Generalize the Enumerable services such as SQS, SNS, Kinesis, Firehose, S3 and Glue Tables and creates intuitive mappings between them

[sam-goodwin]

This change generalizes the Enumerable services such as SQS, SNS, Kinesis, Firehose, S3 and Glue Tables and creates intuitive mappings between them.

Closes #1

Example (see: https://github.com/sam-goodwin/punchcard/blob/to-glue/examples/lib/stream-processing.ts)

/**
 * Persist Kinesis Stream data as a tome-series Glue Table.
 * 
 * Kinesis Stream -> Firehose Delivery Stream -> S3 (staging) -> Lambda -> S3 (partitioned by `year`, `month`, `day`, `hour` and `minute`)
 *                                                                      -> Glue Catalog
 */
const database = new glue.Database(stack, 'Database', {
  databaseName: 'my_database'
});
stream
  .toS3(stack, 'ToS3').enumerable()
  .toGlueTable(stack, 'ToGlue', {
    database,
    tableName: 'my_table',
    columns: stream.type.shape,
    partition: {
      // Glue Table partition keys: minutely using the timestamp field
      keys: {
        year: integer(),
        month: integer(),
        day: integer(),
        hour: integer(),
        minute: integer()
      },
      get: record => ({
        // define the mapping of a record to its Glue Table partition keys
        year: record.timestamp.getUTCFullYear(),
        month: record.timestamp.getUTCMonth(),
        day: record.timestamp.getUTCDate(),
        hour: record.timestamp.getUTCHours(),
        minute: record.timestamp.getUTCMinutes(),
      })
    }
  });