feat(benchmark): enhance BenchmarkResult to track duration in nanoseconds and add field count metrics

[Gezi-lzq]

No description provided.

[Copilot]

Pull Request Overview

This PR enhances the BenchmarkResult class to track duration in nanoseconds instead of milliseconds and adds comprehensive field count metrics to support more granular performance analysis of data processing operations.

• Upgrades timing precision from milliseconds to nanoseconds for better accuracy
• Adds field count tracking to measure processing cost per field
• Updates field count calculation algorithms with new cost models for strings and binary data

Reviewed Changes

Copilot reviewed 10 out of 10 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
AvroRecordBinderTest.java	Updates test expectations to match new field counting algorithm
DebeziumUnwrapTransform.java	Improves schema caching with proper key-based lookup and thread-safe fingerprint computation
RecordAssembler.java	Minor optimization caching field count to avoid repeated schema field size lookups
PerfTestCase.java	Modifies benchmark to return field counts and use nanosecond precision timing
FieldsPerformanceTest.java	Updates output formatting to display new field count metrics and removes logging dependency
FieldProfileRunner.java	New comprehensive field profiling tool for detailed per-field-type performance analysis
BenchmarkResult.java	Core changes adding nanosecond duration tracking and field count metrics with derived calculations
FieldMetric.java	Completely redesigned field counting algorithm with new cost models for different data types
RecordBinder.java	Fixes map field counting logic and removes redundant field existence checks
AvroValueAdapter.java	Refactors map conversion logic with better error handling and code deduplication

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}