Skip to content

Ledger & Crypto APIs

In this section, we show you how to integrate the AEA with the Fetch.ai and third-party ledgers.

Ledger support

For a ledger to be considered supported in the framework, three abstract base classes need to be implemented:

  • the LedgerApi class wraps the API to talk to the ledger and its helper methods
  • the Crypto class wraps the API to perform cryptographic operations for the relevant ledger
  • the FaucetApi class wraps the API to talk to a faucet on a testnet

These three classes have their own registries, which allow the developer to import the relevant object where needed.

Ledger plug-in architecture

The AEA framework provides a plug-in mechanism to support ledger functionalities in an easily extendible way. At import time, the framework will load all the crypto plug-ins available in the current Python environment.

A crypto plug-in is a Python package which declares some specific setuptools "entry points" in its setup.py script. In particular, there are three types of entry points the framework looks up:

  • aea.ledger_apis, which points to instantiable classes implementing the LedgerApi interface;
  • aea.cryptos, which points to instantiable classes implementing the Crypto interface;
  • aea.faucet_apis, which points to instantiable classes implementing the FaucetApi interface.

This is an example of setup.py script for a ledger plug-in aea-ledger-myledger:

# sample ./setup.py file
from setuptools import setup

setup(
    name="aea-ledger-myledger",
    packages=["aea_ledger_myledger"],
    # plugins must depend on 'aea'  
    install_requires=["aea"], # add other dependencies...
    # the following makes a plugin available to aea
    entry_points={
        "aea.cryptos": ["myledger = aea_ledger_myledger:MyLedgerCrypto"],
        "aea.ledger_apis": ["myledger = aea_ledger_myledger:MyLedgerApi"],
        "aea.faucet_apis": ["myledger = aea_ledger_myledger:MyLedgerFaucetApi"],
    },
    # PyPI classifier for AEA plugins
    classifiers=["Framework :: AEA"],
)

By convention, such plug-in packages should be named aea-ledger-${LEDGER_ID}, and the importable package name aea_ledger_${LEDGER_ID}. In the example above, the package name is aea-ledger-myledger, and the importable package name is aea_ledger_myledger.

You can search for AEA ledger plug-ins on PyPI: https://pypi.org/search/?q=aea-ledger

Maintained plug-ins

At the moment, the framework natively supports the following three ledgers:

However, support for additional ledgers can be added to the framework at runtime.

Examples

  • Examples of how to interact with the crypto registry:
from aea.crypto.registries import crypto_registry, make_crypto, register_crypto

# by default we can use the native cryptos
fetchai_crypto = make_crypto("fetchai")

# we can check what cryptos are registered
crypto_registry.supported_ids

# we can also add a new crypto to the registry
register_crypto(id_="my_ledger_id", entry_point="some.dotted.path:MyLedgerCrypto")

# and then make it anywhere
my_ledger_crypto = make_crypto("my_ledger_id")
  • Examples of how to interact with the ledger API registry:
from aea.crypto.registries import ledger_apis_registry, make_ledger_api, register_ledger_api

# by default we can use the native ledger apis
CONFIG = {"network": "testnet"}
fetchai_ledger_api = make_ledger_api("fetchai", **CONFIG)

# we can check what ledger apis are registered
ledger_apis_registry.supported_ids

# we can also add a new ledger api to the registry
register_ledger_api(id_="my_ledger_id", entry_point="some.dotted.path:MyLedgerApi")

# and then make it anywhere
my_ledger_api = make_ledger_api("my_ledger_id")
  • Examples of how to interact with the faucet API registry:
from aea.crypto.registries import faucet_apis_registry, make_faucet_api, register_faucet_api

# by default we can use the native faucet apis
CONFIG = dict(poll_interval=1.0)
fetchai_faucet_api = make_faucet_api("fetchai", **CONFIG)

# we can check what faucet apis are registered
faucet_apis_registry.supported_ids

# we can also add a new faucet api to the registry
register_faucet_api(id_="my_ledger_id", entry_point="some.dotted.path:MyLedgerFaucetApi")

# and then make it anywhere
my_faucet_api = make_faucet_api("my_ledger_id")

The framework wraps all LedgerApi classes and exposes them in the LedgerApis classes. The framework also wraps the crypto APIs to create identities on both ledgers and exposes them in the Wallet.

The separation between the Crypto and LedgerApi is fundamental to the framework design. In particular, the object which holds the private key is separated from the object which interacts with the ledger. This design pattern is repeated throughout the framework: the decision maker is the only entity with access to the AEA's Wallet whilst LedgerApis are accessible by all skills.

Agent Land - Fetch.ai testnet for agents

Agent Land is our stable, public testnet for the Fetch Ledger v2. As such, most developers will be interacting with this testnet. This is specifically designed and supported for AEA development.

Parameter Value
Chain ID agent-land
Denomination atestfet
Decimals 18
Version v0.2.x
RPC Endpoint https://rpc-agent-land.fetch.ai:443
REST Endpoint https://rest-agent-land.fetch.ai:443
Block Explorer https://explore-agent-land.fetch.ai
Token Faucet Use block explorer

You can access more details on GitHub.

The configurations can be specified for the fetchai/ledger:0.18.0 connection.

CosmWasm supporting chains

The Fetch.ai networks use CosmWasm for smart contract support.

Currently, to use the smart contract functionality of the Fetch.ai network you have to install a CLI tool which is used by the AEA framework to perform some necessary actions for the smart contract functionality on-chain.

  1. Install Rust using the following command:
curl --proto '=https' --tlsv1.2 https://sh.rustup.rs -sSf | sh
  1. Update the configuration for Rust:
rustup default stable
cargo version
# If this is lower than 1.44.1+, update with:
# rustup update stable

rustup target list --installed
rustup target add wasm32-unknown-unknown
  1. Install fetchd:
git clone https://github.com/fetchai/fetchd.git
cd fetchd
git checkout release/v0.2.x
make install

# Check if fetchcli is properly installed
fetchcli version
# Version should be >=0.2.5
  1. Configure fetchcli:
fetchcli config chain-id agent-land
fetchcli config trust-node false
fetchcli config node https://rpc-agent-land.fetch.ai:443
fetchcli config output json
fetchcli config indent true
fetchcli config broadcast-mode block

Now fetchcli will be ready for use on your system.