POST request handling and indexing improvements

[ikreymer]

Description

This PR updates the indexing of POST and other non-GET requests to include the request body in the URL as a query.
This allows for improved replay fidelity where request->response matching can not rely on the URL alone, and must also compare (parts) of the request body. The changes include:

• POST application/json request are parsed and primitive json values are added to query string. Duplicate json values are also added with a special suffix, eg. {"a", "b", "c": {"a": "b"}} will be converted to query string a=b&a.2_=b (the suffix is hopefully one that is not commonly used).
• POST text/plain data is attempted to be parsed as json, then as binary
• PUT requests treated same as POST
• For all non-GET requests, __wb_method=<method> is also added, so post requests will have a __wb_method=POST in the query

This indexing approach is compatible with cdxj-indexer and the replay used in wabac.js/ReplayWeb.page

This is a breaking change for existing collections that already have POST requests. These will need to be re-indexed to get accurate POST request replay.

Motivation and Context

Many sites require accurate replay of POST and PUT requests to be able to successfully match requests to responses and accurate replay pages.

Screenshots (if appropriate):

Types of changes

• Replay fix (fixes a replay specific issue)
• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)

Checklist:

• My change requires a change to the documentation.
• I have updated the documentation accordingly.
• I have added or updated tests to cover my changes.
• All new and existing tests passed.