMr9mU894i8JXAFqpgoFtx6qnV1LFPSfVc3Y34N4h4LS
... done
Compiling <path-to-examples>/examples/aqua-examples/near-integration/services/.fluence/aqua/deployed.app.aqua... done

Currently deployed services listed in <path-to-examples>/examples/aqua-examples/near-integration/services/.fluence/app.yaml:

nearAdapter:
  default:
    - blueprintId: ef2c354407c29317657c47403630a0217601934f4ceda58b7642ef8985a983df
      serviceId: 5bc6970c-1f7a-4e24-bc93-519dd4c40daf
      peerId: 12D3KooWCMr9mU894i8JXAFqpgoFtx6qnV1LFPSfVc3Y34N4h4LS

Please note the helper generated in Aqua by the CLI in the .fluence/aqua/deployed.app.aqua file for future use in our Aqua. Let's have a look at our Aqua script in ./src/aqua/main.aqua:

-- aqua/main.aqua
func node_report(network_id: string) -> Result:
    services <- App.services()
    on services.nearAdapter.default!.peerId:
        NearRpcServices services.nearAdapter.default!.serviceId
        res <- NearRpcServices.node_status(network_id)
    <- res

Which we can run with the fluence CLI:

fluence run -f 'node_report("testnet")'

Which results in the following output:

Running:
  function: node_report("testnet")
  relay: /dns4/kras-02.fluence.dev/tcp/19001/wss/p2p/12D3KooWHLxVhUQyAuZe6AHMB29P7wkvTNMn7eDMcsqimJYLKREf
... done

Result:

{
  "stderr": "",
  "stdout": "{\"jsonrpc\":\"2.0\",\"result\":{\"chain_id\":\"testnet\",\"latest_protocol_version\":55,\"protocol_version\":54,\"rpc_addr\":\"0.0.0.0:4040\",\"sync_info\":{\"earliest_block_hash\":\"87rXaRN96eVGijmoxXMvKm9XAas1RedpHgh5ifaMfQne\",\"earliest_block_height\":96939089,\"earliest_block_time\":\"2022-08-07T05:20:29.139629926Z\",\"epoch_id\":\"9fvV3KdWb71CtFj6shiFrAgBfq4Zqk16reWBXSGRhgZy\",\"epoch_start_height\":97111890,\"latest_block_hash\":\"2DcwXKkZf175VExkDpryH73p99rHQjdu54u4Y5vfambK\",\"latest_block_height\":97136905,\"latest_block_time\":\"2022-08-09T10:30:51.592058687Z\",\"latest_state_root\":\"AmXivm3DbmRqU9fYQe6tDuC3ujz2UbdeYNeLmSRrxzsZ\",\"syncing\":false},\"validator_account_id\":null,\"validators\":[{\"account_id\":\"legends.pool.f863973.m0\",\"is_slashed\":false},
  ...
  ...
  ...
  {\"account_id\":\"stakingfacilities.pool.f863973.m0\",\"is_slashed\":false},{\"account_id\":\"kuutamo.pool.f863973.m0\",\"is_slashed\":false},{\"account_id\":\"gargoyle.pool.f863973.m0\",\"is_slashed\":false}],\"version\":{\"build\":\"crates-0.14.0-148-g5228fb106\",\"rustc_version\":\"1.61.0\",\"version\":\"1.28.0-rc.3\"}},\"id\":\"dontcare\"}"
}

Give the already implemented view_account_report and tx_status_report functions a try or add more methods from the RPC API -- we are looking forward to your pull requests.

Summary

We created portable Wasm modules to function as an adapter to NEAR's JSON-RPC framework, which can be distributed as hosted services to Rust peer nodes. These services may be used on their own or in conjunction with a specialized peer node, see above, taking care of signing tasks while shielding the secret (wallet) keys from preying eyes. Regardless of the implementation route taken, Aqua allows us to seamlessly compose and reuse our services regardless of the chosen deployment option.