Refactor AudioSegmentProcessor processAudioData

[ysdede]

Refactored processAudioData in AudioSegmentProcessor.ts to reduce complexity by extracting logic into helper methods. Added a behavior test to ensure no regression.

---

PR created automatically by Jules for task 11059173089517581752 started by @ysdede

[ysdede]

Review Summary (Post r1-to-master merge)

Status: MERGEABLE - Safe to merge into staging

Changes: Decomposes the monolithic processAudioData() method in AudioSegmentProcessor into focused helper methods:

• updateNoiseAndEnergy() - noise floor and recent energy tracking
• handleProactiveSplitting() - long segment proactive split logic
• updateStateMachine() - dispatcher for state transitions
• handleSilenceToSpeech() - silence to speech transition
• handleSpeechToSilence() - speech to silence transition with confirmed silence check
• finishSpeechSegment() - segment creation and stats recording
• handleStateContinuation() - continuing in current state

Assessment: Good structural refactor. The original processAudioData() was ~120 lines with deeply nested conditionals. Each extracted method is now 10-25 lines and has a clear single responsibility. No behavioral changes.

Note: Overlaps with PR #87 (which refactors AudioSegmentProcessor, specifically findSpeechStart). They touch different methods in the same file, so merge #85 first, then #87.

Recommendation: MERGE - Clean decomposition, merge before #87

[google-labs-jules]

Review Summary (Post r1-to-master merge)

Status: MERGEABLE - Safe to merge into staging

Changes: Decomposes the monolithic processAudioData() method in AudioSegmentProcessor into focused helper methods:

• updateNoiseAndEnergy() - noise floor and recent energy tracking
• handleProactiveSplitting() - long segment proactive split logic
• updateStateMachine() - dispatcher for state transitions
• handleSilenceToSpeech() - silence to speech transition
• handleSpeechToSilence() - speech to silence transition with confirmed silence check
• finishSpeechSegment() - segment creation and stats recording
• handleStateContinuation() - continuing in current state

Assessment: Good structural refactor. The original processAudioData() was ~120 lines with deeply nested conditionals. Each extracted method is now 10-25 lines and has a clear single responsibility. No behavioral changes.

Note: Overlaps with PR #87 (which refactors AudioSegmentProcessor, specifically findSpeechStart). They touch different methods in the same file, so merge #85 first, then #87.

Recommendation: MERGE - Clean decomposition, merge before #87

Acknowledged. Thank you for the review. I have noted the overlap with PR #87 and that this PR (#85) should be merged first.