v1.18.1 — TTS: strip JA full-stops before Kokoro misaki phonemizer

[amariichi]

Summary

Fixes an audible artifact at the end of Japanese sentences spoken through Kokoro: misaki's pyopenjtalk-backed G2P maps the JA full-stop 「。」 (and its fullwidth ASCII twin 「．」) to an actual phoneme rather than silence, so Kokoro rendered chunk endings as a short "ye"-like sound.

Diagnosis

Confirmed on hardware:

• あ。 → user heard あ plus an extra sound
• あ。。。。。 → user heard あ followed by five repeated イヤ-like sounds (one per period), demonstrating each 。 reaches misaki and produces a phoneme

Fix

• New module tts-worker/src/tts_worker/kokoro_text.py exposes strip_japanese_silent_punctuation (regex [。．]+ → empty).
• KokoroEngine._to_ja_phonemes strips the input before calling misaki, and returns '' if the chunk becomes empty after stripping.
• KokoroEngine.synthesize_chunks now skips chunks that come back empty from _to_ja_phonemes (avoids misaki ja g2p returned empty phoneme output).
• Other JA punctuation (「、」「！」「？」「・」「…」) is intentionally left in place — those either drive prosodic pausing or have not been observed to produce artifacts.
• Shared text normalization (tts_worker.shared_text) is untouched, so the existing "preserve JA punctuation" contract there still holds; the strip is engine-local to Kokoro, parallel to how qwen3_text.py houses Qwen3-specific text prep.

Tests

tts-worker/tests/test_kokoro_text.py (8 tests): trailing 「。」, repeated 「。」 runs, internal+trailing, fullwidth 「．」, preservation of other JA punctuation, punctuation-only input → empty, empty input passthrough, ASCII passthrough.

Test plan

• Ran new + existing tts-worker tests locally — 25/25 pass.
• Live hardware test on AtomS3R after a fresh Kokoro restart: あ。 now sounds like あ only, no trailing artifact.
• Reviewer: confirm synthesize_chunks still finalizes correctly when every chunk happens to be punctuation-only (rare; the chunk loop simply produces no audio and returns the zero-sample fallback).

🤖 Generated with Claude Code