1. A, C, & E: ClearImageErrors, ImageError for Unknown Serials

[ingalls]

Background

• I have gone through and ESified the main API. I have hit a couple dependency issues with serverless as modules are increasingly going ES-Only. Unfortunately Serverless does support ES Modules directly and can cause issues at load as it directly injects handling code into the main runtime at build. This is not possible to work around at this time as the injected code is CommonJS.
• Switching to image based lambdas also allows us to choose a run-time of choice as AWS has been increasing the time it takes to roll out a new Package Runtime (ie: Node@20). Their documentation pushes folks to use a Dockerized approach for this problem.
• Finally, this unifies style between Exif-API, ANIML-Ingest, and now ANIML-API and future proofs the installation of dependencies as they go ES-Only.

Icon	Note
🟡	Implemented - Not tested
🟢	Tested

1.A Context 🟢

Add an API Endpoint to allow a user to clear Image Errors by a given Batch ID. I've confirmed this functionality works by manually creating an ImageError entry and calling the API per the below example.

Example

{
    query: `
        mutation ClearImageErrors($input: ClearImageErrorsInput!){
            clearImageErrors(input: $input) {
                message
            }
        }
    `,
    variables: {
        input: {
            batch: '123'
        }
    }
}

This PR should be merged after the following:
Ref: #102

1.C Context 🟢

Reject all images w/ unknown SNs in the image-ingest lambda, and write those errors to the ImageError collection.

1.E Redrive Functionality 🟢

Example

Allow user's to perform a redrive, moving SQS Messages from the Dead Letter Queue (DLQ) back into the processing queue.

{
  query: `
      mutation RedriveBatch($input: RedriveBatchInput!){
          redriveBatch(input: $input) {
              message
          }
      }
  `,
  variables: {
      input: {
          batch: 'batch-1d8cb6ca-06a9-4f43-8d6e-4c4324c182ed'
      }
  }