feat(plugins-soniox): surface per-run language segments

[rosetta-livekit-bot]

Summary

Fixes #5685 (and the follow-up source-side symptom raised in the comment thread, which @chenghao-mou approved bundling into the same PR).

Both halves are the same plugin bug: _TokenAccumulator._lang_segments is built per-run by the existing coalescing logic but then dropped in send_endpoint_transcript (and the interim path). The fix surfaces it through new SpeechData fields on the target side, and stops dropping it on the source side in non-translation mode.

Changes

• stt.SpeechData: add target_languages / target_texts (symmetric to existing source_languages / source_texts). Same LanguageCode coercion in __post_init__. Default None, so the addition is strictly additive for every other plugin.
• Soniox plugin, translation mode: populate target_* from final._lang_segments on FINAL_TRANSCRIPT and INTERIM_TRANSCRIPT / PREFLIGHT_TRANSCRIPT. Consumers now see the per-run target breakdown for code-switched two-way translation, e.g. target_languages=["en", "es"] / target_texts=["Hello, how are you?", " Estoy bien, gracias."] for the translation of "Hello, ¿cómo estás? I'm doing fine, gracias.".
• Soniox plugin, non-translation mode: populate source_* from the same accumulator (previously None). A code-switched ja + en utterance now surfaces source_languages=["ja", "en"] / source_texts=["こんにちは、私の名はサムです。", " My name is Sam."] -- matches what the SpeechData docstring already promised for "multi-language detection services".
• Refactor: extract a _lang_segments_to_fields helper to DRY the conversion across both modes and both event paths; the four duplicated inline list comprehensions collapse to one named operation. The predicate that distinguishes source from target became data-presence-based (final_original._lang_segments) rather than config-based (is_translation_mode is not None), which is what unified both halves cleanly.

SpeechData.text and SpeechData.language are unchanged for back-compat (still the full concatenation and the first translated/detected language, respectively).

Test plan

• 14 new unit tests in tests/test_plugin_soniox_stt.py covering:
  • SpeechData.__post_init__ target_languages coercion (strings → LanguageCode, None stays None, existing LanguageCode passthrough)
  • _TokenAccumulator._lang_segments per-run coalescing
  • _lang_segments_to_fields helper edge cases (empty → (None, None), non-empty → parallel lists with LanguageCode coercion)
  • Two-way translation, code-switched (the issue's canonical example)
  • One-way translation (single target run)
  • "none" untranslated chunk inside a translated utterance (asymmetric per-run list lengths)
  • Interim path: translation mode merging final + non-final per run on both sides
  • Interim path: non-translation mode populates source_* from final + non-final merged
  • Non-translation single-language: source_* populated, target_* None
  • Non-translation code-switched JA+EN: source_* carries the per-run breakdown
• Live-verified end-to-end against the real Soniox WebSocket API in console mode:
  • Translation mode, code-switched "Hello, ¿cómo estás? I'm doing fine, gracias." → text="Hello, how are you? Estoy bien, gracias.", target_languages=["en", "es"], target_texts=["Hello, how are you?", " Estoy bien, gracias."], "".join(target_texts) == text. Source side unchanged.
  • Non-translation mode, code-switched " こんにちは、私の名はサムです。 My name is Sam." → text=" こんにちは、私の名はサムです。 My name is Sam.", source_languages=["ja", "en"], source_texts=[" こんにちは、私の名はサムです。", " My name is Sam."], target_* correctly None. Interim events also surface the multi-language source breakdown progressively as the user code-switches.
• ruff format clean, ruff check clean, no new mypy --strict errors introduced in changed files.

Follow-ups (intentionally not in this PR)

• The final / final_original accumulator names are honest about routing today but the new target_* fields make their two-mode roles more glaring (final is "primary user-facing accumulator", final_original is "source-side accumulator that's empty in non-translation mode"). Worth a separate behavior-preserving rename PR to final_primary / final_source.
• The new target_* fields are wired in Soniox only; other translation-capable plugins (Gladia, Deepgram v2, AWS) can adopt them in follow-up PRs.

[changeset-bot]

🦋 Changeset detected

Latest commit: 96ea001

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 34 packages

Name	Type
@livekit/agents	Patch
@livekit/agents-plugin-soniox	Patch
@livekit/agents-plugin-anam	Patch
@livekit/agents-plugin-assemblyai	Patch
@livekit/agents-plugin-baseten	Patch
@livekit/agents-plugin-bey	Patch
@livekit/agents-plugin-cartesia	Patch
@livekit/agents-plugin-cerebras	Patch
@livekit/agents-plugin-deepgram	Patch
@livekit/agents-plugin-elevenlabs	Patch
@livekit/agents-plugin-fishaudio	Patch
@livekit/agents-plugin-google	Patch
@livekit/agents-plugin-hedra	Patch
@livekit/agents-plugin-hume	Patch
@livekit/agents-plugin-inworld	Patch
@livekit/agents-plugin-lemonslice	Patch
@livekit/agents-plugin-liveavatar	Patch
@livekit/agents-plugin-livekit	Patch
@livekit/agents-plugin-minimax	Patch
@livekit/agents-plugin-mistral	Patch
@livekit/agents-plugin-mistralai	Patch
@livekit/agents-plugin-neuphonic	Patch
@livekit/agents-plugin-openai	Patch
@livekit/agents-plugin-perplexity	Patch
@livekit/agents-plugin-phonic	Patch
@livekit/agents-plugin-resemble	Patch
@livekit/agents-plugin-rime	Patch
@livekit/agents-plugin-runway	Patch
@livekit/agents-plugin-sarvam	Patch
@livekit/agents-plugin-silero	Patch
@livekit/agents-plugin-tavus	Patch
@livekit/agents-plugins-test	Patch
@livekit/agents-plugin-trugen	Patch
@livekit/agents-plugin-xai	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR