chore:DynamoDB vs S3 for TextractGenericAsyncSfnTask

Currently, the DynamoDB table used in TextractGenericSyncSfnTask is only used for the StepFunction Token hand off between the two Async Processing Lambdas and logging into CW. We can use a rather frugal way for this by using a temp file under a prefix something like processing_temp/, where we create a JSON file with naming convention {job_tag}.json with the same fields as the DDB table, and simply read that file to get the necessary info in subsequent processing. If we want to do it in a DB/Query way, we can use S3 Select to query the prefix and filter in where clause by job_tag but this adds un-necessary complexities especially since we can simply get the file by job_tag value and read it as a JSON object/dict.