Refactor: Java wrapper call instrumentation into focused helpers

[shreejaykurhade]

This PR refactors wrap_target_calls_with_treesitter in codeflash/languages/java/instrumentation.py to reduce complexity without changing instrumentation behavior. Issue #526

The original function handled several responsibilities in one place:

• filtering and ordering target calls
• precomputing byte/character offsets for replacement
• building timing/capture statements for instrumented calls
• generating replacement code for expression-statement calls
• generating prefix blocks for embedded calls
• applying back-to-front source rewrites

That made the function harder to review and riskier to modify, especially in a part of the codebase with a lot of exact emitted-source expectations.

What Changed

The repeated preparation and code-generation logic is now split into focused private helpers:

• _prepare_valid_calls
• _build_call_statements
• _build_expression_statement_replacement
• _build_embedded_call_prefix_lines

wrap_target_calls_with_treesitter now acts as the orchestration layer for:

• parsing the body with tree-sitter
• collecting candidate calls
• preparing valid call metadata
• iterating through calls in reverse order
• applying expression or embedded-call replacements

Why This Is Better

This improves maintainability by:

• reducing the responsibility surface of the main wrapper function
• separating metadata preparation from emitted-code construction
• making the expression-statement and embedded-call paths easier to reason about in isolation
• lowering the risk of future edits to one instrumentation path affecting another

This is intended to be a behavior-preserving refactor, not a logic change.

Behavior / Compatibility

No instrumentation behavior is intended to change.

In particular, this refactor preserves:

• lambda and complex-expression filtering
• call ordering and reverse-application semantics
• void vs non-void handling
• precise-timing vs non-precise-timing behavior
• emitted marker / serialization / SQLite-write structure

Validation

uv run pytest -q tests/test_languages/test_java/test_instrumentation.py
uv run ruff check codeflash/languages/java/instrumentation.py

Results:

Both Passed

Risk

Low.

This change is a structural refactor in an area already covered by exact-string Java instrumentation tests, and the refactor preserves the original control flow rather than redesigning the instrumentation logic.

[shreejaykurhade]

@Saga4 @KRRT7 please review.