Integrates OpenSprinkler Pi to the AWS Alexa so that you can command your sprinkler.
This is how your commands are sent to the OSPI server:
Echo -> Alexa -> Lambda -> AWS IoT -> IoT client -> OSPI server
The project is organised into 2 parts, thing and controller.
A Thing is the Raspberry Pi running the IoT client and OSPI server.
The Alexa service is a controller that sends commands to the Thing. Other examples could be a Mobile App.
Deploying a Thing requires AWS IoT Thing provisioning then setting up our Thing to connect to AWS IoT.
Install the depdencies to setup AWS IoT, in this case it's boto3.
sudo pip install requirements.txt
Configure AWS credentials by following the instructions for AWS CLI configuration.
Provision our Thing:
cd thing/ospi
python setup-aws-iot.py
If all went well you'll see the following printed to the console:
- Certificate ARN
- Policy ARN
- Thing ARN
The certificate and keys for use on your Thing will also be saved to:
- opensprinklerpi-private.pem.key
- opensprinklerpi-public.pem.key (unused)
- opensprinklerpi-certificate.pem.crt
TIP: Use these values if you want to delete the provisioned resources.
Deploying to a Raspberry Pi Thing is simple, copy the thing/ospi directory onto your Raspberry Pi, then execute the following:
sudo python setup.py install
If successful you'll have:
iotcommand available- sample configuration at /etc/iot/iot.json
- systemd unit installed as iot.service
After successful installation you'll need to configure the IoT client.
Set the endpoint value to your AWS IoT endpoint located within the AWS IoT Dashboard.
You'll need to download the Verisign ca-certificate and save as ca-certificate.pem.crt.
curl -o ca-certificate.pem.crt https://www.symantec.com/content/en/us/enterprise/verisign/roots/VeriSign-Class%203-Public-Primary-Certification-Authority-G5.pem
Move ca-certificate.pem.crt, opensprinklerpi-private.pem.key and opensprinklerpi-certificate.pem.crt into /etc/iot/.
If you prefer to keep your certificates and key in an alternative location, you may their path within iot.json.
MD5 your OSPI password and set as ospi-password-hash in /etc/iot/iot.json:
echo "yourpassword" | md5
Specify your OSPI instance base URL, this would be the URL that you'd load to view it's web page eg. http://raspberry-pi.local:8080 or http://localhost:8080 if you're running the iot client on the same device as OSPI.
Ensure each of the following values in iot.json is modified to suit your setup:
- client-id
- endpoint
- ospi-password-hash
- ospi-base-url
A systemd unit configuration is deployed to the /lib/systemd/system directory it can be enabled:
sudo systemctl enable iot.service
The service can be started immediately with:
sudo systemctl start iot.service
Alexa Skill setup is a manual operation however the creation of the Lambda is automated.
Create an Alexa skill within developer.amazon.com. Use INTENT.txt, UTTERANCES.txt for field values within the skill configuration.
Setup the Alexa skill handler:
cd controller/alexa
pip install -r requirements.txt -t .
python setup-aws-alexa-lambda.py
TIP: The Role ARN and Lambda ARN are printed if you need to delete these resources.
- Open the AWS IoT platform and select "Test"
- Subscribe to Topic "home/sprinkler"
- Open Alexa Skill > Test
- Enter Utterance "turn on my sprinkler for 5 seconds"
- Confirm a message is received in AWS IoT Platform Test console
If connection fails with a timeout or a certificate error it's likely your AWS IoT certificate isn't setup correctly, please ensure a policy is attached and specifies the appropriate permissions and a Thing is attached.