I'm trying to mock a call to S3 using the aws-sdk-mock package. I want to mock getObject on a bucket, so that instead of createReadStream() creating a network stream from a bucket, it creates a readable stream from a local file in my test suite.

Here's how I'm mocking the s3 call:

s3Spy = AWS.mock("S3", "getObject", {
    createReadStream: () => {
        return fs.createReadStream(testDataFile)
    }
});

Here's the line where it is used:

const stream = s3.getObject(params).createReadStream();

I can trace stream with a console.log call, and it's clearly a readable:

console.log(stream)
Readable {
  _readableState: ReadableState {
    ...
  },
  _events: [Object: null prototype] {},
  _eventsCount: 0,
  _maxListeners: undefined,
  _read: [Function (anonymous)],
  [Symbol(kCapture)]: false
}

However, if I inspect it further, it's not a file readable:

console.log(stream.path)
undefined

Which createReadStream am I calling if it's not the one in my mock? How can I write this so that I pull in the mocked data that I want to use?