fix(riva): preserve long-dictation leading segments on interim divergence

[rbright]

Summary

• track audio_processed for interim Riva responses in stream state
• preserve divergent one-shot interim segments when audio progression indicates a real boundary
• keep existing chain/stability/punctuation boundary rules and add regression coverage for both commit and replace paths

Test Plan

• go test ./apps/sotto/internal/riva
• just ci-check
• nix build 'path:.#sotto'

Notes

• this is aimed at long dictation with pauses where early segments were being dropped
• heuristic added: commit one-shot divergent interim when audio advanced by >=0.75s and prior interim has >=3 words

Summary by CodeRabbit

• Improvements
  • Refined interim transcript handling during speech recognition to improve real-time text segmentation accuracy.
  • Enhanced transcript boundary detection to better account for audio processing progress during transcription.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

Configuration used: defaults

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 8307e65 and 8028b0c.

📒 Files selected for processing (4)

• apps/sotto/internal/riva/client.go
• apps/sotto/internal/riva/client_test.go
• apps/sotto/internal/riva/stream_receive.go
• apps/sotto/internal/riva/transcript_segments.go

---

📝 Walkthrough

Walkthrough

The changes enhance interim boundary detection in the Riva audio stream processing by introducing audio progress tracking. A new field tracks the last processed audio position, enabling the boundary detection function to consider meaningful audio advancement alongside existing chain updates and stability metrics.

Changes

Cohort / File(s)	Summary
Stream State Enhancement apps/sotto/internal/riva/client.go	Added lastInterimAudioProcessed field to Stream struct to track the audio position of the most recent interim transcript.
Boundary Detection Logic apps/sotto/internal/riva/transcript_segments.go	Expanded shouldCommitInterimBoundary signature to accept audio processing parameters; added advancedAudioBoundary helper function and new threshold constants (interimBoundaryAudioAdvanceSeconds, minInterimWordsForAudioBoundary) to enforce boundaries based on meaningful audio progress and word count.
Stream Processing Integration apps/sotto/internal/riva/stream_receive.go	Integrated audio progress tracking by resetting lastInterimAudioProcessed on final results and updating it during interim processing; passes both previous and current audio values to shouldCommitInterimBoundary.
Test Coverage apps/sotto/internal/riva/client_test.go	Added two new tests (TestRecordResponseCommitsOneShotInterimOnAudioAdvance, TestRecordResponseKeepsOneShotInterimWhenAudioAdvanceIsSmall) to validate interim boundary behavior; updated existing test calls to accommodate the expanded shouldCommitInterimBoundary signature.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~25 minutes

Possibly related PRs

• fix(riva): rebuild segment assembly for long dictation #4: Shares modifications to shouldCommitInterimBoundary logic and Stream state management for interim boundary handling.

Poem

🐰 A rabbit hops through audio streams,
Tracking time with sonic dreams,
Three words and 0.75 seconds more,
Mark boundaries we can't ignore! ✨

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 11.11% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly describes the main change: preserving segments during interim divergence in the Riva service, which aligns with the PR's objective to retain long-dictation leading segments when audio progression indicates a real boundary.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

• 📝 Generate docstrings (stacked PR)
• 📝 Generate docstrings (commit on current branch)

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch fix/riva-audio-boundary-segments

Warning

There were issues while running some tools. Please review the errors and either fix the tool's configuration or disable the tool if it's a critical failure.

🔧 golangci-lint (2.5.0)

Error: can't load config: unsupported version of the configuration: "" See https://golangci-lint.run/docs/product/migration-guide for migration instructions
The command is terminated due to an error: can't load config: unsupported version of the configuration: "" See https://golangci-lint.run/docs/product/migration-guide for migration instructions

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}