Azure IoTCentral SDK for MicroPython

An Azure IoT Central device client library for Micropython.

This repository contains code for the Azure IoT Central SDK for Micropython. This enables micropython developers to easily create device solutions that semealessly connect to Azure IoT Central applications. It can run on various boards with some tweaks for low-memory devices.

Prerequisites

• Micropython 1.12+ (recommended)

Import iotc

In most of the micropython capable boards, is sufficient to import library or install it if missing through upip.

try:
    import iotc
except:
    import upip
    upip.install('micropython-iotc')
    import iotc

The same commands apply when running through Micropython REPL.

NOTE: for low-end devices like the ESP8266, importing as external module can cause out-of-memory exception during execution because of the limited amount of heap space. For this kind of boards, putting the library on flash memory as a frozen module might be the only available option.

Details on how to build a custom firmware for specific boards with frozen modules can be found on official micropython github repository and website.

Samples

Check out the sample repository for example code showing how the SDK can be used in the various scenarios:

Connecting

Currently only connection through Shared Access Keys is supported. You can use both device keys or group keys.

Init

from iotc import IoTCConnectType
id_scope = 'scopeID'
device_id = 'device_id'
sasKey = 'masterKey' # or use device key directly
conn_type=IoTCConnectType.SYMM_KEY # or use DEVICE_KEY if working with device keys
client = IoTCClient(id_scope, device_id, conn_type, sasKey)

You can pass a logger instance to have your custom log implementation. (see #Logging)

e.g.

from iotc import ConsoleLogger,IoTCLogLevel
logger = ConsoleLogger(IoTCLogLevel.ALL)
client = IoTCClient(id_scope, device_id, conn_type, sasKey, logger)

Connect

client.connect()

After successfull connection, IOTC context is available for further commands.

Operations

Send telemetry

client.send_telemetry(payload,properties=None)

e.g. Send telemetry every 3 seconds

while client.is_connected():
    print('Sending telemetry')
    client.send_telemetry({'temperature':randint(0,20),'pressure':randint(0,20),'acceleration':{'x':randint(0,20),'y':randint(0,20)}})
    sleep(3)

An optional properties object can be included in the send methods, to specify additional properties for the message (e.g. timestamp,etc... ). Properties can be custom or part of the reserved ones (see list here).

NOTE: Payload content type and encoding are set by default to 'application/json' and 'utf-8'. Alternative values can be set using these functions:
iotc.set_content_type(content_type) # .e.g 'text/plain' iotc.set_content_encoding(content_encoding) # .e.g 'ascii'

Send property update

client.send_property({'fieldName':'fieldValue'})

Listen to events

Due to limitations of the Mqtt library for micropython, you must explictely declare your will to listen for incoming messages. This client implements a non-blocking way of receiving messages so if no messages are present, it will not wait for them and continue execution.

To make sure your client receives all messages just call listen() function in your main loop. Be aware that some sleeping time (200 ms +) is needed in order to let the underlying library listen for messages and release the socket.

while client.is_connected():
    client.listen() # listen for incoming messages
    client.send_telemetry(...)
    sleep(3)

You also need to subscribe to specific events to effectively process messages, otherwise client would just skip them (see below).

Listen to properties update

Subscribe to properties update event before calling connect():

client.on(IoTCEvents.PROPERTIES, callback)

To provide property sync aknowledgement, the callback must return the property value if has been successfully applied or nothing.

e.g.

def on_props(prop_name, prop_value):
    if prop_value>10:
        # process property
        return prop_value

client.on(IoTCEvents.PROPERTIES, on_props)

Listen to commands

Subscribe to command events before calling connect():

client.on(IoTCEvents.COMMANDS, callback)

To provide feedbacks for the command like execution result or progress, the client can call the ack function available in the callback.

The function accepts 2 arguments: the command instance and a custom response message.

def on_commands(command, ack):
    print(command.name)
    ack(command, 'Command received')

client.on(IoTCEvents.COMMANDS, on_commands)

Logging

The default log prints to serial console operations status and errors. This is the API_ONLY logging level. The function set_log_level() can be used to change options or disable logs. It accepts a IoTCLogLevel value among the following:

• IoTCLogLevel.DISABLED (log disabled)
• IoTCLogLevel.API_ONLY (information and errors, default)
• IoTCLogLevel.ALL (all messages, debug and underlying errors)

The device client also accepts an optional Logger instance to redirect logs to other targets than console. The custom class must implement three methods:

• info(message)
• debug(message)
• set_log_level(message);

Cache Credentials

The IoT Central device client accepts a storage manager to cache connection credentials. This allows to skip unnecessary device re-provisioning and requests to provisioning service. When valid credentials are present, device connects directly to IoT Central, otherwise it asks provisioning service for new credentials and eventually cache them.

Provided class must have the following methods:

def retrieve(self):
    # fetch credentials. return a Credentials object

def persist(self, credentials):
    # save credentials object

One-touch device provisioning and approval

A device can send custom data during provision process: if a device is aware of its IoT Central template Id, then it can be automatically provisioned.

How to set IoTC template ID in your device

Device template id (a.k.a Model Id) is used when obtaining authorization codes for new devices and automatically assign them to the right template. By providing template id during credentials generation, user doesn't need to manually migrate or assign device from IoT Central site.

In order to get the unique identifier, open configuration page for required model under "Device templates" section.

Click on "View Identity" and in next screen copy model urn.

Then call this method before connect():

client.set_model_id(model_id)

Automatic approval (default)

By default device auto-approval in IoT Central is enabled, which means that administrators don't need to approve device registration to complete the provisioning process when device is not already created.

Manual approval

If auto-approval is disabled, administrators need to manually approve new devices. This can be done from explorer page after selecting the device

License

This samples is licensed with the MIT license. For more information, see LICENSE