SAT: perform extra validation on integer fields.

[alafanechere]

What

SAT perform schema validation of records against the catalog schema with the jsonschema library.
According to jsonschema a float value with a zero fractional part is a valid integer :

>>> from jsonschema import Draft7Validator
>>> schema = {'$schema': 'http://json-schema.org/draft-07/schema#', 'type': 'object', 'properties': {'metrics.conversion': {'type': ['null', 'integer']},}}
>>> record = {'metrics.conversion': 1.0}
>>> Draft7Validator(schema).is_valid(record)
True
>>> record = {'metrics.conversion': 1.1}
>>> Draft7Validator(schema).is_valid(record)
False

This relaxed definition is problematic because database destinations are more strict in terms of typing.
This situation prevents us from identifying schema inconsistencies in SAT when a field is declared as integers but the actual record values in our test account are 1.0, 2.0, 3.0 etc. It led this PR that introduced a schema inconsistency to be merged with successful acceptance tests.

How

A. Implement a custom check in the verify_records_schema function to make sure that a field defined as integer is an integer in python term's...
B. Find a way to force this validation with jsonschema