README.md

EULAgy

Get a EULA read to you.

---

What EULAgy does

EULAgy takes a text file, converts it to an mp3 and generates a csv based on the Amazon Comprehend findings.

EULAgy is a utility that can be plugged into an application that will stream an mp3. Or an application that wants to analyze the key phrases in a text file.

And (shhhh, don't tell) EULAgy will work for any text file, not just EULAs :).

• EULAgy uses Amazon Polly to create an mp3 from EULA text
• EULAgy uses Amazon Comprehend to create a csv from EULA text
• EULAgy is designed to work on AWS Lambda, but can also be run from the command line

NOTE: EULAgy is an MVP and work to make it more flexible are in progress. In particular, code changes are required to remove hard-coded eulagy bucket name throughout the application!

Using

Environment variables

EULAgy requires a populated BUCKET_NAME environment variable for tests and production

AWS Lambda

See Deploy below. Place txt files in the root S3 bucket -- output will appear in the uploaded folder of the S3 bucket.

Develop

1. Install npm
2. npm install
3. npm test to validate setup!
4. npm run liveTest to run tests that actually hit S3
   • requires configured AWS CLI setup on box running tests -- (You'll need to have AWS credentials created, e.g., https://docs.aws.amazon.com/sdkref/latest/guide/file-format.html)
   • will incur AWS charges (up to several cents per full suite run in the cheapest AWS region)

More on testing

• Comprehend usage in the suite comprises over 98% of the cost and is executed from comprehendUtils.test.ts and app.test.ts
• app.test.ts executes the entire processing chain
• See the coverage directory for code coverage numbers after the tests are run

I chose to create "live" tests that hit AWS services instead of mocking. The application is small enough and the tests are inexpensive enough for it to be worth not spending my time mocking and potentially retooling mocks if anything changes on the AWS side. The live tests give me greater confidence that a packaged release will work.

Deploy

S3 bucket

1. Create a bucket in S3
2. Create a folder named uploaded in the bucket (see below for helper scripting)
3. Create an AWS role with AWSLambdaExecute, AmazonPollyFullAccess, AmazonS3FullAccess and ComprehendFullAccess
4. Add this policy to the bucket

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Principal": {
                "AWS": "arn:aws:iam::<user_id>:role/<role_name>
            },
            "Action": "s3:*",
            "Resource": "arn:aws:s3:::<bucket_name>/*"
        }
    ]
}

5. Replace bucket references in code with your bucket name (I'm working on making this configurable)

Lambda

1. Create a new Lambda function with the role you created
2. Give the function a Timeout of 10 seconds
3. Run tsc, then npm run zpack_x to create the artifact (a zip file) in the build directory. On Windows and WSL you may need to execute the script steps manually.
4. Upload the artifact to your Lambda function in AWS

Helper functions

Create bucket named eulagy

Bucket name defaults to eulagy. Will not create a bucket that already exists.

node lib/buckets createb [name]

List buckets

node lib/buckets