Directly use InputStream for parsing response JSON.

[findgriffin]

Description

This change is a big refactor and the main party trick is that we can use HttpResponse<InputStream and then pass that to FaunaParser instead of having to do toString -> ObjectMapper.read(..).

Some other things are mixed in here like some follow-on fixes from the previous PR #143 and the QueryTags parse method was incorrectly expecting an array, I've implemented an E2E test to verify that the a=b,c=d format is what we really get.

I expect to have a follow-on PR that will completely remove ConstraintFailureWire, ErrorInfoWire and QueryResponseWire, associated dead code, and refactor the tests. I stopped short of doing that in this PR because a lot of tests need to be changed.

Motivation and context

I started out with the goal of removing ObjectMapper and the associated *Wire classes (ConstraintFailureWire, ErrorInfoWire, and QueryResponseWire) and I didn't get all the way there, but this PR is already pretty big.

How was the change tested?

Many modified unit tests, and I added some E2E tests.

Screenshots (if appropriate):

Change types

• • Bug fix (non-breaking change that fixes an issue)
• • New feature (non-breaking change that adds functionality)
• • Breaking change (backwards-incompatible fix or feature)

Checklist:

• • My code follows the code style of this project.
• • My change requires a change to Fauna documentation.
• • My change requires a change to the README, and I have updated it accordingly.

---

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[pnwpedro]

A few required and a few optional pieces of feedback.

In the desc, you say:

QueryTags parse method was incorrectly expecting an array

I don't see a parse method in main. AFAICT, the QueryTags deserializer was doing the right thing expecting a comma-delimited key-value pair syntax (foo=bar,baz=moo)