appointment-bot

Ethical bot that checks for website changes (like appointments or products being available) and alerts you via SMS when it notices one.

tl;dr: This repo has two lambda functions in it. One checks a website and reports if it has changed from a user-defined state. The other function sends an SMS to the user when triggered.

This can be used to monitor websites for a variety of reasons. You might be looking to make an appointment, or just want to know if a site changes (New blogpost or video, etc.) You can basically think of this as a lightweight notification layer for a website. I built this because I needed to book an appointment on an ancient government website that had no waitlist or schedule for appointment availability.

Limitations: This bot is very simple, and assumes the target site has some kind of static element to it. It can only handle basic logins, and can't handle MFA. Most importantly, by design it doesn't take any direct action on its findings.

Bots that scoop up appointments or buy out tickets are lame as fuck and make the world worse, don't build those.

Lambda set up + scheduling

This setup assumes you have some familiarity with AWS or aren't afraid to play around and test things out.

Tl;dr: Set up two lambdas for each python script. Trigger the website checker function via an eventbridge schedule, and give it permission to execute other lambdas.

You will need:

• An AWS account
• A Twilio API key + credits
• A lambda layer that contains the requests and twilio libraries
  • If you don't know how to make a layer, or want instructions for doing it via the CLI, see below.

1. Create a lambda function using python 3.9, add the alert_function.py code, and apply a layer containing twilio to it
2. Input your own API Key + messaging into the script and deploy it
3. Create another lambda function using python 3.9, add the checker_function.py code, and apply a layer containing requests to it
4. Add the first lambda function as a destination for this one. You don't have to properly configure this, it's just a quick way to grant this function permission to invoke other functions
5. Go to Eventbridge > Schedules and create a schedule that targets the chekcer_function.py function on a schedule that makes sense. I'd recommend 1 hour with +/- 15 minutes of randomness, which the GUI lets you add in

That's largely it. All that's left is to customize the scripts with your own credentials and targets. There are comments within the scripts that describe what to do.

Making a lambda layer from AWS Cloudshell

The below instructions let you build a lambda function entirely from the AWS CLI. This can save the hassle of doing it on your local machine. Someone will probably say you should do this as two separate layers. Maybe you should. This is the lazy way.

What we'll do here: Spin up a python venv, make a directory and download the packages we need there, zip up that folder, make an S3 bucket, move the zip to that S3 bucket, and then make a lambda layer out of the zip in that S3 bucket.

• Open AWS Cloudshell (Make sure you select the right region)
• Execute the following commands to make a director and start a python virtual environment
  • mkdir layers
  • cd layers
  • python3 -m venv venv
  • source venv/bin/activate
• Now, in the virtual environment, execute these commands
  • mkdir my_layer_name
  • cd my_layer_name
  • pip install requests -t .
  • pip install twilio -t .
  • cd ..
• Cool, all of the stuff we need is in the subdirectory, so let's zip it up, make a bucket, put it in that bucket, and turn it into a layer. You'll need to make your own S3 bucket name.
  • zip -r my_layer_name.zip my_layer_name
  • aws s3api create-bucket \ --bucket a-cool-and-unique-bucket-name \ --region us-east-1
  • aws s3 cp lazy_layer.zip s3://a-cool-and-unique-bucket-name/
  • aws lambda publish-layer-version \ --layer-name twilio_requests_layer \ --description "Python Layer for the requests and twilio libraries" \ --license-info "MIT" \ --content S3Bucket=a-cool-and-unique-bucket-name,S3Key=my_layer_name.zip \ --compatible-runtimes python3.9

You should now see the lambda layer available.