Add TestLogData, remove public LogDataImpl and LogDataBuilder from log SDK

[jack-berg]

I'm working on refactoring some of the Log SDK to allow modifying log records in LogProcessor for #4550 (related spec PR 2681). I have a commit that includes all the necessary changes to make this happen, but in an effort to make the review more manageable, have split out one of the changes into this PR.

Currently there is a a single LogData interface which is accepted by LogProcessor and LogExporter. The goal is to instead pass a ReadWriteLogRecord interface (akin to ReadWriteSpan) to LogProcessor, while reserving LogData for LogExporter#export. This PR removes the public LogDataImpl and LogDataBuilder classes from the Log SDK, replacing them with a TestLogData autovalue class in :sdk:logs-testing for testing purposes. This makes LogData much more in line with SpanData.

[codecov]

Codecov Report

Merging #4635 (83d76cd) into main (2ef73c8) will decrease coverage by 0.00%.
The diff coverage is 100.00%.

@@             Coverage Diff              @@
##               main    #4635      +/-   ##
============================================
- Coverage     90.05%   90.04%   -0.01%     
+ Complexity     5066     5056      -10     
============================================
  Files           581      581              
  Lines         15618    15608      -10     
  Branches       1497     1498       +1     
============================================
- Hits          14065    14055      -10     
  Misses         1098     1098              
  Partials        455      455              

Impacted Files	Coverage Δ
...io/opentelemetry/sdk/testing/logs/TestLogData.java	100.00% <100.00%> (ø)
.../java/io/opentelemetry/sdk/logs/SdkLogBuilder.java	100.00% <100.00%> (ø)
...ain/java/io/opentelemetry/sdk/logs/SdkLogData.java	100.00% <100.00%> (ø)
.../java/io/opentelemetry/sdk/logs/SdkLogEmitter.java	100.00% <100.00%> (ø)
...metry/sdk/metrics/export/PeriodicMetricReader.java	87.14% <0.00%> (-2.86%)	⬇️
...y/exporter/internal/marshal/CodedOutputStream.java	71.00% <0.00%> (+1.18%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 2ef73c8...83d76cd. Read the comment docs.

[jkwatson]

seems reasonable to me. Just one question about some of the test code