feat(native): expose context batch sizing controls

[leehack]

Summary

• Add ModelParams.batchSize / ModelParams.microBatchSize so native n_batch and n_ubatch can be tuned independently of contextSize while preserving legacy defaults.
• Wire resolved batch sizing into native context creation and prompt decode chunking to support smaller CPU-friendly batch settings safely.
• Add unit coverage for batch-size resolution and update docs/changelog, including llama.cpp semantics references for n_batch and n_ubatch.

Validation

• dart analyze
• dart test test/unit/core/models/inference/model_params_test.dart
• dart test test/unit/backends/llama_cpp/llama_cpp_service_test.dart

[leehack]

@maiguangyang Let me know if this fulfills your request.

[codecov-commenter]

Codecov Report

❌ Patch coverage is 86.48649% with 5 lines in your changes missing coverage. Please review.
✅ Project coverage is 76.54%. Comparing base (fd4ac69) to head (102c3bb).
⚠️ Report is 4 commits behind head on main.

Files with missing lines	Patch %	Lines
lib/src/backends/llama_cpp/llama_cpp_service.dart	91.42%	3 Missing ⚠️
lib/src/core/models/inference/model_params.dart	0.00%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #71      +/-   ##
==========================================
+ Coverage   76.51%   76.54%   +0.03%     
==========================================
  Files          66       66              
  Lines        8196     8216      +20     
==========================================
+ Hits         6271     6289      +18     
- Misses       1925     1927       +2     

Flag	Coverage Δ
unittests	76.54% <86.48%> (+0.03%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.