Cache Field

[lmath]

To load into big query, we need to take some data that is in a json following some particular schema, and transform it into something suitable to be understood by big query. Each time we were:

1. looking up the schema associated with the schema key associated with the data
2. converting the schema to a Field

We expect that for a given schema key, that its corresponding Field remains the same unless the schema has changed. So, this field can be cached, as long as we have a means to ensure that if the schema is updated, so is the field in a timely manner.

To ensure that, we add a TTL on the cache, taking the same approach as described here: snowplow/snowplow-rdb-loader#1086

where for our Field cache,
key = (SchemaKey, TTL)
value = Field