This library is deprecated. The Aries and DIDComm community are pursing support for DIDComm v2, which this library does not support. Much of how we think about DIDComm connections and how to manage those connections has evolved since this library was in use. It is therefore recommended to seek out other libraries.
One such library can be found here: https://github.com/Indicio-tech/didcomm-messaging-python
The didcomm-messaging-python library is focused on DIDComm v2 but contains an updated version of the crypto
module of this library, providing the core pack and unpack operations for DIDComm v1. It also contains more complete interfaces for building out a more complete DIDComm v1 agent. These interfaces may be deprecated in the future as DIDComm v2 is adopted.
This repo contains an example Aries Static Agent Library in Python.
A static agent is a form of agent that can speak DIDComm at a basic level but its keys and connections are configured statically. Static Agents have a direct relationship with a single full agent. Static Agents do not have a wallet.
Examples of static agents may include:
- Remote administration interface for an agent
- IoT devices
- Relays
- OpenAPI to DIDComm translator
A static agent's configuration minimally consists of:
- Its own public and private key
- The public key of its full agent counterpart
- The endpoint of its full agent counterpart
It is up to you to secure your static agent's configuration. The examples included in this repository use command line arguments or environment variables to configure the agent for simplicity and demonstration purposes only. This is not recommended for production environments.
This library makes as few assumptions about it's running environment as possible. This includes few dependencies, assumptions about web frameworks, etc.
- Python 3.6 or higher
You will need to pair the static agent with a full agent capable of basic DIDComm to complete the examples. The Indy Python Reference Agent can be used as the full agent for these examples.
Create and activate python virtual environment:
$ python3 -m venv env
$ source env/bin/activate
Install dependencies and the library into the virtual environment:
$ pip install -e .
If you want to run the included tests, install the
test
feature with pip:pip install -e .[test]
Execute keygen()
:
$ python -c "import aries_staticagent; aries_staticagent.keygen()"
For full agent:
DID: <Base58 encoded DID>
VK: <Base58 encoded verkey>
For static agent:
VK: <the same Base58 encoded verkey>
SK: <Base58 encoded sigkey>
As the output implies, the first section is intended to be entered in on the full agent to configure
a static connection. The second section is used to configure the static agent. The verkey
(VK) in
the first and second section are the same key, representing the key the static agent will use
for the connection. The keygen
script does not generate the keys that the full agent will use.
If using the Indy Python Reference Agent, open the web interface and
click Add Static Connection
. Enter the information output by keygen.py
and a label of your
choice. The endpoint of the static agent is optional and must match the hostname and port you
configure for the static agent if running the web server example. After clicking Add
, a new
dialogue window will open with the information needed to now start up the static agent.
If you are using another agent that supports configuring a static connection, follow the instructions provided by that agent.
Start the static agent (in this case, exapmles/cron.py
):
$ python examples/cron.py --endpoint <the endpoint of the full agent> \
$ --endpointkey <the verkey output by the full agent> \
$ --mypublickey <the verkey output by keygen.py> \
$ --myprivatekey <the sigkey output by keygen.py>
In the full agent's BasicMessages, you should now see a message sent from the static agent script.
TODO: Include screencast of running the example with the Indy Python Reference Agent
Refer to the examples
directory for complete working examples of using this library.
from aries_staticagent import Connection
# endpoint, endpointkey, mypublickey, myprivatekey key are obtained through some form of static
# configuration
conn = Connection((mypublickey, myprivatekey), their_vk=endpointkey, endpoint=endpoint)
This will open a static connection with the full agent reachable at endpoint
and messages packed
for endpointkey
.
With the static agent connection a
, to send messages to the full agent:
conn.send({
"@type": "https://didcomm.org/basicmessage/1.0/message",
"~l10n": {"locale": "en"},
"sent_time": utils.timestamp(),
"content": "The Cron Script has been executed."
})
An asynchronous method is also provided:
await conn.send_async({
"@type": "https://didcomm.org/basicmessage/1.0/message",
"~l10n": {"locale": "en"},
"sent_time": utils.timestamp(),
"content": "The Cron Script has been executed."
})
Transport mechanisms are completely decoupled from the Static Agent Library. This is intended to
allow library consumers to choose which transport is appropriate for their use case. The
examples/webserver_aiohttp.py
example shows how one might use the aiohttp
library as an inbound
transport mechanism for the static agent:
from aiohttp import web
from aries_staticagent import Connection, utils
# ... Configuration omitted
# Create static agent connection
conn = Connection((args.mypublickey, args.myprivatekey), their_vk=args.endpointkey, endpoint=args.endpoint)
# Register a handler for the basicmessage/1.0/message message type
@conn.route("https://didcomm.org/basicmessage/1.0/message")
async def basic_message(msg, conn):
# Respond to the received basic message by sending another basic message back
await conn.send_async({
"@type": "https://didcomm.org/basicmessage/1.0/message",
"~l10n": {"locale": "en"},
"sent_time": utils.timestamp(),
"content": "You said: {}".format(msg['content'])
})
# aiohttp request handler
async def handle(request):
# Read request body and pass to Connection.handle
await conn.handle(await request.read())
raise web.HTTPAccepted()
# Register aiohttp request handler
app = web.Application()
app.add_routes([web.post('/', handle)])
# Start the web server
web.run_app(app, port=args.port)
As seen in this example, registering a handler for a DIDComm message is done using the
@conn.route('<message_type>')
decorator. Passing raw, unpackaged messages to the static agent
connection over the decoupled transport mechanism is done by calling conn.handle(<raw message>)
.
Static agents can only unpack messages sent by the full agent.