An Alexa skill for asking for Metrolink times
If you just want to use the skill, then it's currently pending approval by Amazon, but I shall update this shortly.
This works by scraping beta.tfgm.com in lieu of a proper API. It's definitely not affiliated with TfGM, Metrolink, or anyone like that.
You'll need Python 3.9 installed, and a way of running shell scripts.
./bootstrap.sh
This will set up a Virtualenv containing all of the necessary dependencies.
You will now need to run utilities/update-stop-names.py to generate
the stop names from the published TfGM data.
There aren't really any tests, but you can run ./test.py to run a test
interaction. There's also a skipped test in there which will test that
that all the stop names are accessible on the TfGM site.
If you want to run your own version on a real device, go to the
Alexa developer portal
and create a new skill. The Interaction Model tab can be populated
with information from the interaction-model/ folder.
Put intent-schema.json into the
intent schema box, then create a Custom Slot Type called TRAM_STOP_NAME,
and put stop-names.txt into that field.
Finally, the sample-utterances.txt
should go into the Sample Utterances box.
If you know a way to automate this, please let me know!
If you want to deploy this, then this is all ready for continuous delivery!
You'll need to have some AWS credentials locally
(virtualenv/bin/aws configure) which allow you to create CloudFormation
stacks etc, and you'll need to complete secrets.json
with your skill ID from the Alexa dev console, then you can run
./deploy.sh which will set up all your infrastructure etc as needed.
This is all defined as a template in
infrastruction/cloudformation.py.
Happy tram travelling!