feat(flagd-rpc)!: add events for rpc mode, some breaking config fixes

[aepfli]

Adding Events to the RPC client:

Breaking Change

• renamed resolver type to rpc instead of grpc to be compliant with flagd spec
• changed configuration values to milliseconds instead of seconds, to be compliant with flagd spec
• renamed timeout config value to deadline, to be compliant with flagd spec
• removed property offline source path poll interval as the general approach in other providers is to use retry_backoff_ms

closes: #56

[codecov]

Codecov Report

Attention: Patch coverage is 96.29630% with 4 lines in your changes missing coverage. Please review.

Project coverage is 93.92%. Comparing base (61e42e7) to head (68501d0).
Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
...enfeature/contrib/provider/flagd/resolvers/grpc.py	94.02%	4 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #108      +/-   ##
==========================================
+ Coverage   93.47%   93.92%   +0.45%     
==========================================
  Files          14       14              
  Lines         582      675      +93     
==========================================
+ Hits          544      634      +90     
- Misses         38       41       +3     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[aepfli]

@toddbaert @beeme1mr @open-feature/sdk-python-maintainers please have a look - this is not perfect - and i might made some errors, but i hope this is somewhat going into the right direction

[aepfli]

relates also to

• partially [flagd-RPC] implement event stream, reconnect and caching #55 (no caching)
• [flagd-RPC] Implement events #56
• [flagd-RPC] implement evaluation test suite #57

[toddbaert]

I can see the e2e running without the event scenario being skipped now! Very nice to have this feature working and the e2e tests for confidence.

I also went through all the breaking config changes and to me they seem in line with the spec (at least those that we've fully implemented).

I've changed the PR title slightly to (IMO) make the release notes a bit clearer.

[beeme1mr]

Thanks for adding backwards compatiblity with timeout. The other breaking changes should have virtually no impact because the currently released in-process mode only supported local file watching.

Awesome job! 🥳

[gruebel]

nice work, just one adjustment would be great.

[gruebel]

nothing you have to fix, but env_or_default() should return the correct type, when cast is set. will try to fix it in a separate PR.

[gruebel]

@beeme1mr feel free to merge when you are ok with your comments.