Fixes #637: Added a schema.json file

[Combo1]

This PR adds a schema.json.

The file is not used to verify the json file creation, but is only for documentation purposes, such that the user knows what a field means.

#637

[fabclmnt]

Hi @Combo1, thanks for contributing and choosing and open issue. I do have a couple of comments though.

Fundamentally, the PR doesn't say why this shape of fix. #637 asks for
documentation of the JSON report fields, which means that content belongs in
docs/reference/ as a page on the JSON report, not a JSON file somewhere in the library.

On the schema itself:

• The structure doesn't match the report. variables is a top-level
  field of BaseDescription, a sibling of table, not nested inside it.
  time_index_analysis is a TimeIndexAnalysis object or None, not a
  string. See src/data_profiling/model/description.py.
• It covers 3 of the 11 top-level fields. scatter, correlations,
  missing, alerts, package, sample and duplicates are absent.
• Nothing validates it. As written it's a hand-maintained copy of the
  serialization that will drift silently. If we keep a schema, it needs
  a test asserting a real to_json() payload conforms to it.

If you want to take this on properly: a docs/reference page describing
the JSON fields, generated from or checked against the dataclasses in
model/description.py, would close #637.